Update to C API 0.99.14

subprojects/tskit/tskit/convert.c

@@ -42,7 +42,7 @@
  * pointless. */
 
 typedef struct {
-    size_t precision;
+    unsigned int precision;
     tsk_flags_t options;
     char *newick;
     tsk_id_t *traversal_stack;
@@ -98,7 +98,7 @@ tsk_newick_converter_run(
                 }
                 /* We do this for ms-compatability. This should be a configurable option
                  * via the flags attribute */
-                label = v + 1;
+                label = (int) v + 1;
                 r = snprintf(buffer + s, buffer_size - s, "%d", label);
                 if (r < 0) {
                     ret = TSK_ERR_IO;
@@ -145,16 +145,17 @@ tsk_newick_converter_run(
 
 static int
 tsk_newick_converter_init(tsk_newick_converter_t *self, const tsk_tree_t *tree,
-    size_t precision, tsk_flags_t options)
+    unsigned int precision, tsk_flags_t options)
 {
     int ret = 0;
 
-    memset(self, 0, sizeof(tsk_newick_converter_t));
+    tsk_memset(self, 0, sizeof(tsk_newick_converter_t));
     self->precision = precision;
     self->options = options;
     self->tree = tree;
-    self->traversal_stack = malloc(tsk_treeseq_get_num_nodes(self->tree->tree_sequence)
-                                   * sizeof(*self->traversal_stack));
+    self->traversal_stack
+        = tsk_malloc(tsk_treeseq_get_num_nodes(self->tree->tree_sequence)
+                     * sizeof(*self->traversal_stack));
     if (self->traversal_stack == NULL) {
         ret = TSK_ERR_NO_MEMORY;
         goto out;
@@ -171,7 +172,7 @@ tsk_newick_converter_free(tsk_newick_converter_t *self)
 }
 
 int
-tsk_convert_newick(const tsk_tree_t *tree, tsk_id_t root, size_t precision,
+tsk_convert_newick(const tsk_tree_t *tree, tsk_id_t root, unsigned int precision,
     tsk_flags_t options, size_t buffer_size, char *buffer)
 {
     int ret = 0;

subprojects/tskit/tskit/convert.h

@@ -32,7 +32,7 @@ extern "C" {
 
 #include <tskit/trees.h>
 
-int tsk_convert_newick(const tsk_tree_t *tree, tsk_id_t root, size_t precision,
+int tsk_convert_newick(const tsk_tree_t *tree, tsk_id_t root, unsigned int precision,
     tsk_flags_t options, size_t buffer_size, char *buffer);
 
 #ifdef __cplusplus

subprojects/tskit/tskit/core.c

@@ -1,7 +1,7 @@
 /*
  * MIT License
  *
- * Copyright (c) 2019 Tskit Developers
+ * Copyright (c) 2019-2021 Tskit Developers
  * Copyright (c) 2015-2018 University of Oxford
  *
  * Permission is hereby granted, free of charge, to any person obtaining a copy
@@ -175,6 +175,9 @@ tsk_strerror_internal(int err)
         case TSK_ERR_BOTH_COLUMNS_REQUIRED:
             ret = "Both columns in a related pair must be provided";
             break;
+        case TSK_ERR_BAD_COLUMN_TYPE:
+            ret = "An incompatible type for a column was found in the file";
+            break;
 
         /* Out of bounds errors */
         case TSK_ERR_BAD_OFFSET:
@@ -400,6 +403,9 @@ tsk_strerror_internal(int err)
         case TSK_ERR_BAD_GENOTYPE:
             ret = "Bad genotype value provided";
             break;
+        case TSK_ERR_BAD_ANCESTRAL_STATE:
+            ret = "Bad ancestral state specified";
+            break;
 
