diff --git a/common/misc/diffseq.h b/common/misc/diffseq.h new file mode 100644 index 0000000..b949793 --- /dev/null +++ b/common/misc/diffseq.h @@ -0,0 +1,532 @@ +/* Analyze differences between two vectors. + + Copyright (C) 1988-1989, 1992-1995, 2001-2004, 2006-2019 Free Software + Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + + +/* The basic idea is to consider two vectors as similar if, when + transforming the first vector into the second vector through a + sequence of edits (inserts and deletes of one element each), + this sequence is short - or equivalently, if the ordered list + of elements that are untouched by these edits is long. For a + good introduction to the subject, read about the "Levenshtein + distance" in Wikipedia. + + The basic algorithm is described in: + "An O(ND) Difference Algorithm and its Variations", Eugene W. Myers, + Algorithmica Vol. 1, 1986, pp. 251-266, + . + See especially section 4.2, which describes the variation used below. + + The basic algorithm was independently discovered as described in: + "Algorithms for Approximate String Matching", Esko Ukkonen, + Information and Control Vol. 64, 1985, pp. 100-118, + . + + Unless the 'find_minimal' flag is set, this code uses the TOO_EXPENSIVE + heuristic, by Paul Eggert, to limit the cost to O(N**1.5 log N) + at the price of producing suboptimal output for large inputs with + many differences. */ + +/* Before including this file, you need to define: + ELEMENT The element type of the vectors being compared. + EQUAL A two-argument macro that tests two elements for + equality. + OFFSET A signed integer type sufficient to hold the + difference between two indices. Usually + something like ptrdiff_t. + EXTRA_CONTEXT_FIELDS Declarations of fields for 'struct context'. + NOTE_DELETE(ctxt, xoff) Record the removal of the object xvec[xoff]. + NOTE_INSERT(ctxt, yoff) Record the insertion of the object yvec[yoff]. + EARLY_ABORT(ctxt) (Optional) A boolean expression that triggers an + early abort of the computation. + USE_HEURISTIC (Optional) Define if you want to support the + heuristic for large vectors. + It is also possible to use this file with abstract arrays. In this case, + xvec and yvec are not represented in memory. They only exist conceptually. + In this case, the list of defines above is amended as follows: + ELEMENT Undefined. + EQUAL Undefined. + XVECREF_YVECREF_EQUAL(ctxt, xoff, yoff) + A three-argument macro: References xvec[xoff] and + yvec[yoff] and tests these elements for equality. + Before including this file, you also need to include: + #include + #include + #include "minmax.h" + */ + +/* Maximum value of type OFFSET. */ +#define OFFSET_MAX \ + ((((OFFSET)1 << (sizeof (OFFSET) * CHAR_BIT - 2)) - 1) * 2 + 1) + +/* Default to no early abort. */ +#ifndef EARLY_ABORT +# define EARLY_ABORT(ctxt) false +#endif + +#define MAX(a,b) (((a) > (b)) ? (a) : (b)) +#define MIN(a,b) (((a) < (b)) ? (a) : (b)) + +/* Use this to suppress gcc's "...may be used before initialized" warnings. + Beware: The Code argument must not contain commas. */ +#ifndef IF_LINT +# if defined GCC_LINT || defined lint +# define IF_LINT(Code) Code +# else +# define IF_LINT(Code) /* empty */ +# endif +#endif + +/* As above, but when Code must contain one comma. */ +#ifndef IF_LINT2 +# if defined GCC_LINT || defined lint +# define IF_LINT2(Code1, Code2) Code1, Code2 +# else +# define IF_LINT2(Code1, Code2) /* empty */ +# endif +#endif + +/* + * Context of comparison operation. + */ +struct context +{ + #ifdef ELEMENT + /* Vectors being compared. */ + ELEMENT const *xvec; + ELEMENT const *yvec; + #endif + + /* Extra fields. */ + EXTRA_CONTEXT_FIELDS + + /* Vector, indexed by diagonal, containing 1 + the X coordinate of the point + furthest along the given diagonal in the forward search of the edit + matrix. */ + OFFSET *fdiag; + + /* Vector, indexed by diagonal, containing the X coordinate of the point + furthest along the given diagonal in the backward search of the edit + matrix. */ + OFFSET *bdiag; + + #ifdef USE_HEURISTIC + /* This corresponds to the diff --speed-large-files flag. With this + heuristic, for vectors with a constant small density of changes, + the algorithm is linear in the vector size. */ + bool heuristic; + #endif + + /* Edit scripts longer than this are too expensive to compute. */ + OFFSET too_expensive; + + /* Snakes bigger than this are considered "big". */ + #define SNAKE_LIMIT 20 +}; + +struct partition +{ + /* Midpoints of this partition. */ + OFFSET xmid; + OFFSET ymid; + + /* True if low half will be analyzed minimally. */ + bool lo_minimal; + + /* Likewise for high half. */ + bool hi_minimal; +}; + + +/* Find the midpoint of the shortest edit script for a specified portion + of the two vectors. + + Scan from the beginnings of the vectors, and simultaneously from the ends, + doing a breadth-first search through the space of edit-sequence. + When the two searches meet, we have found the midpoint of the shortest + edit sequence. + + If FIND_MINIMAL is true, find the minimal edit script regardless of + expense. Otherwise, if the search is too expensive, use heuristics to + stop the search and report a suboptimal answer. + + Set PART->(xmid,ymid) to the midpoint (XMID,YMID). The diagonal number + XMID - YMID equals the number of inserted elements minus the number + of deleted elements (counting only elements before the midpoint). + + Set PART->lo_minimal to true iff the minimal edit script for the + left half of the partition is known; similarly for PART->hi_minimal. + + This function assumes that the first elements of the specified portions + of the two vectors do not match, and likewise that the last elements do not + match. The caller must trim matching elements from the beginning and end + of the portions it is going to specify. + + If we return the "wrong" partitions, the worst this can do is cause + suboptimal diff output. It cannot cause incorrect diff output. */ + +static void +diag (OFFSET xoff, OFFSET xlim, OFFSET yoff, OFFSET ylim, bool find_minimal, + struct partition *part, struct context *ctxt) +{ + OFFSET *const fd = ctxt->fdiag; /* Give the compiler a chance. */ + OFFSET *const bd = ctxt->bdiag; /* Additional help for the compiler. */ +#ifdef ELEMENT + ELEMENT const *const xv = ctxt->xvec; /* Still more help for the compiler. */ + ELEMENT const *const yv = ctxt->yvec; /* And more and more . . . */ + #define XREF_YREF_EQUAL(x,y) EQUAL (xv[x], yv[y]) +#else + #define XREF_YREF_EQUAL(x,y) XVECREF_YVECREF_EQUAL (ctxt, x, y) +#endif + const OFFSET dmin = xoff - ylim; /* Minimum valid diagonal. */ + const OFFSET dmax = xlim - yoff; /* Maximum valid diagonal. */ + const OFFSET fmid = xoff - yoff; /* Center diagonal of top-down search. */ + const OFFSET bmid = xlim - ylim; /* Center diagonal of bottom-up search. */ + OFFSET fmin = fmid; + OFFSET fmax = fmid; /* Limits of top-down search. */ + OFFSET bmin = bmid; + OFFSET bmax = bmid; /* Limits of bottom-up search. */ + OFFSET c; /* Cost. */ + bool odd = (fmid - bmid) & 1; /* True if southeast corner is on an odd + diagonal with respect to the northwest. */ + + fd[fmid] = xoff; + bd[bmid] = xlim; + + for (c = 1;; ++c) + { + OFFSET d; /* Active diagonal. */ + bool big_snake = false; + + /* Extend the top-down search by an edit step in each diagonal. */ + if (fmin > dmin) + fd[--fmin - 1] = -1; + else + ++fmin; + if (fmax < dmax) + fd[++fmax + 1] = -1; + else + --fmax; + for (d = fmax; d >= fmin; d -= 2) + { + OFFSET x; + OFFSET y; + OFFSET tlo = fd[d - 1]; + OFFSET thi = fd[d + 1]; + OFFSET x0 = tlo < thi ? thi : tlo + 1; + + for (x = x0, y = x0 - d; + x < xlim && y < ylim && XREF_YREF_EQUAL (x, y); + x++, y++) + continue; + if (x - x0 > SNAKE_LIMIT) + big_snake = true; + fd[d] = x; + if (odd && bmin <= d && d <= bmax && bd[d] <= x) + { + part->xmid = x; + part->ymid = y; + part->lo_minimal = part->hi_minimal = true; + return; + } + } + + /* Similarly extend the bottom-up search. */ + if (bmin > dmin) + bd[--bmin - 1] = OFFSET_MAX; + else + ++bmin; + if (bmax < dmax) + bd[++bmax + 1] = OFFSET_MAX; + else + --bmax; + for (d = bmax; d >= bmin; d -= 2) + { + OFFSET x; + OFFSET y; + OFFSET tlo = bd[d - 1]; + OFFSET thi = bd[d + 1]; + OFFSET x0 = tlo < thi ? tlo : thi - 1; + + for (x = x0, y = x0 - d; + xoff < x && yoff < y && XREF_YREF_EQUAL (x - 1, y - 1); + x--, y--) + continue; + if (x0 - x > SNAKE_LIMIT) + big_snake = true; + bd[d] = x; + if (!odd && fmin <= d && d <= fmax && x <= fd[d]) + { + part->xmid = x; + part->ymid = y; + part->lo_minimal = part->hi_minimal = true; + return; + } + } + + if (find_minimal) + continue; + +#ifdef USE_HEURISTIC + bool heuristic = ctxt->heuristic; +#else + bool heuristic = false; +#endif + + /* Heuristic: check occasionally for a diagonal that has made lots + of progress compared with the edit distance. If we have any + such, find the one that has made the most progress and return it + as if it had succeeded. + + With this heuristic, for vectors with a constant small density + of changes, the algorithm is linear in the vector size. */ + + if (200 < c && big_snake && heuristic) + { + { + OFFSET best = 0; + + for (d = fmax; d >= fmin; d -= 2) + { + OFFSET dd = d - fmid; + OFFSET x = fd[d]; + OFFSET y = x - d; + OFFSET v = (x - xoff) * 2 - dd; + + if (v > 12 * (c + (dd < 0 ? -dd : dd))) + { + if (v > best + && xoff + SNAKE_LIMIT <= x && x < xlim + && yoff + SNAKE_LIMIT <= y && y < ylim) + { + /* We have a good enough best diagonal; now insist + that it end with a significant snake. */ + int k; + + for (k = 1; XREF_YREF_EQUAL (x - k, y - k); k++) + if (k == SNAKE_LIMIT) + { + best = v; + part->xmid = x; + part->ymid = y; + break; + } + } + } + } + if (best > 0) + { + part->lo_minimal = true; + part->hi_minimal = false; + return; + } + } + + { + OFFSET best = 0; + + for (d = bmax; d >= bmin; d -= 2) + { + OFFSET dd = d - bmid; + OFFSET x = bd[d]; + OFFSET y = x - d; + OFFSET v = (xlim - x) * 2 + dd; + + if (v > 12 * (c + (dd < 0 ? -dd : dd))) + { + if (v > best + && xoff < x && x <= xlim - SNAKE_LIMIT + && yoff < y && y <= ylim - SNAKE_LIMIT) + { + /* We have a good enough best diagonal; now insist + that it end with a significant snake. */ + int k; + + for (k = 0; XREF_YREF_EQUAL (x + k, y + k); k++) + if (k == SNAKE_LIMIT - 1) + { + best = v; + part->xmid = x; + part->ymid = y; + break; + } + } + } + } + if (best > 0) + { + part->lo_minimal = false; + part->hi_minimal = true; + return; + } + } + } + + /* Heuristic: if we've gone well beyond the call of duty, give up + and report halfway between our best results so far. */ + if (c >= ctxt->too_expensive) + { + OFFSET fxybest; + OFFSET fxbest IF_LINT (= 0); + OFFSET bxybest; + OFFSET bxbest IF_LINT (= 0); + + /* Find forward diagonal that maximizes X + Y. */ + fxybest = -1; + for (d = fmax; d >= fmin; d -= 2) + { + OFFSET x = MIN (fd[d], xlim); + OFFSET y = x - d; + if (ylim < y) + { + x = ylim + d; + y = ylim; + } + if (fxybest < x + y) + { + fxybest = x + y; + fxbest = x; + } + } + + /* Find backward diagonal that minimizes X + Y. */ + bxybest = OFFSET_MAX; + for (d = bmax; d >= bmin; d -= 2) + { + OFFSET x = MAX (xoff, bd[d]); + OFFSET y = x - d; + if (y < yoff) + { + x = yoff + d; + y = yoff; + } + if (x + y < bxybest) + { + bxybest = x + y; + bxbest = x; + } + } + + /* Use the better of the two diagonals. */ + if ((xlim + ylim) - bxybest < fxybest - (xoff + yoff)) + { + part->xmid = fxbest; + part->ymid = fxybest - fxbest; + part->lo_minimal = true; + part->hi_minimal = false; + } + else + { + part->xmid = bxbest; + part->ymid = bxybest - bxbest; + part->lo_minimal = false; + part->hi_minimal = true; + } + return; + } + } + #undef XREF_YREF_EQUAL +} + + +/* Compare in detail contiguous subsequences of the two vectors + which are known, as a whole, to match each other. + + The subsequence of vector 0 is [XOFF, XLIM) and likewise for vector 1. + + Note that XLIM, YLIM are exclusive bounds. All indices into the vectors + are origin-0. + + If FIND_MINIMAL, find a minimal difference no matter how + expensive it is. + + The results are recorded by invoking NOTE_DELETE and NOTE_INSERT. + + Return false if terminated normally, or true if terminated through early + abort. */ + +static bool +compareseq (OFFSET xoff, OFFSET xlim, OFFSET yoff, OFFSET ylim, + bool find_minimal, struct context *ctxt) +{ +#ifdef ELEMENT + ELEMENT const *xv = ctxt->xvec; /* Help the compiler. */ + ELEMENT const *yv = ctxt->yvec; + #define XREF_YREF_EQUAL(x,y) EQUAL (xv[x], yv[y]) +#else + #define XREF_YREF_EQUAL(x,y) XVECREF_YVECREF_EQUAL (ctxt, x, y) +#endif + + /* Slide down the bottom initial diagonal. */ + while (xoff < xlim && yoff < ylim && XREF_YREF_EQUAL (xoff, yoff)) + { + xoff++; + yoff++; + } + + /* Slide up the top initial diagonal. */ + while (xoff < xlim && yoff < ylim && XREF_YREF_EQUAL (xlim - 1, ylim - 1)) + { + xlim--; + ylim--; + } + + /* Handle simple cases. */ + if (xoff == xlim) + while (yoff < ylim) + { + NOTE_INSERT (ctxt, yoff); + if (EARLY_ABORT (ctxt)) + return true; + yoff++; + } + else if (yoff == ylim) + while (xoff < xlim) + { + NOTE_DELETE (ctxt, xoff); + if (EARLY_ABORT (ctxt)) + return true; + xoff++; + } + else + { + struct partition part IF_LINT2 (= { .xmid = 0, .ymid = 0 }); + + /* Find a point of correspondence in the middle of the vectors. */ + diag (xoff, xlim, yoff, ylim, find_minimal, &part, ctxt); + + /* Use the partitions to split this problem into subproblems. */ + if (compareseq (xoff, part.xmid, yoff, part.ymid, part.lo_minimal, ctxt)) + return true; + if (compareseq (part.xmid, xlim, part.ymid, ylim, part.hi_minimal, ctxt)) + return true; + } + + return false; + #undef XREF_YREF_EQUAL +} + +#undef ELEMENT +#undef EQUAL +#undef OFFSET +#undef EXTRA_CONTEXT_FIELDS +#undef NOTE_DELETE +#undef NOTE_INSERT +#undef EARLY_ABORT +#undef USE_HEURISTIC +#undef XVECREF_YVECREF_EQUAL +#undef OFFSET_MAX diff --git a/common/misc/fstrcmp.c b/common/misc/fstrcmp.c new file mode 100644 index 0000000..c61eab9 --- /dev/null +++ b/common/misc/fstrcmp.c @@ -0,0 +1,248 @@ +/* Functions to make fuzzy comparisons between strings + Copyright (C) 1988-1989, 1992-1993, 1995, 2001-2003, 2006, 2008-2019 Free + Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + + +#include + +/* Specification. */ +#include "fstrcmp.h" + +#include +#include +#include +#include +#include +#include +#include + +#include "glthread/lock.h" +#include "glthread/tls.h" +#include "minmax.h" +#include "xalloc.h" + + +#define ELEMENT char +#define EQUAL(x,y) ((x) == (y)) +#define OFFSET ptrdiff_t +#define EXTRA_CONTEXT_FIELDS \ + /* The number of edits beyond which the computation can be aborted. */ \ + ptrdiff_t edit_count_limit; \ + /* The number of edits (= number of elements inserted, plus the number of \ + elements deleted), temporarily minus edit_count_limit. */ \ + ptrdiff_t edit_count; +#define NOTE_DELETE(ctxt, xoff) ctxt->edit_count++ +#define NOTE_INSERT(ctxt, yoff) ctxt->edit_count++ +#define EARLY_ABORT(ctxt) ctxt->edit_count > 0 +/* We don't need USE_HEURISTIC, since it is unlikely in typical uses of + fstrcmp(). */ +#include "diffseq.h" + + +/* Because fstrcmp is typically called multiple times, attempt to minimize + the number of memory allocations performed. Thus, let a call reuse the + memory already allocated by the previous call, if it is sufficient. + To make it multithread-safe, without need for a lock that protects the + already allocated memory, store the allocated memory per thread. Free + it only when the thread exits. */ + +static gl_tls_key_t buffer_key; /* TLS key for a 'ptrdiff_t *' */ +static gl_tls_key_t bufmax_key; /* TLS key for a 'uintptr_t' */ + +static void +keys_init (void) +{ + gl_tls_key_init (buffer_key, free); + gl_tls_key_init (bufmax_key, NULL); + /* The per-thread initial values are NULL and 0, respectively. */ +} + +/* Ensure that keys_init is called once only. */ +gl_once_define(static, keys_init_once) + + +/* In the code below, branch probabilities were measured by Ralf Wildenhues, + by running "msgmerge LL.po coreutils.pot" with msgmerge 0.18 for many + values of LL. The probability indicates that the condition evaluates + to true; whether that leads to a branch or a non-branch in the code, + depends on the compiler's reordering of basic blocks. */ + + +double +fstrcmp_bounded (const char *string1, const char *string2, double lower_bound) +{ + struct context ctxt; + size_t xvec_length = strlen (string1); + size_t yvec_length = strlen (string2); + size_t length_sum = xvec_length + yvec_length; + ptrdiff_t i; + + ptrdiff_t fdiag_len; + ptrdiff_t *buffer; + uintptr_t bufmax; + + /* short-circuit obvious comparisons */ + if (xvec_length == 0 || yvec_length == 0) /* Prob: 1% */ + return length_sum == 0; + + if (! (xvec_length <= length_sum + && length_sum <= MIN (UINTPTR_MAX, PTRDIFF_MAX) - 3)) + xalloc_die (); + + if (lower_bound > 0) + { + /* Compute a quick upper bound. + Each edit is an insertion or deletion of an element, hence modifies + the length of the sequence by at most 1. + Therefore, when starting from a sequence X and ending at a sequence Y, + with N edits, | yvec_length - xvec_length | <= N. (Proof by + induction over N.) + So, at the end, we will have + edit_count >= | xvec_length - yvec_length |. + and hence + result + = (xvec_length + yvec_length - edit_count) + / (xvec_length + yvec_length) + <= (xvec_length + yvec_length - | yvec_length - xvec_length |) + / (xvec_length + yvec_length) + = 2 * min (xvec_length, yvec_length) / (xvec_length + yvec_length). + */ + ptrdiff_t length_min = MIN (xvec_length, yvec_length); + volatile double upper_bound = 2.0 * length_min / length_sum; + + if (upper_bound < lower_bound) /* Prob: 74% */ + /* Return an arbitrary value < LOWER_BOUND. */ + return 0.0; + +#if CHAR_BIT <= 8 + /* When X and Y are both small, avoid the overhead of setting up an + array of size 256. */ + if (length_sum >= 20) /* Prob: 99% */ + { + /* Compute a less quick upper bound. + Each edit is an insertion or deletion of a character, hence + modifies the occurrence count of a character by 1 and leaves the + other occurrence counts unchanged. + Therefore, when starting from a sequence X and ending at a + sequence Y, and denoting the occurrence count of C in X with + OCC (X, C), with N edits, + sum_C | OCC (X, C) - OCC (Y, C) | <= N. + (Proof by induction over N.) + So, at the end, we will have + edit_count >= sum_C | OCC (X, C) - OCC (Y, C) |, + and hence + result + = (xvec_length + yvec_length - edit_count) + / (xvec_length + yvec_length) + <= (xvec_length + yvec_length - sum_C | OCC(X,C) - OCC(Y,C) |) + / (xvec_length + yvec_length). + */ + ptrdiff_t occ_diff[UCHAR_MAX + 1]; /* array C -> OCC(X,C) - OCC(Y,C) */ + ptrdiff_t sum; + double dsum; + + /* Determine the occurrence counts in X. */ + memset (occ_diff, 0, sizeof (occ_diff)); + for (i = xvec_length - 1; i >= 0; i--) + occ_diff[(unsigned char) string1[i]]++; + /* Subtract the occurrence counts in Y. */ + for (i = yvec_length - 1; i >= 0; i--) + occ_diff[(unsigned char) string2[i]]--; + /* Sum up the absolute values. */ + sum = 0; + for (i = 0; i <= UCHAR_MAX; i++) + { + ptrdiff_t d = occ_diff[i]; + sum += (d >= 0 ? d : -d); + } + + dsum = sum; + upper_bound = 1.0 - dsum / length_sum; + + if (upper_bound < lower_bound) /* Prob: 66% */ + /* Return an arbitrary value < LOWER_BOUND. */ + return 0.0; + } +#endif + } + + /* set the info for each string. */ + ctxt.xvec = string1; + ctxt.yvec = string2; + + /* Set TOO_EXPENSIVE to be approximate square root of input size, + bounded below by 4096. */ + ctxt.too_expensive = 1; + for (i = xvec_length + yvec_length; i != 0; i >>= 2) + ctxt.too_expensive <<= 1; + if (ctxt.too_expensive < 4096) + ctxt.too_expensive = 4096; + + /* Allocate memory for fdiag and bdiag from a thread-local pool. */ + fdiag_len = length_sum + 3; + gl_once (keys_init_once, keys_init); + buffer = gl_tls_get (buffer_key); + bufmax = (uintptr_t) gl_tls_get (bufmax_key); + if (fdiag_len > bufmax) + { + /* Need more memory. */ + bufmax = 2 * bufmax; + if (fdiag_len > bufmax) + bufmax = fdiag_len; + /* Calling xrealloc would be a waste: buffer's contents does not need + to be preserved. */ + free (buffer); + buffer = xnmalloc (bufmax, 2 * sizeof *buffer); + gl_tls_set (buffer_key, buffer); + gl_tls_set (bufmax_key, (void *) (uintptr_t) bufmax); + } + ctxt.fdiag = buffer + yvec_length + 1; + ctxt.bdiag = ctxt.fdiag + fdiag_len; + + /* The edit_count is only ever increased. The computation can be aborted + when + (xvec_length + yvec_length - edit_count) / (xvec_length + yvec_length) + < lower_bound, + or equivalently + edit_count > (xvec_length + yvec_length) * (1 - lower_bound) + or equivalently + edit_count > floor((xvec_length + yvec_length) * (1 - lower_bound)). + We need to add an epsilon inside the floor(...) argument, to neutralize + rounding errors. */ + ctxt.edit_count_limit = + (lower_bound < 1.0 + ? (ptrdiff_t) (length_sum * (1.0 - lower_bound + 0.000001)) + : 0); + + /* Now do the main comparison algorithm */ + ctxt.edit_count = - ctxt.edit_count_limit; + if (compareseq (0, xvec_length, 0, yvec_length, 0, &ctxt)) /* Prob: 98% */ + /* The edit_count passed the limit. Hence the result would be + < lower_bound. We can return any value < lower_bound instead. */ + return 0.0; + ctxt.edit_count += ctxt.edit_count_limit; + + /* The result is + ((number of chars in common) / (average length of the strings)). + The numerator is + = xvec_length - (number of calls to NOTE_DELETE) + = yvec_length - (number of calls to NOTE_INSERT) + = 1/2 * (xvec_length + yvec_length - (number of edits)). + This is admittedly biased towards finding that the strings are + similar, however it does produce meaningful results. */ + return ((double) (xvec_length + yvec_length - ctxt.edit_count) + / (xvec_length + yvec_length)); +} diff --git a/common/misc/fstrcmp.h b/common/misc/fstrcmp.h new file mode 100644 index 0000000..119498d --- /dev/null +++ b/common/misc/fstrcmp.h @@ -0,0 +1,45 @@ +/* Fuzzy string comparison. + Copyright (C) 1995, 2000, 2002-2003, 2006, 2008-2019 Free Software + Foundation, Inc. + + This file was written by Peter Miller + +This program is free software: you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 3 of the License, or +(at your option) any later version. + +This program is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with this program. If not, see . */ + +#ifndef _FSTRCMP_H +#define _FSTRCMP_H + +#ifdef __cplusplus +extern "C" { +#endif + +/* Fuzzy compare of S1 and S2. Return a measure for the similarity of S1 + and S1. The higher the result, the more similar the strings are. + The result is bounded between 0 (meaning very dissimilar strings) and + 1 (meaning identical strings). */ +extern double fstrcmp (const char *s1, const char *s2); + +/* Like fstrcmp (S1, S2), except that if the result is < LOWER_BOUND, an + arbitrary other value < LOWER_BOUND can be returned. */ +extern double fstrcmp_bounded (const char *s1, const char *s2, + double lower_bound); + +/* A shortcut for fstrcmp. Avoids a function call. */ +#define fstrcmp(s1,s2) fstrcmp_bounded (s1, s2, 0.0) + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/common/misc/glthread/lock.c b/common/misc/glthread/lock.c new file mode 100644 index 0000000..e471188 --- /dev/null +++ b/common/misc/glthread/lock.c @@ -0,0 +1,729 @@ +/* Locking in multithreaded situations. + Copyright (C) 2005-2019 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see . */ + +/* Written by Bruno Haible , 2005. + Based on GCC's gthr-posix.h, gthr-posix95.h. */ + +#include + +#include "glthread/lock.h" + +/* ========================================================================= */ + +#if USE_ISOC_THREADS || USE_ISOC_AND_POSIX_THREADS + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +int +glthread_lock_init (gl_lock_t *lock) +{ + if (mtx_init (&lock->mutex, mtx_plain) != thrd_success) + return ENOMEM; + lock->init_needed = 0; + return 0; +} + +int +glthread_lock_lock (gl_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_lock (&lock->mutex) != thrd_success) + return EAGAIN; + return 0; +} + +int +glthread_lock_unlock (gl_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_unlock (&lock->mutex) != thrd_success) + return EINVAL; + return 0; +} + +int +glthread_lock_destroy (gl_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + mtx_destroy (&lock->mutex); + return 0; +} + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +int +glthread_rwlock_init (gl_rwlock_t *lock) +{ + if (mtx_init (&lock->lock, mtx_plain) != thrd_success + || cnd_init (&lock->waiting_readers) != thrd_success + || cnd_init (&lock->waiting_writers) != thrd_success) + return ENOMEM; + lock->waiting_writers_count = 0; + lock->runcount = 0; + lock->init_needed = 0; + return 0; +} + +int +glthread_rwlock_rdlock (gl_rwlock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_lock (&lock->lock) != thrd_success) + return EAGAIN; + /* Test whether only readers are currently running, and whether the runcount + field will not overflow, and whether no writer is waiting. The latter + condition is because POSIX recommends that "write locks shall take + precedence over read locks", to avoid "writer starvation". */ + while (!(lock->runcount + 1 > 0 && lock->waiting_writers_count == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_readers. */ + if (cnd_wait (&lock->waiting_readers, &lock->lock) != thrd_success) + { + mtx_unlock (&lock->lock); + return EINVAL; + } + } + lock->runcount++; + if (mtx_unlock (&lock->lock) != thrd_success) + return EINVAL; + return 0; +} + +int +glthread_rwlock_wrlock (gl_rwlock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_lock (&lock->lock) != thrd_success) + return EAGAIN; + /* Test whether no readers or writers are currently running. */ + while (!(lock->runcount == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_writers. */ + lock->waiting_writers_count++; + if (cnd_wait (&lock->waiting_writers, &lock->lock) != thrd_success) + { + lock->waiting_writers_count--; + mtx_unlock (&lock->lock); + return EINVAL; + } + lock->waiting_writers_count--; + } + lock->runcount--; /* runcount becomes -1 */ + if (mtx_unlock (&lock->lock) != thrd_success) + return EINVAL; + return 0; +} + +int +glthread_rwlock_unlock (gl_rwlock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_lock (&lock->lock) != thrd_success) + return EAGAIN; + if (lock->runcount < 0) + { + /* Drop a writer lock. */ + if (!(lock->runcount == -1)) + { + mtx_unlock (&lock->lock); + return EINVAL; + } + lock->runcount = 0; + } + else + { + /* Drop a reader lock. */ + if (!(lock->runcount > 0)) + { + mtx_unlock (&lock->lock); + return EINVAL; + } + lock->runcount--; + } + if (lock->runcount == 0) + { + /* POSIX recommends that "write locks shall take precedence over read + locks", to avoid "writer starvation". */ + if (lock->waiting_writers_count > 0) + { + /* Wake up one of the waiting writers. */ + if (cnd_signal (&lock->waiting_writers) != thrd_success) + { + mtx_unlock (&lock->lock); + return EINVAL; + } + } + else + { + /* Wake up all waiting readers. */ + if (cnd_broadcast (&lock->waiting_readers) != thrd_success) + { + mtx_unlock (&lock->lock); + return EINVAL; + } + } + } + if (mtx_unlock (&lock->lock) != thrd_success) + return EINVAL; + return 0; +} + +int +glthread_rwlock_destroy (gl_rwlock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + mtx_destroy (&lock->lock); + cnd_destroy (&lock->waiting_readers); + cnd_destroy (&lock->waiting_writers); + return 0; +} + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +int +glthread_recursive_lock_init (gl_recursive_lock_t *lock) +{ + if (mtx_init (&lock->mutex, mtx_plain | mtx_recursive) != thrd_success) + return ENOMEM; + lock->init_needed = 0; + return 0; +} + +int +glthread_recursive_lock_lock (gl_recursive_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_lock (&lock->mutex) != thrd_success) + return EAGAIN; + return 0; +} + +int +glthread_recursive_lock_unlock (gl_recursive_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + if (mtx_unlock (&lock->mutex) != thrd_success) + return EINVAL; + return 0; +} + +int +glthread_recursive_lock_destroy (gl_recursive_lock_t *lock) +{ + if (lock->init_needed) + call_once (&lock->init_once, lock->init_func); + mtx_destroy (&lock->mutex); + return 0; +} + +/* -------------------------- gl_once_t datatype -------------------------- */ + +#endif + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +# if HAVE_PTHREAD_RWLOCK && (HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER || (defined PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP && (__GNU_LIBRARY__ > 1))) + +# ifdef PTHREAD_RWLOCK_INITIALIZER + +# if !HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER + /* glibc with bug https://sourceware.org/bugzilla/show_bug.cgi?id=13701 */ + +int +glthread_rwlock_init_for_glibc (pthread_rwlock_t *lock) +{ + pthread_rwlockattr_t attributes; + int err; + + err = pthread_rwlockattr_init (&attributes); + if (err != 0) + return err; + /* Note: PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP is the only value that + causes the writer to be preferred. PTHREAD_RWLOCK_PREFER_WRITER_NP does not + do this; see + http://man7.org/linux/man-pages/man3/pthread_rwlockattr_setkind_np.3.html */ + err = pthread_rwlockattr_setkind_np (&attributes, + PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP); + if (err == 0) + err = pthread_rwlock_init(lock, &attributes); + /* pthread_rwlockattr_destroy always returns 0. It cannot influence the + return value. */ + pthread_rwlockattr_destroy (&attributes); + return err; +} + +# endif +# else + +int +glthread_rwlock_init_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_rwlock_init (&lock->rwlock, NULL); + if (err != 0) + return err; + lock->initialized = 1; + return 0; +} + +int +glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_rwlock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_rwlock_rdlock (&lock->rwlock); +} + +int +glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_rwlock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_rwlock_wrlock (&lock->rwlock); +} + +int +glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock) +{ + if (!lock->initialized) + return EINVAL; + return pthread_rwlock_unlock (&lock->rwlock); +} + +int +glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock) +{ + int err; + + if (!lock->initialized) + return EINVAL; + err = pthread_rwlock_destroy (&lock->rwlock); + if (err != 0) + return err; + lock->initialized = 0; + return 0; +} + +# endif + +# else + +int +glthread_rwlock_init_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_init (&lock->lock, NULL); + if (err != 0) + return err; + err = pthread_cond_init (&lock->waiting_readers, NULL); + if (err != 0) + return err; + err = pthread_cond_init (&lock->waiting_writers, NULL); + if (err != 0) + return err; + lock->waiting_writers_count = 0; + lock->runcount = 0; + return 0; +} + +int +glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + /* Test whether only readers are currently running, and whether the runcount + field will not overflow, and whether no writer is waiting. The latter + condition is because POSIX recommends that "write locks shall take + precedence over read locks", to avoid "writer starvation". */ + while (!(lock->runcount + 1 > 0 && lock->waiting_writers_count == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_readers. */ + err = pthread_cond_wait (&lock->waiting_readers, &lock->lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + lock->runcount++; + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + /* Test whether no readers or writers are currently running. */ + while (!(lock->runcount == 0)) + { + /* This thread has to wait for a while. Enqueue it among the + waiting_writers. */ + lock->waiting_writers_count++; + err = pthread_cond_wait (&lock->waiting_writers, &lock->lock); + if (err != 0) + { + lock->waiting_writers_count--; + pthread_mutex_unlock (&lock->lock); + return err; + } + lock->waiting_writers_count--; + } + lock->runcount--; /* runcount becomes -1 */ + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_lock (&lock->lock); + if (err != 0) + return err; + if (lock->runcount < 0) + { + /* Drop a writer lock. */ + if (!(lock->runcount == -1)) + { + pthread_mutex_unlock (&lock->lock); + return EINVAL; + } + lock->runcount = 0; + } + else + { + /* Drop a reader lock. */ + if (!(lock->runcount > 0)) + { + pthread_mutex_unlock (&lock->lock); + return EINVAL; + } + lock->runcount--; + } + if (lock->runcount == 0) + { + /* POSIX recommends that "write locks shall take precedence over read + locks", to avoid "writer starvation". */ + if (lock->waiting_writers_count > 0) + { + /* Wake up one of the waiting writers. */ + err = pthread_cond_signal (&lock->waiting_writers); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + else + { + /* Wake up all waiting readers. */ + err = pthread_cond_broadcast (&lock->waiting_readers); + if (err != 0) + { + pthread_mutex_unlock (&lock->lock); + return err; + } + } + } + return pthread_mutex_unlock (&lock->lock); +} + +int +glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock) +{ + int err; + + err = pthread_mutex_destroy (&lock->lock); + if (err != 0) + return err; + err = pthread_cond_destroy (&lock->waiting_readers); + if (err != 0) + return err; + err = pthread_cond_destroy (&lock->waiting_writers); + if (err != 0) + return err; + return 0; +} + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +# if HAVE_PTHREAD_MUTEX_RECURSIVE + +# if defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER || defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_mutexattr_t attributes; + int err; + + err = pthread_mutexattr_init (&attributes); + if (err != 0) + return err; + err = pthread_mutexattr_settype (&attributes, PTHREAD_MUTEX_RECURSIVE); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutex_init (lock, &attributes); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutexattr_destroy (&attributes); + if (err != 0) + return err; + return 0; +} + +# else + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_mutexattr_t attributes; + int err; + + err = pthread_mutexattr_init (&attributes); + if (err != 0) + return err; + err = pthread_mutexattr_settype (&attributes, PTHREAD_MUTEX_RECURSIVE); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutex_init (&lock->recmutex, &attributes); + if (err != 0) + { + pthread_mutexattr_destroy (&attributes); + return err; + } + err = pthread_mutexattr_destroy (&attributes); + if (err != 0) + return err; + lock->initialized = 1; + return 0; +} + +int +glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock) +{ + if (!lock->initialized) + { + int err; + + err = pthread_mutex_lock (&lock->guard); + if (err != 0) + return err; + if (!lock->initialized) + { + err = glthread_recursive_lock_init_multithreaded (lock); + if (err != 0) + { + pthread_mutex_unlock (&lock->guard); + return err; + } + } + err = pthread_mutex_unlock (&lock->guard); + if (err != 0) + return err; + } + return pthread_mutex_lock (&lock->recmutex); +} + +int +glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock) +{ + if (!lock->initialized) + return EINVAL; + return pthread_mutex_unlock (&lock->recmutex); +} + +int +glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock) +{ + int err; + + if (!lock->initialized) + return EINVAL; + err = pthread_mutex_destroy (&lock->recmutex); + if (err != 0) + return err; + lock->initialized = 0; + return 0; +} + +# endif + +# else + +int +glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock) +{ + int err; + + err = pthread_mutex_init (&lock->mutex, NULL); + if (err != 0) + return err; + lock->owner = (pthread_t) 0; + lock->depth = 0; + return 0; +} + +int +glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock) +{ + pthread_t self = pthread_self (); + if (lock->owner != self) + { + int err; + + err = pthread_mutex_lock (&lock->mutex); + if (err != 0) + return err; + lock->owner = self; + } + if (++(lock->depth) == 0) /* wraparound? */ + { + lock->depth--; + return EAGAIN; + } + return 0; +} + +int +glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != pthread_self ()) + return EPERM; + if (lock->depth == 0) + return EINVAL; + if (--(lock->depth) == 0) + { + lock->owner = (pthread_t) 0; + return pthread_mutex_unlock (&lock->mutex); + } + else + return 0; +} + +int +glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock) +{ + if (lock->owner != (pthread_t) 0) + return EBUSY; + return pthread_mutex_destroy (&lock->mutex); +} + +# endif + +/* -------------------------- gl_once_t datatype -------------------------- */ + +static const pthread_once_t fresh_once = PTHREAD_ONCE_INIT; + +int +glthread_once_singlethreaded (pthread_once_t *once_control) +{ + /* We don't know whether pthread_once_t is an integer type, a floating-point + type, a pointer type, or a structure type. */ + char *firstbyte = (char *)once_control; + if (*firstbyte == *(const char *)&fresh_once) + { + /* First time use of once_control. Invert the first byte. */ + *firstbyte = ~ *(const char *)&fresh_once; + return 1; + } + else + return 0; +} + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +#endif + +/* ========================================================================= */ diff --git a/common/misc/glthread/lock.h b/common/misc/glthread/lock.h new file mode 100644 index 0000000..c190e2e --- /dev/null +++ b/common/misc/glthread/lock.h @@ -0,0 +1,775 @@ +/* Locking in multithreaded situations. + Copyright (C) 2005-2019 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see . */ + +/* Written by Bruno Haible , 2005. + Based on GCC's gthr-posix.h, gthr-posix95.h, gthr-win32.h. */ + +/* This file contains locking primitives for use with a given thread library. + It does not contain primitives for creating threads or for other + synchronization primitives. + + Normal (non-recursive) locks: + Type: gl_lock_t + Declaration: gl_lock_define(extern, name) + Initializer: gl_lock_define_initialized(, name) + Initialization: gl_lock_init (name); + Taking the lock: gl_lock_lock (name); + Releasing the lock: gl_lock_unlock (name); + De-initialization: gl_lock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_lock_init (&name); + Taking the lock: err = glthread_lock_lock (&name); + Releasing the lock: err = glthread_lock_unlock (&name); + De-initialization: err = glthread_lock_destroy (&name); + + Read-Write (non-recursive) locks: + Type: gl_rwlock_t + Declaration: gl_rwlock_define(extern, name) + Initializer: gl_rwlock_define_initialized(, name) + Initialization: gl_rwlock_init (name); + Taking the lock: gl_rwlock_rdlock (name); + gl_rwlock_wrlock (name); + Releasing the lock: gl_rwlock_unlock (name); + De-initialization: gl_rwlock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_rwlock_init (&name); + Taking the lock: err = glthread_rwlock_rdlock (&name); + err = glthread_rwlock_wrlock (&name); + Releasing the lock: err = glthread_rwlock_unlock (&name); + De-initialization: err = glthread_rwlock_destroy (&name); + + Recursive locks: + Type: gl_recursive_lock_t + Declaration: gl_recursive_lock_define(extern, name) + Initializer: gl_recursive_lock_define_initialized(, name) + Initialization: gl_recursive_lock_init (name); + Taking the lock: gl_recursive_lock_lock (name); + Releasing the lock: gl_recursive_lock_unlock (name); + De-initialization: gl_recursive_lock_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_recursive_lock_init (&name); + Taking the lock: err = glthread_recursive_lock_lock (&name); + Releasing the lock: err = glthread_recursive_lock_unlock (&name); + De-initialization: err = glthread_recursive_lock_destroy (&name); + + Once-only execution: + Type: gl_once_t + Initializer: gl_once_define(extern, name) + Execution: gl_once (name, initfunction); + Equivalent functions with control of error handling: + Execution: err = glthread_once (&name, initfunction); +*/ + + +#ifndef _LOCK_H +#define _LOCK_H + +#include +#include + +#if !defined c11_threads_in_use +# if HAVE_THREADS_H && USE_POSIX_THREADS_WEAK +# include +# pragma weak thrd_exit +# define c11_threads_in_use() (thrd_exit != NULL) +# else +# define c11_threads_in_use() 0 +# endif +#endif + +/* ========================================================================= */ + +#if USE_ISOC_THREADS || USE_ISOC_AND_POSIX_THREADS + +/* Use the ISO C threads library. */ + +# include + +# ifdef __cplusplus +extern "C" { +# endif + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef struct + { + int volatile init_needed; + once_flag init_once; + void (*init_func) (void); + mtx_t mutex; + } + gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_lock_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + static void _atomic_init_##NAME (void); \ + STORAGECLASS gl_lock_t NAME = \ + { 1, ONCE_FLAG_INIT, _atomic_init_##NAME }; \ + static void _atomic_init_##NAME (void) \ + { \ + if (glthread_lock_init (&(NAME))) \ + abort (); \ + } +extern int glthread_lock_init (gl_lock_t *lock); +extern int glthread_lock_lock (gl_lock_t *lock); +extern int glthread_lock_unlock (gl_lock_t *lock); +extern int glthread_lock_destroy (gl_lock_t *lock); + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +typedef struct + { + int volatile init_needed; + once_flag init_once; + void (*init_func) (void); + mtx_t lock; /* protects the remaining fields */ + cnd_t waiting_readers; /* waiting readers */ + cnd_t waiting_writers; /* waiting writers */ + unsigned int waiting_writers_count; /* number of waiting writers */ + int runcount; /* number of readers running, or -1 when a writer runs */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + static void _atomic_init_##NAME (void); \ + STORAGECLASS gl_rwlock_t NAME = \ + { 1, ONCE_FLAG_INIT, _atomic_init_##NAME }; \ + static void _atomic_init_##NAME (void) \ + { \ + if (glthread_rwlock_init (&(NAME))) \ + abort (); \ + } +extern int glthread_rwlock_init (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy (gl_rwlock_t *lock); + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +typedef struct + { + int volatile init_needed; + once_flag init_once; + void (*init_func) (void); + mtx_t mutex; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + static void _atomic_init_##NAME (void); \ + STORAGECLASS gl_recursive_lock_t NAME = \ + { 1, ONCE_FLAG_INIT, _atomic_init_##NAME }; \ + static void _atomic_init_##NAME (void) \ + { \ + if (glthread_recursive_lock_init (&(NAME))) \ + abort (); \ + } +extern int glthread_recursive_lock_init (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy (gl_recursive_lock_t *lock); + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef once_flag gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS once_flag NAME = ONCE_FLAG_INIT; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (call_once (ONCE_CONTROL, INITFUNCTION), 0) + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* Use the POSIX threads library. */ + +# include + +# ifdef __cplusplus +extern "C" { +# endif + +# if PTHREAD_IN_USE_DETECTION_HARD + +/* The pthread_in_use() detection needs to be done at runtime. */ +# define pthread_in_use() \ + glthread_in_use () +extern int glthread_in_use (void); + +# endif + +# if USE_POSIX_THREADS_WEAK + +/* Use weak references to the POSIX threads library. */ + +/* Weak references avoid dragging in external libraries if the other parts + of the program don't use them. Here we use them, because we don't want + every program that uses libintl to depend on libpthread. This assumes + that libpthread would not be loaded after libintl; i.e. if libintl is + loaded first, by an executable that does not depend on libpthread, and + then a module is dynamically loaded that depends on libpthread, libintl + will not be multithread-safe. */ + +/* The way to test at runtime whether libpthread is present is to test + whether a function pointer's value, such as &pthread_mutex_init, is + non-NULL. However, some versions of GCC have a bug through which, in + PIC mode, &foo != NULL always evaluates to true if there is a direct + call to foo(...) in the same function. To avoid this, we test the + address of a function in libpthread that we don't use. */ + +# pragma weak pthread_mutex_init +# pragma weak pthread_mutex_lock +# pragma weak pthread_mutex_unlock +# pragma weak pthread_mutex_destroy +# pragma weak pthread_rwlock_init +# pragma weak pthread_rwlock_rdlock +# pragma weak pthread_rwlock_wrlock +# pragma weak pthread_rwlock_unlock +# pragma weak pthread_rwlock_destroy +# pragma weak pthread_once +# pragma weak pthread_cond_init +# pragma weak pthread_cond_wait +# pragma weak pthread_cond_signal +# pragma weak pthread_cond_broadcast +# pragma weak pthread_cond_destroy +# pragma weak pthread_mutexattr_init +# pragma weak pthread_mutexattr_settype +# pragma weak pthread_mutexattr_destroy +# pragma weak pthread_rwlockattr_init +# if __GNU_LIBRARY__ > 1 +# pragma weak pthread_rwlockattr_setkind_np +# endif +# pragma weak pthread_rwlockattr_destroy +# ifndef pthread_self +# pragma weak pthread_self +# endif + +# if !PTHREAD_IN_USE_DETECTION_HARD + /* Considering all platforms with USE_POSIX_THREADS_WEAK, only few symbols + can be used to determine whether libpthread is in use. These are: + pthread_mutexattr_gettype + pthread_rwlockattr_destroy + pthread_rwlockattr_init + */ +# pragma weak pthread_mutexattr_gettype +# define pthread_in_use() \ + (pthread_mutexattr_gettype != NULL || c11_threads_in_use ()) +# endif + +# else + +# if !PTHREAD_IN_USE_DETECTION_HARD +# define pthread_in_use() 1 +# endif + +# endif + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef pthread_mutex_t gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + PTHREAD_MUTEX_INITIALIZER +# define glthread_lock_init(LOCK) \ + (pthread_in_use () ? pthread_mutex_init (LOCK, NULL) : 0) +# define glthread_lock_lock(LOCK) \ + (pthread_in_use () ? pthread_mutex_lock (LOCK) : 0) +# define glthread_lock_unlock(LOCK) \ + (pthread_in_use () ? pthread_mutex_unlock (LOCK) : 0) +# define glthread_lock_destroy(LOCK) \ + (pthread_in_use () ? pthread_mutex_destroy (LOCK) : 0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +# if HAVE_PTHREAD_RWLOCK && (HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER || (defined PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP && (__GNU_LIBRARY__ > 1))) + +# ifdef PTHREAD_RWLOCK_INITIALIZER + +typedef pthread_rwlock_t gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_rwlock_t NAME = gl_rwlock_initializer; +# if HAVE_PTHREAD_RWLOCK_RDLOCK_PREFER_WRITER +# define gl_rwlock_initializer \ + PTHREAD_RWLOCK_INITIALIZER +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? pthread_rwlock_init (LOCK, NULL) : 0) +# else /* glibc with bug https://sourceware.org/bugzilla/show_bug.cgi?id=13701 */ +# define gl_rwlock_initializer \ + PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_for_glibc (LOCK) : 0) +extern int glthread_rwlock_init_for_glibc (pthread_rwlock_t *lock); +# endif +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_rdlock (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_wrlock (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? pthread_rwlock_unlock (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? pthread_rwlock_destroy (LOCK) : 0) + +# else + +typedef struct + { + int initialized; + pthread_mutex_t guard; /* protects the initialization */ + pthread_rwlock_t rwlock; /* read-write lock */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { 0, PTHREAD_MUTEX_INITIALIZER } +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_multithreaded (LOCK) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_rdlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_wrlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_unlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? glthread_rwlock_destroy_multithreaded (LOCK) : 0) +extern int glthread_rwlock_init_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock); + +# endif + +# else + +typedef struct + { + pthread_mutex_t lock; /* protects the remaining fields */ + pthread_cond_t waiting_readers; /* waiting readers */ + pthread_cond_t waiting_writers; /* waiting writers */ + unsigned int waiting_writers_count; /* number of waiting writers */ + int runcount; /* number of readers running, or -1 when a writer runs */ + } + gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, PTHREAD_COND_INITIALIZER, PTHREAD_COND_INITIALIZER, 0, 0 } +# define glthread_rwlock_init(LOCK) \ + (pthread_in_use () ? glthread_rwlock_init_multithreaded (LOCK) : 0) +# define glthread_rwlock_rdlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_rdlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_wrlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_wrlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_unlock(LOCK) \ + (pthread_in_use () ? glthread_rwlock_unlock_multithreaded (LOCK) : 0) +# define glthread_rwlock_destroy(LOCK) \ + (pthread_in_use () ? glthread_rwlock_destroy_multithreaded (LOCK) : 0) +extern int glthread_rwlock_init_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_rdlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_wrlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_unlock_multithreaded (gl_rwlock_t *lock); +extern int glthread_rwlock_destroy_multithreaded (gl_rwlock_t *lock); + +# endif + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +# if HAVE_PTHREAD_MUTEX_RECURSIVE + +# if defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER || defined PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP + +typedef pthread_mutex_t gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS pthread_mutex_t NAME = gl_recursive_lock_initializer; +# ifdef PTHREAD_RECURSIVE_MUTEX_INITIALIZER +# define gl_recursive_lock_initializer \ + PTHREAD_RECURSIVE_MUTEX_INITIALIZER +# else +# define gl_recursive_lock_initializer \ + PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP +# endif +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? pthread_mutex_lock (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? pthread_mutex_unlock (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? pthread_mutex_destroy (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); + +# else + +typedef struct + { + pthread_mutex_t recmutex; /* recursive mutex */ + pthread_mutex_t guard; /* protects the initialization */ + int initialized; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_lock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_unlock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_destroy_multithreaded (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock); + +# endif + +# else + +/* Old versions of POSIX threads on Solaris did not have recursive locks. + We have to implement them ourselves. */ + +typedef struct + { + pthread_mutex_t mutex; + pthread_t owner; + unsigned long depth; + } + gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + { PTHREAD_MUTEX_INITIALIZER, (pthread_t) 0, 0 } +# define glthread_recursive_lock_init(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_init_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_lock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_lock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_unlock(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_unlock_multithreaded (LOCK) : 0) +# define glthread_recursive_lock_destroy(LOCK) \ + (pthread_in_use () ? glthread_recursive_lock_destroy_multithreaded (LOCK) : 0) +extern int glthread_recursive_lock_init_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_lock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_unlock_multithreaded (gl_recursive_lock_t *lock); +extern int glthread_recursive_lock_destroy_multithreaded (gl_recursive_lock_t *lock); + +# endif + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef pthread_once_t gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS pthread_once_t NAME = PTHREAD_ONCE_INIT; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (pthread_in_use () \ + ? pthread_once (ONCE_CONTROL, INITFUNCTION) \ + : (glthread_once_singlethreaded (ONCE_CONTROL) ? (INITFUNCTION (), 0) : 0)) +extern int glthread_once_singlethreaded (pthread_once_t *once_control); + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +# define WIN32_LEAN_AND_MEAN /* avoid including junk */ +# include + +# include "windows-mutex.h" +# include "windows-rwlock.h" +# include "windows-recmutex.h" +# include "windows-once.h" + +# ifdef __cplusplus +extern "C" { +# endif + +/* We can use CRITICAL_SECTION directly, rather than the native Windows Event, + Mutex, Semaphore types, because + - we need only to synchronize inside a single process (address space), + not inter-process locking, + - we don't need to support trylock operations. (TryEnterCriticalSection + does not work on Windows 95/98/ME. Packages that need trylock usually + define their own mutex type.) */ + +/* There is no way to statically initialize a CRITICAL_SECTION. It needs + to be done lazily, once only. For this we need spinlocks. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef glwthread_mutex_t gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_lock_t NAME; +# define gl_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_lock_t NAME = gl_lock_initializer; +# define gl_lock_initializer \ + GLWTHREAD_MUTEX_INIT +# define glthread_lock_init(LOCK) \ + (glwthread_mutex_init (LOCK), 0) +# define glthread_lock_lock(LOCK) \ + glwthread_mutex_lock (LOCK) +# define glthread_lock_unlock(LOCK) \ + glwthread_mutex_unlock (LOCK) +# define glthread_lock_destroy(LOCK) \ + glwthread_mutex_destroy (LOCK) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +typedef glwthread_rwlock_t gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME; +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_rwlock_t NAME = gl_rwlock_initializer; +# define gl_rwlock_initializer \ + GLWTHREAD_RWLOCK_INIT +# define glthread_rwlock_init(LOCK) \ + (glwthread_rwlock_init (LOCK), 0) +# define glthread_rwlock_rdlock(LOCK) \ + glwthread_rwlock_rdlock (LOCK) +# define glthread_rwlock_wrlock(LOCK) \ + glwthread_rwlock_wrlock (LOCK) +# define glthread_rwlock_unlock(LOCK) \ + glwthread_rwlock_unlock (LOCK) +# define glthread_rwlock_destroy(LOCK) \ + glwthread_rwlock_destroy (LOCK) + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +typedef glwthread_recmutex_t gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME; +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) \ + STORAGECLASS gl_recursive_lock_t NAME = gl_recursive_lock_initializer; +# define gl_recursive_lock_initializer \ + GLWTHREAD_RECMUTEX_INIT +# define glthread_recursive_lock_init(LOCK) \ + (glwthread_recmutex_init (LOCK), 0) +# define glthread_recursive_lock_lock(LOCK) \ + glwthread_recmutex_lock (LOCK) +# define glthread_recursive_lock_unlock(LOCK) \ + glwthread_recmutex_unlock (LOCK) +# define glthread_recursive_lock_destroy(LOCK) \ + glwthread_recmutex_destroy (LOCK) + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef glwthread_once_t gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_once_t NAME = GLWTHREAD_ONCE_INIT; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (glwthread_once (ONCE_CONTROL, INITFUNCTION), 0) + +# ifdef __cplusplus +} +# endif + +#endif + +/* ========================================================================= */ + +#if !(USE_ISOC_THREADS || USE_POSIX_THREADS || USE_ISOC_AND_POSIX_THREADS || USE_WINDOWS_THREADS) + +/* Provide dummy implementation if threads are not supported. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +typedef int gl_lock_t; +# define gl_lock_define(STORAGECLASS, NAME) +# define gl_lock_define_initialized(STORAGECLASS, NAME) +# define glthread_lock_init(NAME) 0 +# define glthread_lock_lock(NAME) 0 +# define glthread_lock_unlock(NAME) 0 +# define glthread_lock_destroy(NAME) 0 + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +typedef int gl_rwlock_t; +# define gl_rwlock_define(STORAGECLASS, NAME) +# define gl_rwlock_define_initialized(STORAGECLASS, NAME) +# define glthread_rwlock_init(NAME) 0 +# define glthread_rwlock_rdlock(NAME) 0 +# define glthread_rwlock_wrlock(NAME) 0 +# define glthread_rwlock_unlock(NAME) 0 +# define glthread_rwlock_destroy(NAME) 0 + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +typedef int gl_recursive_lock_t; +# define gl_recursive_lock_define(STORAGECLASS, NAME) +# define gl_recursive_lock_define_initialized(STORAGECLASS, NAME) +# define glthread_recursive_lock_init(NAME) 0 +# define glthread_recursive_lock_lock(NAME) 0 +# define glthread_recursive_lock_unlock(NAME) 0 +# define glthread_recursive_lock_destroy(NAME) 0 + +/* -------------------------- gl_once_t datatype -------------------------- */ + +typedef int gl_once_t; +# define gl_once_define(STORAGECLASS, NAME) \ + STORAGECLASS gl_once_t NAME = 0; +# define glthread_once(ONCE_CONTROL, INITFUNCTION) \ + (*(ONCE_CONTROL) == 0 ? (*(ONCE_CONTROL) = ~ 0, INITFUNCTION (), 0) : 0) + +#endif + +/* ========================================================================= */ + +/* Macros with built-in error handling. */ + +/* -------------------------- gl_lock_t datatype -------------------------- */ + +#define gl_lock_init(NAME) \ + do \ + { \ + if (glthread_lock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_lock(NAME) \ + do \ + { \ + if (glthread_lock_lock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_unlock(NAME) \ + do \ + { \ + if (glthread_lock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_lock_destroy(NAME) \ + do \ + { \ + if (glthread_lock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* ------------------------- gl_rwlock_t datatype ------------------------- */ + +#define gl_rwlock_init(NAME) \ + do \ + { \ + if (glthread_rwlock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_rdlock(NAME) \ + do \ + { \ + if (glthread_rwlock_rdlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_wrlock(NAME) \ + do \ + { \ + if (glthread_rwlock_wrlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_unlock(NAME) \ + do \ + { \ + if (glthread_rwlock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_rwlock_destroy(NAME) \ + do \ + { \ + if (glthread_rwlock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* --------------------- gl_recursive_lock_t datatype --------------------- */ + +#define gl_recursive_lock_init(NAME) \ + do \ + { \ + if (glthread_recursive_lock_init (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_lock(NAME) \ + do \ + { \ + if (glthread_recursive_lock_lock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_unlock(NAME) \ + do \ + { \ + if (glthread_recursive_lock_unlock (&NAME)) \ + abort (); \ + } \ + while (0) +#define gl_recursive_lock_destroy(NAME) \ + do \ + { \ + if (glthread_recursive_lock_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* -------------------------- gl_once_t datatype -------------------------- */ + +#define gl_once(NAME, INITFUNCTION) \ + do \ + { \ + if (glthread_once (&NAME, INITFUNCTION)) \ + abort (); \ + } \ + while (0) + +/* ========================================================================= */ + +#endif /* _LOCK_H */ diff --git a/common/misc/glthread/threadlib.c b/common/misc/glthread/threadlib.c new file mode 100644 index 0000000..86947f5 --- /dev/null +++ b/common/misc/glthread/threadlib.c @@ -0,0 +1,73 @@ +/* Multithreading primitives. + Copyright (C) 2005-2019 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see . */ + +/* Written by Bruno Haible , 2005. */ + +#include + +/* ========================================================================= */ + +#if USE_POSIX_THREADS || USE_ISOC_AND_POSIX_THREADS + +/* Use the POSIX threads library. */ + +# include +# include + +# if PTHREAD_IN_USE_DETECTION_HARD + +/* The function to be executed by a dummy thread. */ +static void * +dummy_thread_func (void *arg) +{ + return arg; +} + +int +glthread_in_use (void) +{ + static int tested; + static int result; /* 1: linked with -lpthread, 0: only with libc */ + + if (!tested) + { + pthread_t thread; + + if (pthread_create (&thread, NULL, dummy_thread_func, NULL) != 0) + /* Thread creation failed. */ + result = 0; + else + { + /* Thread creation works. */ + void *retval; + if (pthread_join (thread, &retval) != 0) + abort (); + result = 1; + } + tested = 1; + } + return result; +} + +# endif + +#endif + +/* ========================================================================= */ + +/* This declaration is solely to ensure that after preprocessing + this file is never empty. */ +typedef int dummy; diff --git a/common/misc/glthread/tls.c b/common/misc/glthread/tls.c new file mode 100644 index 0000000..90a50e9 --- /dev/null +++ b/common/misc/glthread/tls.c @@ -0,0 +1,41 @@ +/* Thread-local storage in multithreaded situations. + Copyright (C) 2005-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +/* Written by Bruno Haible , 2005. */ + +#include + +#include "glthread/tls.h" + +/* ========================================================================= */ + +#if USE_ISOC_THREADS || USE_ISOC_AND_POSIX_THREADS + +#endif + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +#endif + +/* ========================================================================= */ diff --git a/common/misc/glthread/tls.h b/common/misc/glthread/tls.h new file mode 100644 index 0000000..32775fc --- /dev/null +++ b/common/misc/glthread/tls.h @@ -0,0 +1,228 @@ +/* Thread-local storage in multithreaded situations. + Copyright (C) 2005, 2007-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +/* Written by Bruno Haible , 2005. */ + +/* This file contains thread-local storage primitives for use with a given + thread library. It does not contain primitives for creating threads or + for other multithreading primitives. + + Type: gl_tls_key_t + Initialization: gl_tls_key_init (name, destructor); + Getting per-thread value: gl_tls_get (name) + Setting per-thread value: gl_tls_set (name, pointer); + De-initialization: gl_tls_key_destroy (name); + Equivalent functions with control of error handling: + Initialization: err = glthread_tls_key_init (&name, destructor); + Setting per-thread value: err = glthread_tls_set (&name, pointer); + De-initialization: err = glthread_tls_key_destroy (&name); + + A per-thread value is of type 'void *'. + + A destructor is a function pointer of type 'void (*) (void *)', called + when a thread exits, and taking the last per-thread value as argument. It + is unspecified whether the destructor function is called when the last + per-thread value is NULL. On some platforms, the destructor function is + not called at all. +*/ + + +#ifndef _TLS_H +#define _TLS_H + +#include +#include + +#if !defined c11_threads_in_use +# if HAVE_THREADS_H && USE_POSIX_THREADS_WEAK +# include +# pragma weak thrd_exit +# define c11_threads_in_use() (thrd_exit != NULL) +# else +# define c11_threads_in_use() 0 +# endif +#endif + +/* ========================================================================= */ + +#if USE_ISOC_THREADS || USE_ISOC_AND_POSIX_THREADS + +/* Use the ISO C threads library. */ + +# include + +/* ------------------------- gl_tls_key_t datatype ------------------------- */ + +typedef tss_t gl_tls_key_t; +# define glthread_tls_key_init(KEY, DESTRUCTOR) \ + (tss_create (KEY, DESTRUCTOR) != thrd_success ? EAGAIN : 0) +# define gl_tls_get(NAME) \ + tss_get (NAME) +# define glthread_tls_set(KEY, POINTER) \ + (tss_set (*(KEY), (POINTER)) != thrd_success ? ENOMEM : 0) +# define glthread_tls_key_destroy(KEY) \ + (tss_delete (*(KEY)), 0) + +#endif + +/* ========================================================================= */ + +#if USE_POSIX_THREADS + +/* Use the POSIX threads library. */ + +# include + +# if PTHREAD_IN_USE_DETECTION_HARD + +/* The pthread_in_use() detection needs to be done at runtime. */ +# define pthread_in_use() \ + glthread_in_use () +extern int glthread_in_use (void); + +# endif + +# if USE_POSIX_THREADS_WEAK + +/* Use weak references to the POSIX threads library. */ + +# pragma weak pthread_key_create +# pragma weak pthread_getspecific +# pragma weak pthread_setspecific +# pragma weak pthread_key_delete +# ifndef pthread_self +# pragma weak pthread_self +# endif + +# if !PTHREAD_IN_USE_DETECTION_HARD +# pragma weak pthread_mutexattr_gettype +# define pthread_in_use() \ + (pthread_mutexattr_gettype != NULL || c11_threads_in_use ()) +# endif + +# else + +# if !PTHREAD_IN_USE_DETECTION_HARD +# define pthread_in_use() 1 +# endif + +# endif + +/* ------------------------- gl_tls_key_t datatype ------------------------- */ + +typedef union + { + void *singlethread_value; + pthread_key_t key; + } + gl_tls_key_t; +# define glthread_tls_key_init(KEY, DESTRUCTOR) \ + (pthread_in_use () \ + ? pthread_key_create (&(KEY)->key, DESTRUCTOR) \ + : ((KEY)->singlethread_value = NULL, 0)) +# define gl_tls_get(NAME) \ + (pthread_in_use () \ + ? pthread_getspecific ((NAME).key) \ + : (NAME).singlethread_value) +# define glthread_tls_set(KEY, POINTER) \ + (pthread_in_use () \ + ? pthread_setspecific ((KEY)->key, (POINTER)) \ + : ((KEY)->singlethread_value = (POINTER), 0)) +# define glthread_tls_key_destroy(KEY) \ + (pthread_in_use () ? pthread_key_delete ((KEY)->key) : 0) + +#endif + +/* ========================================================================= */ + +#if USE_WINDOWS_THREADS + +# define WIN32_LEAN_AND_MEAN /* avoid including junk */ +# include + +# include "windows-tls.h" + +/* ------------------------- gl_tls_key_t datatype ------------------------- */ + +typedef glwthread_tls_key_t gl_tls_key_t; +# define glthread_tls_key_init(KEY, DESTRUCTOR) \ + glwthread_tls_key_create (KEY, DESTRUCTOR) +# define gl_tls_get(NAME) \ + TlsGetValue (NAME) +# define glthread_tls_set(KEY, POINTER) \ + (!TlsSetValue (*(KEY), POINTER) ? EINVAL : 0) +# define glthread_tls_key_destroy(KEY) \ + glwthread_tls_key_delete (*(KEY)) + +#endif + +/* ========================================================================= */ + +#if !(USE_ISOC_THREADS || USE_POSIX_THREADS || USE_ISOC_AND_POSIX_THREADS || USE_WINDOWS_THREADS) + +/* Provide dummy implementation if threads are not supported. */ + +/* ------------------------- gl_tls_key_t datatype ------------------------- */ + +typedef struct + { + void *singlethread_value; + } + gl_tls_key_t; +# define glthread_tls_key_init(KEY, DESTRUCTOR) \ + ((KEY)->singlethread_value = NULL, \ + (void) (DESTRUCTOR), \ + 0) +# define gl_tls_get(NAME) \ + (NAME).singlethread_value +# define glthread_tls_set(KEY, POINTER) \ + ((KEY)->singlethread_value = (POINTER), 0) +# define glthread_tls_key_destroy(KEY) \ + 0 + +#endif + +/* ========================================================================= */ + +/* Macros with built-in error handling. */ + +/* ------------------------- gl_tls_key_t datatype ------------------------- */ + +#define gl_tls_key_init(NAME, DESTRUCTOR) \ + do \ + { \ + if (glthread_tls_key_init (&NAME, DESTRUCTOR)) \ + abort (); \ + } \ + while (0) +#define gl_tls_set(NAME, POINTER) \ + do \ + { \ + if (glthread_tls_set (&NAME, POINTER)) \ + abort (); \ + } \ + while (0) +#define gl_tls_key_destroy(NAME) \ + do \ + { \ + if (glthread_tls_key_destroy (&NAME)) \ + abort (); \ + } \ + while (0) + +/* ========================================================================= */ + +#endif /* _TLS_H */ diff --git a/common/misc/intprops.h b/common/misc/intprops.h new file mode 100644 index 0000000..0c3963d --- /dev/null +++ b/common/misc/intprops.h @@ -0,0 +1,584 @@ +/* intprops.h -- properties of integer types + + Copyright (C) 2001-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify it + under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +/* Written by Paul Eggert. */ + +#ifndef _GL_INTPROPS_H +#define _GL_INTPROPS_H + +#include + +/* Return a value with the common real type of E and V and the value of V. + Do not evaluate E. */ +#define _GL_INT_CONVERT(e, v) ((1 ? 0 : (e)) + (v)) + +/* Act like _GL_INT_CONVERT (E, -V) but work around a bug in IRIX 6.5 cc; see + . */ +#define _GL_INT_NEGATE_CONVERT(e, v) ((1 ? 0 : (e)) - (v)) + +/* The extra casts in the following macros work around compiler bugs, + e.g., in Cray C 5.0.3.0. */ + +/* True if the arithmetic type T is an integer type. bool counts as + an integer. */ +#define TYPE_IS_INTEGER(t) ((t) 1.5 == 1) + +/* True if the real type T is signed. */ +#define TYPE_SIGNED(t) (! ((t) 0 < (t) -1)) + +/* Return 1 if the real expression E, after promotion, has a + signed or floating type. Do not evaluate E. */ +#define EXPR_SIGNED(e) (_GL_INT_NEGATE_CONVERT (e, 1) < 0) + + +/* Minimum and maximum values for integer types and expressions. */ + +/* The width in bits of the integer type or expression T. + Do not evaluate T. + Padding bits are not supported; this is checked at compile-time below. */ +#define TYPE_WIDTH(t) (sizeof (t) * CHAR_BIT) + +/* The maximum and minimum values for the integer type T. */ +#define TYPE_MINIMUM(t) ((t) ~ TYPE_MAXIMUM (t)) +#define TYPE_MAXIMUM(t) \ + ((t) (! TYPE_SIGNED (t) \ + ? (t) -1 \ + : ((((t) 1 << (TYPE_WIDTH (t) - 2)) - 1) * 2 + 1))) + +/* The maximum and minimum values for the type of the expression E, + after integer promotion. E is not evaluated. */ +#define _GL_INT_MINIMUM(e) \ + (EXPR_SIGNED (e) \ + ? ~ _GL_SIGNED_INT_MAXIMUM (e) \ + : _GL_INT_CONVERT (e, 0)) +#define _GL_INT_MAXIMUM(e) \ + (EXPR_SIGNED (e) \ + ? _GL_SIGNED_INT_MAXIMUM (e) \ + : _GL_INT_NEGATE_CONVERT (e, 1)) +#define _GL_SIGNED_INT_MAXIMUM(e) \ + (((_GL_INT_CONVERT (e, 1) << (TYPE_WIDTH ((e) + 0) - 2)) - 1) * 2 + 1) + +/* Work around OpenVMS incompatibility with C99. */ +#if !defined LLONG_MAX && defined __INT64_MAX +# define LLONG_MAX __INT64_MAX +# define LLONG_MIN __INT64_MIN +#endif + +/* This include file assumes that signed types are two's complement without + padding bits; the above macros have undefined behavior otherwise. + If this is a problem for you, please let us know how to fix it for your host. + This assumption is tested by the intprops-tests module. */ + +/* Does the __typeof__ keyword work? This could be done by + 'configure', but for now it's easier to do it by hand. */ +#if (2 <= __GNUC__ \ + || (1210 <= __IBMC__ && defined __IBM__TYPEOF__) \ + || (0x5110 <= __SUNPRO_C && !__STDC__)) +# define _GL_HAVE___TYPEOF__ 1 +#else +# define _GL_HAVE___TYPEOF__ 0 +#endif + +/* Return 1 if the integer type or expression T might be signed. Return 0 + if it is definitely unsigned. This macro does not evaluate its argument, + and expands to an integer constant expression. */ +#if _GL_HAVE___TYPEOF__ +# define _GL_SIGNED_TYPE_OR_EXPR(t) TYPE_SIGNED (__typeof__ (t)) +#else +# define _GL_SIGNED_TYPE_OR_EXPR(t) 1 +#endif + +/* Bound on length of the string representing an unsigned integer + value representable in B bits. log10 (2.0) < 146/485. The + smallest value of B where this bound is not tight is 2621. */ +#define INT_BITS_STRLEN_BOUND(b) (((b) * 146 + 484) / 485) + +/* Bound on length of the string representing an integer type or expression T. + Subtract 1 for the sign bit if T is signed, and then add 1 more for + a minus sign if needed. + + Because _GL_SIGNED_TYPE_OR_EXPR sometimes returns 1 when its argument is + unsigned, this macro may overestimate the true bound by one byte when + applied to unsigned types of size 2, 4, 16, ... bytes. */ +#define INT_STRLEN_BOUND(t) \ + (INT_BITS_STRLEN_BOUND (TYPE_WIDTH (t) - _GL_SIGNED_TYPE_OR_EXPR (t)) \ + + _GL_SIGNED_TYPE_OR_EXPR (t)) + +/* Bound on buffer size needed to represent an integer type or expression T, + including the terminating null. */ +#define INT_BUFSIZE_BOUND(t) (INT_STRLEN_BOUND (t) + 1) + + +/* Range overflow checks. + + The INT__RANGE_OVERFLOW macros return 1 if the corresponding C + operators might not yield numerically correct answers due to + arithmetic overflow. They do not rely on undefined or + implementation-defined behavior. Their implementations are simple + and straightforward, but they are a bit harder to use than the + INT__OVERFLOW macros described below. + + Example usage: + + long int i = ...; + long int j = ...; + if (INT_MULTIPLY_RANGE_OVERFLOW (i, j, LONG_MIN, LONG_MAX)) + printf ("multiply would overflow"); + else + printf ("product is %ld", i * j); + + Restrictions on *_RANGE_OVERFLOW macros: + + These macros do not check for all possible numerical problems or + undefined or unspecified behavior: they do not check for division + by zero, for bad shift counts, or for shifting negative numbers. + + These macros may evaluate their arguments zero or multiple times, + so the arguments should not have side effects. The arithmetic + arguments (including the MIN and MAX arguments) must be of the same + integer type after the usual arithmetic conversions, and the type + must have minimum value MIN and maximum MAX. Unsigned types should + use a zero MIN of the proper type. + + These macros are tuned for constant MIN and MAX. For commutative + operations such as A + B, they are also tuned for constant B. */ + +/* Return 1 if A + B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_ADD_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? (a) < (min) - (b) \ + : (max) - (b) < (a)) + +/* Return 1 if A - B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_SUBTRACT_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? (max) + (b) < (a) \ + : (a) < (min) + (b)) + +/* Return 1 if - A would overflow in [MIN,MAX] arithmetic. + See above for restrictions. */ +#define INT_NEGATE_RANGE_OVERFLOW(a, min, max) \ + ((min) < 0 \ + ? (a) < - (max) \ + : 0 < (a)) + +/* Return 1 if A * B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Avoid && and || as they tickle + bugs in Sun C 5.11 2010/08/13 and other compilers; see + . */ +#define INT_MULTIPLY_RANGE_OVERFLOW(a, b, min, max) \ + ((b) < 0 \ + ? ((a) < 0 \ + ? (a) < (max) / (b) \ + : (b) == -1 \ + ? 0 \ + : (min) / (b) < (a)) \ + : (b) == 0 \ + ? 0 \ + : ((a) < 0 \ + ? (a) < (min) / (b) \ + : (max) / (b) < (a))) + +/* Return 1 if A / B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Do not check for division by zero. */ +#define INT_DIVIDE_RANGE_OVERFLOW(a, b, min, max) \ + ((min) < 0 && (b) == -1 && (a) < - (max)) + +/* Return 1 if A % B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Do not check for division by zero. + Mathematically, % should never overflow, but on x86-like hosts + INT_MIN % -1 traps, and the C standard permits this, so treat this + as an overflow too. */ +#define INT_REMAINDER_RANGE_OVERFLOW(a, b, min, max) \ + INT_DIVIDE_RANGE_OVERFLOW (a, b, min, max) + +/* Return 1 if A << B would overflow in [MIN,MAX] arithmetic. + See above for restrictions. Here, MIN and MAX are for A only, and B need + not be of the same type as the other arguments. The C standard says that + behavior is undefined for shifts unless 0 <= B < wordwidth, and that when + A is negative then A << B has undefined behavior and A >> B has + implementation-defined behavior, but do not check these other + restrictions. */ +#define INT_LEFT_SHIFT_RANGE_OVERFLOW(a, b, min, max) \ + ((a) < 0 \ + ? (a) < (min) >> (b) \ + : (max) >> (b) < (a)) + +/* True if __builtin_add_overflow (A, B, P) and __builtin_sub_overflow + (A, B, P) work when P is non-null. */ +#if 5 <= __GNUC__ && !defined __ICC +# define _GL_HAS_BUILTIN_ADD_OVERFLOW 1 +#elif defined __has_builtin +# define _GL_HAS_BUILTIN_ADD_OVERFLOW __has_builtin (__builtin_add_overflow) +#else +# define _GL_HAS_BUILTIN_ADD_OVERFLOW 0 +#endif + +/* True if __builtin_mul_overflow (A, B, P) works when P is non-null. */ +#ifdef __clang__ +/* Work around Clang bug . */ +# define _GL_HAS_BUILTIN_MUL_OVERFLOW 0 +#else +# define _GL_HAS_BUILTIN_MUL_OVERFLOW _GL_HAS_BUILTIN_ADD_OVERFLOW +#endif + +/* True if __builtin_add_overflow_p (A, B, C) works, and similarly for + __builtin_mul_overflow_p and __builtin_mul_overflow_p. */ +#define _GL_HAS_BUILTIN_OVERFLOW_P (7 <= __GNUC__) + +/* The _GL*_OVERFLOW macros have the same restrictions as the + *_RANGE_OVERFLOW macros, except that they do not assume that operands + (e.g., A and B) have the same type as MIN and MAX. Instead, they assume + that the result (e.g., A + B) has that type. */ +#if _GL_HAS_BUILTIN_OVERFLOW_P +# define _GL_ADD_OVERFLOW(a, b, min, max) \ + __builtin_add_overflow_p (a, b, (__typeof__ ((a) + (b))) 0) +# define _GL_SUBTRACT_OVERFLOW(a, b, min, max) \ + __builtin_sub_overflow_p (a, b, (__typeof__ ((a) - (b))) 0) +# define _GL_MULTIPLY_OVERFLOW(a, b, min, max) \ + __builtin_mul_overflow_p (a, b, (__typeof__ ((a) * (b))) 0) +#else +# define _GL_ADD_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? INT_ADD_RANGE_OVERFLOW (a, b, min, max) \ + : (a) < 0 ? (b) <= (a) + (b) \ + : (b) < 0 ? (a) <= (a) + (b) \ + : (a) + (b) < (b)) +# define _GL_SUBTRACT_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? INT_SUBTRACT_RANGE_OVERFLOW (a, b, min, max) \ + : (a) < 0 ? 1 \ + : (b) < 0 ? (a) - (b) <= (a) \ + : (a) < (b)) +# define _GL_MULTIPLY_OVERFLOW(a, b, min, max) \ + (((min) == 0 && (((a) < 0 && 0 < (b)) || ((b) < 0 && 0 < (a)))) \ + || INT_MULTIPLY_RANGE_OVERFLOW (a, b, min, max)) +#endif +#define _GL_DIVIDE_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? (b) == _GL_INT_NEGATE_CONVERT (min, 1) && (a) < - (max) \ + : (a) < 0 ? (b) <= (a) + (b) - 1 \ + : (b) < 0 && (a) + (b) <= (a)) +#define _GL_REMAINDER_OVERFLOW(a, b, min, max) \ + ((min) < 0 ? (b) == _GL_INT_NEGATE_CONVERT (min, 1) && (a) < - (max) \ + : (a) < 0 ? (a) % (b) != ((max) - (b) + 1) % (b) \ + : (b) < 0 && ! _GL_UNSIGNED_NEG_MULTIPLE (a, b, max)) + +/* Return a nonzero value if A is a mathematical multiple of B, where + A is unsigned, B is negative, and MAX is the maximum value of A's + type. A's type must be the same as (A % B)'s type. Normally (A % + -B == 0) suffices, but things get tricky if -B would overflow. */ +#define _GL_UNSIGNED_NEG_MULTIPLE(a, b, max) \ + (((b) < -_GL_SIGNED_INT_MAXIMUM (b) \ + ? (_GL_SIGNED_INT_MAXIMUM (b) == (max) \ + ? (a) \ + : (a) % (_GL_INT_CONVERT (a, _GL_SIGNED_INT_MAXIMUM (b)) + 1)) \ + : (a) % - (b)) \ + == 0) + +/* Check for integer overflow, and report low order bits of answer. + + The INT__OVERFLOW macros return 1 if the corresponding C operators + might not yield numerically correct answers due to arithmetic overflow. + The INT__WRAPV macros compute the low-order bits of the sum, + difference, and product of two C integers, and return 1 if these + low-order bits are not numerically correct. + These macros work correctly on all known practical hosts, and do not rely + on undefined behavior due to signed arithmetic overflow. + + Example usage, assuming A and B are long int: + + if (INT_MULTIPLY_OVERFLOW (a, b)) + printf ("result would overflow\n"); + else + printf ("result is %ld (no overflow)\n", a * b); + + Example usage with WRAPV flavor: + + long int result; + bool overflow = INT_MULTIPLY_WRAPV (a, b, &result); + printf ("result is %ld (%s)\n", result, + overflow ? "after overflow" : "no overflow"); + + Restrictions on these macros: + + These macros do not check for all possible numerical problems or + undefined or unspecified behavior: they do not check for division + by zero, for bad shift counts, or for shifting negative numbers. + + These macros may evaluate their arguments zero or multiple times, so the + arguments should not have side effects. + + The WRAPV macros are not constant expressions. They support only + +, binary -, and *. Because the WRAPV macros convert the result, + they report overflow in different circumstances than the OVERFLOW + macros do. + + These macros are tuned for their last input argument being a constant. + + Return 1 if the integer expressions A * B, A - B, -A, A * B, A / B, + A % B, and A << B would overflow, respectively. */ + +#define INT_ADD_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_ADD_OVERFLOW) +#define INT_SUBTRACT_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_SUBTRACT_OVERFLOW) +#if _GL_HAS_BUILTIN_OVERFLOW_P +# define INT_NEGATE_OVERFLOW(a) INT_SUBTRACT_OVERFLOW (0, a) +#else +# define INT_NEGATE_OVERFLOW(a) \ + INT_NEGATE_RANGE_OVERFLOW (a, _GL_INT_MINIMUM (a), _GL_INT_MAXIMUM (a)) +#endif +#define INT_MULTIPLY_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_MULTIPLY_OVERFLOW) +#define INT_DIVIDE_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_DIVIDE_OVERFLOW) +#define INT_REMAINDER_OVERFLOW(a, b) \ + _GL_BINARY_OP_OVERFLOW (a, b, _GL_REMAINDER_OVERFLOW) +#define INT_LEFT_SHIFT_OVERFLOW(a, b) \ + INT_LEFT_SHIFT_RANGE_OVERFLOW (a, b, \ + _GL_INT_MINIMUM (a), _GL_INT_MAXIMUM (a)) + +/* Return 1 if the expression A B would overflow, + where OP_RESULT_OVERFLOW (A, B, MIN, MAX) does the actual test, + assuming MIN and MAX are the minimum and maximum for the result type. + Arguments should be free of side effects. */ +#define _GL_BINARY_OP_OVERFLOW(a, b, op_result_overflow) \ + op_result_overflow (a, b, \ + _GL_INT_MINIMUM (_GL_INT_CONVERT (a, b)), \ + _GL_INT_MAXIMUM (_GL_INT_CONVERT (a, b))) + +/* Store the low-order bits of A + B, A - B, A * B, respectively, into *R. + Return 1 if the result overflows. See above for restrictions. */ +#if _GL_HAS_BUILTIN_ADD_OVERFLOW +# define INT_ADD_WRAPV(a, b, r) __builtin_add_overflow (a, b, r) +# define INT_SUBTRACT_WRAPV(a, b, r) __builtin_sub_overflow (a, b, r) +#else +# define INT_ADD_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, +, _GL_INT_ADD_RANGE_OVERFLOW) +# define INT_SUBTRACT_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, -, _GL_INT_SUBTRACT_RANGE_OVERFLOW) +#endif +#if _GL_HAS_BUILTIN_MUL_OVERFLOW +# if (9 < __GNUC__ + (3 <= __GNUC_MINOR__) \ + || (__GNUC__ == 8 && 4 <= __GNUC_MINOR__)) +# define INT_MULTIPLY_WRAPV(a, b, r) __builtin_mul_overflow (a, b, r) +# else + /* Work around GCC bug 91450. */ +# define INT_MULTIPLY_WRAPV(a, b, r) \ + ((!_GL_SIGNED_TYPE_OR_EXPR (*(r)) && EXPR_SIGNED (a) && EXPR_SIGNED (b) \ + && _GL_INT_MULTIPLY_RANGE_OVERFLOW (a, b, 0, (__typeof__ (*(r))) -1)) \ + ? ((void) __builtin_mul_overflow (a, b, r), 1) \ + : __builtin_mul_overflow (a, b, r)) +# endif +#else +# define INT_MULTIPLY_WRAPV(a, b, r) \ + _GL_INT_OP_WRAPV (a, b, r, *, _GL_INT_MULTIPLY_RANGE_OVERFLOW) +#endif + +/* Nonzero if this compiler has GCC bug 68193 or Clang bug 25390. See: + https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68193 + https://llvm.org/bugs/show_bug.cgi?id=25390 + For now, assume all versions of GCC-like compilers generate bogus + warnings for _Generic. This matters only for compilers that + lack relevant builtins. */ +#if __GNUC__ +# define _GL__GENERIC_BOGUS 1 +#else +# define _GL__GENERIC_BOGUS 0 +#endif + +/* Store the low-order bits of A B into *R, where OP specifies + the operation and OVERFLOW the overflow predicate. Return 1 if the + result overflows. See above for restrictions. */ +#if 201112 <= __STDC_VERSION__ && !_GL__GENERIC_BOGUS +# define _GL_INT_OP_WRAPV(a, b, r, op, overflow) \ + (_Generic \ + (*(r), \ + signed char: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + signed char, SCHAR_MIN, SCHAR_MAX), \ + unsigned char: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + unsigned char, 0, UCHAR_MAX), \ + short int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + short int, SHRT_MIN, SHRT_MAX), \ + unsigned short int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + unsigned short int, 0, USHRT_MAX), \ + int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + int, INT_MIN, INT_MAX), \ + unsigned int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + unsigned int, 0, UINT_MAX), \ + long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX), \ + unsigned long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + unsigned long int, 0, ULONG_MAX), \ + long long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + long long int, LLONG_MIN, LLONG_MAX), \ + unsigned long long int: \ + _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + unsigned long long int, 0, ULLONG_MAX))) +#else +/* Store the low-order bits of A B into *R, where OP specifies + the operation and OVERFLOW the overflow predicate. If *R is + signed, its type is ST with bounds SMIN..SMAX; otherwise its type + is UT with bounds U..UMAX. ST and UT are narrower than int. + Return 1 if the result overflows. See above for restrictions. */ +# if _GL_HAVE___TYPEOF__ +# define _GL_INT_OP_WRAPV_SMALLISH(a,b,r,op,overflow,st,smin,smax,ut,umax) \ + (TYPE_SIGNED (__typeof__ (*(r))) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, st, smin, smax) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, ut, 0, umax)) +# else +# define _GL_INT_OP_WRAPV_SMALLISH(a,b,r,op,overflow,st,smin,smax,ut,umax) \ + (overflow (a, b, smin, smax) \ + ? (overflow (a, b, 0, umax) \ + ? (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a,b,op,unsigned,st), 1) \ + : (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a,b,op,unsigned,st)) < 0) \ + : (overflow (a, b, 0, umax) \ + ? (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a,b,op,unsigned,st)) >= 0 \ + : (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a,b,op,unsigned,st), 0))) +# endif + +# define _GL_INT_OP_WRAPV(a, b, r, op, overflow) \ + (sizeof *(r) == sizeof (signed char) \ + ? _GL_INT_OP_WRAPV_SMALLISH (a, b, r, op, overflow, \ + signed char, SCHAR_MIN, SCHAR_MAX, \ + unsigned char, UCHAR_MAX) \ + : sizeof *(r) == sizeof (short int) \ + ? _GL_INT_OP_WRAPV_SMALLISH (a, b, r, op, overflow, \ + short int, SHRT_MIN, SHRT_MAX, \ + unsigned short int, USHRT_MAX) \ + : sizeof *(r) == sizeof (int) \ + ? (EXPR_SIGNED (*(r)) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + int, INT_MIN, INT_MAX) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned int, \ + unsigned int, 0, UINT_MAX)) \ + : _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow)) +# ifdef LLONG_MAX +# define _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow) \ + (sizeof *(r) == sizeof (long int) \ + ? (EXPR_SIGNED (*(r)) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + unsigned long int, 0, ULONG_MAX)) \ + : (EXPR_SIGNED (*(r)) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + long long int, LLONG_MIN, LLONG_MAX) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long long int, \ + unsigned long long int, 0, ULLONG_MAX))) +# else +# define _GL_INT_OP_WRAPV_LONGISH(a, b, r, op, overflow) \ + (EXPR_SIGNED (*(r)) \ + ? _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + long int, LONG_MIN, LONG_MAX) \ + : _GL_INT_OP_CALC (a, b, r, op, overflow, unsigned long int, \ + unsigned long int, 0, ULONG_MAX)) +# endif +#endif + +/* Store the low-order bits of A B into *R, where the operation + is given by OP. Use the unsigned type UT for calculation to avoid + overflow problems. *R's type is T, with extrema TMIN and TMAX. + T must be a signed integer type. Return 1 if the result overflows. */ +#define _GL_INT_OP_CALC(a, b, r, op, overflow, ut, t, tmin, tmax) \ + (overflow (a, b, tmin, tmax) \ + ? (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a, b, op, ut, t), 1) \ + : (*(r) = _GL_INT_OP_WRAPV_VIA_UNSIGNED (a, b, op, ut, t), 0)) + +/* Return the low-order bits of A B, where the operation is given + by OP. Use the unsigned type UT for calculation to avoid undefined + behavior on signed integer overflow, and convert the result to type T. + UT is at least as wide as T and is no narrower than unsigned int, + T is two's complement, and there is no padding or trap representations. + Assume that converting UT to T yields the low-order bits, as is + done in all known two's-complement C compilers. E.g., see: + https://gcc.gnu.org/onlinedocs/gcc/Integers-implementation.html + + According to the C standard, converting UT to T yields an + implementation-defined result or signal for values outside T's + range. However, code that works around this theoretical problem + runs afoul of a compiler bug in Oracle Studio 12.3 x86. See: + https://lists.gnu.org/r/bug-gnulib/2017-04/msg00049.html + As the compiler bug is real, don't try to work around the + theoretical problem. */ + +#define _GL_INT_OP_WRAPV_VIA_UNSIGNED(a, b, op, ut, t) \ + ((t) ((ut) (a) op (ut) (b))) + +/* Return true if the numeric values A + B, A - B, A * B fall outside + the range TMIN..TMAX. Arguments should be integer expressions + without side effects. TMIN should be signed and nonpositive. + TMAX should be positive, and should be signed unless TMIN is zero. */ +#define _GL_INT_ADD_RANGE_OVERFLOW(a, b, tmin, tmax) \ + ((b) < 0 \ + ? (((tmin) \ + ? ((EXPR_SIGNED (_GL_INT_CONVERT (a, (tmin) - (b))) || (b) < (tmin)) \ + && (a) < (tmin) - (b)) \ + : (a) <= -1 - (b)) \ + || ((EXPR_SIGNED (a) ? 0 <= (a) : (tmax) < (a)) && (tmax) < (a) + (b))) \ + : (a) < 0 \ + ? (((tmin) \ + ? ((EXPR_SIGNED (_GL_INT_CONVERT (b, (tmin) - (a))) || (a) < (tmin)) \ + && (b) < (tmin) - (a)) \ + : (b) <= -1 - (a)) \ + || ((EXPR_SIGNED (_GL_INT_CONVERT (a, b)) || (tmax) < (b)) \ + && (tmax) < (a) + (b))) \ + : (tmax) < (b) || (tmax) - (b) < (a)) +#define _GL_INT_SUBTRACT_RANGE_OVERFLOW(a, b, tmin, tmax) \ + (((a) < 0) == ((b) < 0) \ + ? ((a) < (b) \ + ? !(tmin) || -1 - (tmin) < (b) - (a) - 1 \ + : (tmax) < (a) - (b)) \ + : (a) < 0 \ + ? ((!EXPR_SIGNED (_GL_INT_CONVERT ((a) - (tmin), b)) && (a) - (tmin) < 0) \ + || (a) - (tmin) < (b)) \ + : ((! (EXPR_SIGNED (_GL_INT_CONVERT (tmax, b)) \ + && EXPR_SIGNED (_GL_INT_CONVERT ((tmax) + (b), a))) \ + && (tmax) <= -1 - (b)) \ + || (tmax) + (b) < (a))) +#define _GL_INT_MULTIPLY_RANGE_OVERFLOW(a, b, tmin, tmax) \ + ((b) < 0 \ + ? ((a) < 0 \ + ? (EXPR_SIGNED (_GL_INT_CONVERT (tmax, b)) \ + ? (a) < (tmax) / (b) \ + : ((INT_NEGATE_OVERFLOW (b) \ + ? _GL_INT_CONVERT (b, tmax) >> (TYPE_WIDTH (b) - 1) \ + : (tmax) / -(b)) \ + <= -1 - (a))) \ + : INT_NEGATE_OVERFLOW (_GL_INT_CONVERT (b, tmin)) && (b) == -1 \ + ? (EXPR_SIGNED (a) \ + ? 0 < (a) + (tmin) \ + : 0 < (a) && -1 - (tmin) < (a) - 1) \ + : (tmin) / (b) < (a)) \ + : (b) == 0 \ + ? 0 \ + : ((a) < 0 \ + ? (INT_NEGATE_OVERFLOW (_GL_INT_CONVERT (a, tmin)) && (a) == -1 \ + ? (EXPR_SIGNED (b) ? 0 < (b) + (tmin) : -1 - (tmin) < (b) - 1) \ + : (tmin) / (a) < (b)) \ + : (tmax) / (b) < (a))) + +#endif /* _GL_INTPROPS_H */ diff --git a/common/misc/mbchar.c b/common/misc/mbchar.c new file mode 100644 index 0000000..aa8012b --- /dev/null +++ b/common/misc/mbchar.c @@ -0,0 +1,37 @@ +/* Copyright (C) 2001, 2006, 2009-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + + +#include + +#define MBCHAR_INLINE _GL_EXTERN_INLINE + +#include + +#include "mbchar.h" + +#if IS_BASIC_ASCII + +/* Bit table of characters in the ISO C "basic character set". */ +const unsigned int is_basic_table [UCHAR_MAX / 32 + 1] = +{ + 0x00001a00, /* '\t' '\v' '\f' */ + 0xffffffef, /* ' '...'#' '%'...'?' */ + 0xfffffffe, /* 'A'...'Z' '[' '\\' ']' '^' '_' */ + 0x7ffffffe /* 'a'...'z' '{' '|' '}' '~' */ + /* The remaining bits are 0. */ +}; + +#endif /* IS_BASIC_ASCII */ diff --git a/common/misc/mbchar.h b/common/misc/mbchar.h new file mode 100644 index 0000000..0c35fe3 --- /dev/null +++ b/common/misc/mbchar.h @@ -0,0 +1,364 @@ +/* Multibyte character data type. + Copyright (C) 2001, 2005-2007, 2009-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +/* Written by Bruno Haible . */ + +/* A multibyte character is a short subsequence of a char* string, + representing a single wide character. + + We use multibyte characters instead of wide characters because of + the following goals: + 1) correct multibyte handling, i.e. operate according to the LC_CTYPE + locale, + 2) ease of maintenance, i.e. the maintainer needs not know all details + of the ISO C 99 standard, + 3) don't fail grossly if the input is not in the encoding set by the + locale, because often different encodings are in use in the same + countries (ISO-8859-1/UTF-8, EUC-JP/Shift_JIS, ...), + 4) fast in the case of ASCII characters, + 5) portability, i.e. don't make unportable assumptions about wchar_t. + + Multibyte characters are only accessed through the mb* macros. + + mb_ptr (mbc) + return a pointer to the beginning of the multibyte sequence. + + mb_len (mbc) + returns the number of bytes occupied by the multibyte sequence. + Always > 0. + + mb_iseq (mbc, sc) + returns true if mbc is the standard ASCII character sc. + + mb_isnul (mbc) + returns true if mbc is the nul character. + + mb_cmp (mbc1, mbc2) + returns a positive, zero, or negative value depending on whether mbc1 + sorts after, same or before mbc2. + + mb_casecmp (mbc1, mbc2) + returns a positive, zero, or negative value depending on whether mbc1 + sorts after, same or before mbc2, modulo upper/lowercase conversion. + + mb_equal (mbc1, mbc2) + returns true if mbc1 and mbc2 are equal. + + mb_caseequal (mbc1, mbc2) + returns true if mbc1 and mbc2 are equal modulo upper/lowercase conversion. + + mb_isalnum (mbc) + returns true if mbc is alphanumeric. + + mb_isalpha (mbc) + returns true if mbc is alphabetic. + + mb_isascii(mbc) + returns true if mbc is plain ASCII. + + mb_isblank (mbc) + returns true if mbc is a blank. + + mb_iscntrl (mbc) + returns true if mbc is a control character. + + mb_isdigit (mbc) + returns true if mbc is a decimal digit. + + mb_isgraph (mbc) + returns true if mbc is a graphic character. + + mb_islower (mbc) + returns true if mbc is lowercase. + + mb_isprint (mbc) + returns true if mbc is a printable character. + + mb_ispunct (mbc) + returns true if mbc is a punctuation character. + + mb_isspace (mbc) + returns true if mbc is a space character. + + mb_isupper (mbc) + returns true if mbc is uppercase. + + mb_isxdigit (mbc) + returns true if mbc is a hexadecimal digit. + + mb_width (mbc) + returns the number of columns on the output device occupied by mbc. + Always >= 0. + + mb_putc (mbc, stream) + outputs mbc on stream, a byte oriented FILE stream opened for output. + + mb_setascii (&mbc, sc) + assigns the standard ASCII character sc to mbc. + + mb_copy (&destmbc, &srcmbc) + copies srcmbc to destmbc. + + Here are the function prototypes of the macros. + + extern const char * mb_ptr (const mbchar_t mbc); + extern size_t mb_len (const mbchar_t mbc); + extern bool mb_iseq (const mbchar_t mbc, char sc); + extern bool mb_isnul (const mbchar_t mbc); + extern int mb_cmp (const mbchar_t mbc1, const mbchar_t mbc2); + extern int mb_casecmp (const mbchar_t mbc1, const mbchar_t mbc2); + extern bool mb_equal (const mbchar_t mbc1, const mbchar_t mbc2); + extern bool mb_caseequal (const mbchar_t mbc1, const mbchar_t mbc2); + extern bool mb_isalnum (const mbchar_t mbc); + extern bool mb_isalpha (const mbchar_t mbc); + extern bool mb_isascii (const mbchar_t mbc); + extern bool mb_isblank (const mbchar_t mbc); + extern bool mb_iscntrl (const mbchar_t mbc); + extern bool mb_isdigit (const mbchar_t mbc); + extern bool mb_isgraph (const mbchar_t mbc); + extern bool mb_islower (const mbchar_t mbc); + extern bool mb_isprint (const mbchar_t mbc); + extern bool mb_ispunct (const mbchar_t mbc); + extern bool mb_isspace (const mbchar_t mbc); + extern bool mb_isupper (const mbchar_t mbc); + extern bool mb_isxdigit (const mbchar_t mbc); + extern int mb_width (const mbchar_t mbc); + extern void mb_putc (const mbchar_t mbc, FILE *stream); + extern void mb_setascii (mbchar_t *new, char sc); + extern void mb_copy (mbchar_t *new, const mbchar_t *old); + */ + +#ifndef _MBCHAR_H +#define _MBCHAR_H 1 + +#include +#include + +/* Tru64 with Desktop Toolkit C has a bug: must be included before + . + BSD/OS 4.1 has a bug: and must be included before + . */ +#include +#include +#include +#include + +#include "config.h" + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef MBCHAR_INLINE +# define MBCHAR_INLINE _GL_INLINE +#endif + +#define MBCHAR_BUF_SIZE 24 + +struct mbchar +{ + const char *ptr; /* pointer to current character */ + size_t bytes; /* number of bytes of current character, > 0 */ + bool wc_valid; /* true if wc is a valid wide character */ + wchar_t wc; /* if wc_valid: the current character */ + char buf[MBCHAR_BUF_SIZE]; /* room for the bytes, used for file input only */ +}; + +/* EOF (not a real character) is represented with bytes = 0 and + wc_valid = false. */ + +typedef struct mbchar mbchar_t; + +/* Access the current character. */ +#define mb_ptr(mbc) ((mbc).ptr) +#define mb_len(mbc) ((mbc).bytes) + +/* Comparison of characters. */ +#define mb_iseq(mbc, sc) ((mbc).wc_valid && (mbc).wc == (sc)) +#define mb_isnul(mbc) ((mbc).wc_valid && (mbc).wc == 0) +#define mb_cmp(mbc1, mbc2) \ + ((mbc1).wc_valid \ + ? ((mbc2).wc_valid \ + ? (int) (mbc1).wc - (int) (mbc2).wc \ + : -1) \ + : ((mbc2).wc_valid \ + ? 1 \ + : (mbc1).bytes == (mbc2).bytes \ + ? memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) \ + : (mbc1).bytes < (mbc2).bytes \ + ? (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) > 0 ? 1 : -1) \ + : (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc2).bytes) >= 0 ? 1 : -1))) +#define mb_casecmp(mbc1, mbc2) \ + ((mbc1).wc_valid \ + ? ((mbc2).wc_valid \ + ? (int) towlower ((mbc1).wc) - (int) towlower ((mbc2).wc) \ + : -1) \ + : ((mbc2).wc_valid \ + ? 1 \ + : (mbc1).bytes == (mbc2).bytes \ + ? memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) \ + : (mbc1).bytes < (mbc2).bytes \ + ? (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) > 0 ? 1 : -1) \ + : (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc2).bytes) >= 0 ? 1 : -1))) +#define mb_equal(mbc1, mbc2) \ + ((mbc1).wc_valid && (mbc2).wc_valid \ + ? (mbc1).wc == (mbc2).wc \ + : (mbc1).bytes == (mbc2).bytes \ + && memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) == 0) +#define mb_caseequal(mbc1, mbc2) \ + ((mbc1).wc_valid && (mbc2).wc_valid \ + ? towlower ((mbc1).wc) == towlower ((mbc2).wc) \ + : (mbc1).bytes == (mbc2).bytes \ + && memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) == 0) + +/* , classification. */ +#define mb_isascii(mbc) \ + ((mbc).wc_valid && (mbc).wc >= 0 && (mbc).wc <= 127) +#define mb_isalnum(mbc) ((mbc).wc_valid && iswalnum ((mbc).wc)) +#define mb_isalpha(mbc) ((mbc).wc_valid && iswalpha ((mbc).wc)) +#define mb_isblank(mbc) ((mbc).wc_valid && iswblank ((mbc).wc)) +#define mb_iscntrl(mbc) ((mbc).wc_valid && iswcntrl ((mbc).wc)) +#define mb_isdigit(mbc) ((mbc).wc_valid && iswdigit ((mbc).wc)) +#define mb_isgraph(mbc) ((mbc).wc_valid && iswgraph ((mbc).wc)) +#define mb_islower(mbc) ((mbc).wc_valid && iswlower ((mbc).wc)) +#define mb_isprint(mbc) ((mbc).wc_valid && iswprint ((mbc).wc)) +#define mb_ispunct(mbc) ((mbc).wc_valid && iswpunct ((mbc).wc)) +#define mb_isspace(mbc) ((mbc).wc_valid && iswspace ((mbc).wc)) +#define mb_isupper(mbc) ((mbc).wc_valid && iswupper ((mbc).wc)) +#define mb_isxdigit(mbc) ((mbc).wc_valid && iswxdigit ((mbc).wc)) + +/* Extra function. */ + +/* Unprintable characters appear as a small box of width 1. */ +#define MB_UNPRINTABLE_WIDTH 1 + +int wcwidth(wchar_t wc); + +MBCHAR_INLINE int +mb_width_aux (wint_t wc) +{ + int w = wcwidth (wc); + /* For unprintable characters, arbitrarily return 0 for control characters + and MB_UNPRINTABLE_WIDTH otherwise. */ + return (w >= 0 ? w : iswcntrl (wc) ? 0 : MB_UNPRINTABLE_WIDTH); +} + +#define mb_width(mbc) \ + ((mbc).wc_valid ? mb_width_aux ((mbc).wc) : MB_UNPRINTABLE_WIDTH) + +/* Output. */ +#define mb_putc(mbc, stream) fwrite ((mbc).ptr, 1, (mbc).bytes, (stream)) + +/* Assignment. */ +#define mb_setascii(mbc, sc) \ + ((mbc)->ptr = (mbc)->buf, (mbc)->bytes = 1, (mbc)->wc_valid = 1, \ + (mbc)->wc = (mbc)->buf[0] = (sc)) + +/* Copying a character. */ +MBCHAR_INLINE void +mb_copy (mbchar_t *new_mbc, const mbchar_t *old_mbc) +{ + if (old_mbc->ptr == &old_mbc->buf[0]) + { + memcpy (&new_mbc->buf[0], &old_mbc->buf[0], old_mbc->bytes); + new_mbc->ptr = &new_mbc->buf[0]; + } + else + new_mbc->ptr = old_mbc->ptr; + new_mbc->bytes = old_mbc->bytes; + if ((new_mbc->wc_valid = old_mbc->wc_valid)) + new_mbc->wc = old_mbc->wc; +} + + +/* is_basic(c) tests whether the single-byte character c is in the + ISO C "basic character set". + This is a convenience function, and is in this file only to share code + between mbiter_multi.h and mbfile_multi.h. */ +#if (' ' == 32) && ('!' == 33) && ('"' == 34) && ('#' == 35) \ + && ('%' == 37) && ('&' == 38) && ('\'' == 39) && ('(' == 40) \ + && (')' == 41) && ('*' == 42) && ('+' == 43) && (',' == 44) \ + && ('-' == 45) && ('.' == 46) && ('/' == 47) && ('0' == 48) \ + && ('1' == 49) && ('2' == 50) && ('3' == 51) && ('4' == 52) \ + && ('5' == 53) && ('6' == 54) && ('7' == 55) && ('8' == 56) \ + && ('9' == 57) && (':' == 58) && (';' == 59) && ('<' == 60) \ + && ('=' == 61) && ('>' == 62) && ('?' == 63) && ('A' == 65) \ + && ('B' == 66) && ('C' == 67) && ('D' == 68) && ('E' == 69) \ + && ('F' == 70) && ('G' == 71) && ('H' == 72) && ('I' == 73) \ + && ('J' == 74) && ('K' == 75) && ('L' == 76) && ('M' == 77) \ + && ('N' == 78) && ('O' == 79) && ('P' == 80) && ('Q' == 81) \ + && ('R' == 82) && ('S' == 83) && ('T' == 84) && ('U' == 85) \ + && ('V' == 86) && ('W' == 87) && ('X' == 88) && ('Y' == 89) \ + && ('Z' == 90) && ('[' == 91) && ('\\' == 92) && (']' == 93) \ + && ('^' == 94) && ('_' == 95) && ('a' == 97) && ('b' == 98) \ + && ('c' == 99) && ('d' == 100) && ('e' == 101) && ('f' == 102) \ + && ('g' == 103) && ('h' == 104) && ('i' == 105) && ('j' == 106) \ + && ('k' == 107) && ('l' == 108) && ('m' == 109) && ('n' == 110) \ + && ('o' == 111) && ('p' == 112) && ('q' == 113) && ('r' == 114) \ + && ('s' == 115) && ('t' == 116) && ('u' == 117) && ('v' == 118) \ + && ('w' == 119) && ('x' == 120) && ('y' == 121) && ('z' == 122) \ + && ('{' == 123) && ('|' == 124) && ('}' == 125) && ('~' == 126) +/* The character set is ISO-646, not EBCDIC. */ +# define IS_BASIC_ASCII 1 + +extern const unsigned int is_basic_table[]; + +MBCHAR_INLINE bool +is_basic (char c) +{ + return (is_basic_table [(unsigned char) c >> 5] >> ((unsigned char) c & 31)) + & 1; +} + +#else + +MBCHAR_INLINE bool +is_basic (char c) +{ + switch (c) + { + case '\t': case '\v': case '\f': + case ' ': case '!': case '"': case '#': case '%': + case '&': case '\'': case '(': case ')': case '*': + case '+': case ',': case '-': case '.': case '/': + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + case ':': case ';': case '<': case '=': case '>': + case '?': + case 'A': case 'B': case 'C': case 'D': case 'E': + case 'F': case 'G': case 'H': case 'I': case 'J': + case 'K': case 'L': case 'M': case 'N': case 'O': + case 'P': case 'Q': case 'R': case 'S': case 'T': + case 'U': case 'V': case 'W': case 'X': case 'Y': + case 'Z': + case '[': case '\\': case ']': case '^': case '_': + case 'a': case 'b': case 'c': case 'd': case 'e': + case 'f': case 'g': case 'h': case 'i': case 'j': + case 'k': case 'l': case 'm': case 'n': case 'o': + case 'p': case 'q': case 'r': case 's': case 't': + case 'u': case 'v': case 'w': case 'x': case 'y': + case 'z': case '{': case '|': case '}': case '~': + return 1; + default: + return 0; + } +} + +#endif + +_GL_INLINE_HEADER_END + +#endif /* _MBCHAR_H */ diff --git a/common/misc/mbfile.c b/common/misc/mbfile.c new file mode 100644 index 0000000..b0a468e --- /dev/null +++ b/common/misc/mbfile.c @@ -0,0 +1,3 @@ +#include +#define MBFILE_INLINE _GL_EXTERN_INLINE +#include "mbfile.h" diff --git a/common/misc/mbfile.h b/common/misc/mbfile.h new file mode 100644 index 0000000..211f073 --- /dev/null +++ b/common/misc/mbfile.h @@ -0,0 +1,252 @@ +/* Multibyte character I/O: macros for multi-byte encodings. + Copyright (C) 2001, 2005, 2009-2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +/* Written by Mitsuru Chinen + and Bruno Haible . */ + +/* The macros in this file implement multi-byte character input from a + stream. + + mb_file_t + is the type for multibyte character input stream, usable for variable + declarations. + + mbf_char_t + is the type for multibyte character or EOF, usable for variable + declarations. + + mbf_init (mbf, stream) + initializes the MB_FILE for reading from stream. + + mbf_getc (mbc, mbf) + reads the next multibyte character from mbf and stores it in mbc. + + mb_iseof (mbc) + returns true if mbc represents the EOF value. + + Here are the function prototypes of the macros. + + extern void mbf_init (mb_file_t mbf, FILE *stream); + extern void mbf_getc (mbf_char_t mbc, mb_file_t mbf); + extern bool mb_iseof (const mbf_char_t mbc); + */ + +#ifndef _MBFILE_H +#define _MBFILE_H 1 + +#include +#include +#include +#include + +/* Tru64 with Desktop Toolkit C has a bug: must be included before + . + BSD/OS 4.1 has a bug: and must be included before + . */ +#include +#include +#include + +#include "mbchar.h" + +#ifndef _GL_INLINE_HEADER_BEGIN + #error "Please include config.h first." +#endif +_GL_INLINE_HEADER_BEGIN +#ifndef MBFILE_INLINE +# define MBFILE_INLINE _GL_INLINE +#endif + +struct mbfile_multi { + FILE *fp; + bool eof_seen; + bool have_pushback; + mbstate_t state; + unsigned int bufcount; + char buf[MBCHAR_BUF_SIZE]; + struct mbchar pushback; +}; + +MBFILE_INLINE void +mbfile_multi_getc (struct mbchar *mbc, struct mbfile_multi *mbf) +{ + size_t bytes; + + /* If EOF has already been seen, don't use getc. This matters if + mbf->fp is connected to an interactive tty. */ + if (mbf->eof_seen) + goto eof; + + /* Return character pushed back, if there is one. */ + if (mbf->have_pushback) + { + mb_copy (mbc, &mbf->pushback); + mbf->have_pushback = false; + return; + } + + /* Before using mbrtowc, we need at least one byte. */ + if (mbf->bufcount == 0) + { + int c = getc (mbf->fp); + if (c == EOF) + { + mbf->eof_seen = true; + goto eof; + } + mbf->buf[0] = (unsigned char) c; + mbf->bufcount++; + } + + /* Handle most ASCII characters quickly, without calling mbrtowc(). */ + if (mbf->bufcount == 1 && mbsinit (&mbf->state) && is_basic (mbf->buf[0])) + { + /* These characters are part of the basic character set. ISO C 99 + guarantees that their wide character code is identical to their + char code. */ + mbc->wc = mbc->buf[0] = mbf->buf[0]; + mbc->wc_valid = true; + mbc->ptr = &mbc->buf[0]; + mbc->bytes = 1; + mbf->bufcount = 0; + return; + } + + /* Use mbrtowc on an increasing number of bytes. Read only as many bytes + from mbf->fp as needed. This is needed to give reasonable interactive + behaviour when mbf->fp is connected to an interactive tty. */ + for (;;) + { + /* We don't know whether the 'mbrtowc' function updates the state when + it returns -2, - this is the ISO C 99 and glibc-2.2 behaviour - or + not - amended ANSI C, glibc-2.1 and Solaris 2.7 behaviour. We + don't have an autoconf test for this, yet. + The new behaviour would allow us to feed the bytes one by one into + mbrtowc. But the old behaviour forces us to feed all bytes since + the end of the last character into mbrtowc. Since we want to retry + with more bytes when mbrtowc returns -2, we must backup the state + before calling mbrtowc, because implementations with the new + behaviour will clobber it. */ + mbstate_t backup_state = mbf->state; + + bytes = mbrtowc (&mbc->wc, &mbf->buf[0], mbf->bufcount, &mbf->state); + + if (bytes == (size_t) -1) + { + /* An invalid multibyte sequence was encountered. */ + /* Return a single byte. */ + bytes = 1; + mbc->wc_valid = false; + break; + } + else if (bytes == (size_t) -2) + { + /* An incomplete multibyte character. */ + mbf->state = backup_state; + if (mbf->bufcount == MBCHAR_BUF_SIZE) + { + /* An overlong incomplete multibyte sequence was encountered. */ + /* Return a single byte. */ + bytes = 1; + mbc->wc_valid = false; + break; + } + else + { + /* Read one more byte and retry mbrtowc. */ + int c = getc (mbf->fp); + if (c == EOF) + { + /* An incomplete multibyte character at the end. */ + mbf->eof_seen = true; + bytes = mbf->bufcount; + mbc->wc_valid = false; + break; + } + mbf->buf[mbf->bufcount] = (unsigned char) c; + mbf->bufcount++; + } + } + else + { + if (bytes == 0) + { + /* A null wide character was encountered. */ + bytes = 1; + assert (mbf->buf[0] == '\0'); + assert (mbc->wc == 0); + } + mbc->wc_valid = true; + break; + } + } + + /* Return the multibyte sequence mbf->buf[0..bytes-1]. */ + mbc->ptr = &mbc->buf[0]; + memcpy (&mbc->buf[0], &mbf->buf[0], bytes); + mbc->bytes = bytes; + + mbf->bufcount -= bytes; + if (mbf->bufcount > 0) + { + /* It's not worth calling memmove() for so few bytes. */ + unsigned int count = mbf->bufcount; + char *p = &mbf->buf[0]; + + do + { + *p = *(p + bytes); + p++; + } + while (--count > 0); + } + return; + +eof: + /* An mbchar_t with bytes == 0 is used to indicate EOF. */ + mbc->ptr = NULL; + mbc->bytes = 0; + mbc->wc_valid = false; + return; +} + +MBFILE_INLINE void +mbfile_multi_ungetc (const struct mbchar *mbc, struct mbfile_multi *mbf) +{ + mb_copy (&mbf->pushback, mbc); + mbf->have_pushback = true; +} + +typedef struct mbfile_multi mb_file_t; + +typedef mbchar_t mbf_char_t; + +#define mbf_init(mbf, stream) \ + ((mbf).fp = (stream), \ + (mbf).eof_seen = false, \ + (mbf).have_pushback = false, \ + memset (&(mbf).state, '\0', sizeof (mbstate_t)), \ + (mbf).bufcount = 0) + +#define mbf_getc(mbc, mbf) mbfile_multi_getc (&(mbc), &(mbf)) + +#define mbf_ungetc(mbc, mbf) mbfile_multi_ungetc (&(mbc), &(mbf)) + +#define mb_iseof(mbc) ((mbc).bytes == 0) + +_GL_INLINE_HEADER_END + +#endif /* _MBFILE_H */ diff --git a/common/misc/xhash.c b/common/misc/xhash.c new file mode 100644 index 0000000..9b2bcdb --- /dev/null +++ b/common/misc/xhash.c @@ -0,0 +1,38 @@ +/* hash - hashing table processing. + + Copyright (C) 2019 Free Software Foundation, Inc. + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include + +/* Specification. */ +#include "hash.h" + +#include "xalloc.h" + +/* Same as hash_initialize, but invokes xalloc_die on memory + exhaustion. */ + +Hash_table * +hash_xinitialize (size_t candidate, const Hash_tuning *tuning, + Hash_hasher hasher, Hash_comparator comparator, + Hash_data_freer data_freer) +{ + Hash_table *res = + hash_initialize (candidate, tuning, hasher, comparator, data_freer); + if (!res) + xalloc_die (); + return res; +}