         /* Genotype decoding errors */
         case TSK_ERR_TOO_MANY_ALLELES:
@@ -541,11 +547,11 @@ void
 tsk_blkalloc_print_state(tsk_blkalloc_t *self, FILE *out)
 {
     fprintf(out, "Block allocator%p::\n", (void *) self);
-    fprintf(out, "\ttop = %d\n", (int) self->top);
-    fprintf(out, "\tchunk_size = %d\n", (int) self->chunk_size);
-    fprintf(out, "\tnum_chunks = %d\n", (int) self->num_chunks);
-    fprintf(out, "\ttotal_allocated = %d\n", (int) self->total_allocated);
-    fprintf(out, "\ttotal_size = %d\n", (int) self->total_size);
+    fprintf(out, "\ttop = %lld\n", (long long) self->top);
+    fprintf(out, "\tchunk_size = %lld\n", (long long) self->chunk_size);
+    fprintf(out, "\tnum_chunks = %lld\n", (long long) self->num_chunks);
+    fprintf(out, "\ttotal_allocated = %lld\n", (long long) self->total_allocated);
+    fprintf(out, "\ttotal_size = %lld\n", (long long) self->total_size);
 }
 
 int TSK_WARN_UNUSED
@@ -564,7 +570,7 @@ tsk_blkalloc_init(tsk_blkalloc_t *self, size_t chunk_size)
 {
     int ret = 0;
 
-    memset(self, 0, sizeof(tsk_blkalloc_t));
+    tsk_memset(self, 0, sizeof(tsk_blkalloc_t));
     if (chunk_size < 1) {
         ret = TSK_ERR_BAD_PARAM_VALUE;
         goto out;
@@ -642,8 +648,8 @@ tsk_blkalloc_free(tsk_blkalloc_t *self)
 
 /* Mirrors the semantics of numpy's searchsorted function. Uses binary
  * search to find the index of the closest value in the array. */
-size_t
-tsk_search_sorted(const double *restrict array, size_t size, double value)
+tsk_size_t
+tsk_search_sorted(const double *restrict array, tsk_size_t size, double value)
 {
     int64_t upper = (int64_t) size;
     int64_t lower = 0;
@@ -663,7 +669,7 @@ tsk_search_sorted(const double *restrict array, size_t size, double value)
         }
     }
     offset = (int64_t)(array[lower] < value);
-    return (size_t)(lower + offset);
+    return (tsk_size_t)(lower + offset);
 }
 
 /* Rounds the specified double to the closest multiple of 10**-num_digits. If
@@ -703,3 +709,70 @@ tsk_is_unknown_time(double val)
     nan_union.f = val;
     return nan_union.i == TSK_UNKNOWN_TIME_HEX;
 }
+
+void *
+tsk_malloc(tsk_size_t size)
+{
+    /* Avoid malloc(0) as it's not portable */
+    if (size == 0) {
+        size = 1;
+    }
+#if TSK_MAX_SIZE > SIZE_MAX
+    if (size > SIZE_MAX) {
+        return NULL;
+    }
+#endif
+    return malloc((size_t) size);
+}
+
+void *
+tsk_realloc(void *ptr, tsk_size_t size)
+{
+    /* We shouldn't ever realloc to a zero size in tskit */
+    tsk_bug_assert(size > 0);
+    return realloc(ptr, (size_t) size);
+}
+
+/* We keep the size argument here as a size_t because we'd have to
+ * cast the outputs of sizeof() otherwise, which would lead to
+ * less readable code. We need to be careful to use calloc within
+ * the library accordingly, so that size can't overflow on 32 bit.
+ */
+void *
+tsk_calloc(tsk_size_t n, size_t size)
+{
+    /* Avoid calloc(0) as it's not portable */
+    if (n == 0) {
+        n = 1;
+    }
+#if TSK_MAX_SIZE > SIZE_MAX
+    if (n > SIZE_MAX) {
+        return NULL;
+    }
+#endif
+    return calloc((size_t) n, size);
+}
+
+void *
+tsk_memset(void *ptr, int fill, tsk_size_t size)
+{
+    return memset(ptr, fill, (size_t) size);
+}
+
+void *
+tsk_memcpy(void *dest, const void *src, tsk_size_t size)
+{
+    return memcpy(dest, src, (size_t) size);
+}
+
+void *
+tsk_memmove(void *dest, const void *src, tsk_size_t size)
+{
+    return memmove(dest, src, (size_t) size);
+}
+
+int
+tsk_memcmp(const void *s1, const void *s2, tsk_size_t size)
+{
+    return memcmp(s1, s2, (size_t) size);
+}

subprojects/tskit/tskit/core.h

@@ -109,6 +109,59 @@ __tsk_nan_f(void)
 }
 #define TSK_UNKNOWN_TIME __tsk_nan_f()
 
+/**
+@brief Tskit Object IDs.
+
+@rst
+All objects in tskit are referred to by integer IDs corresponding to the
+row they occupy in the relevant table. The ``tsk_id_t`` type should be used
+when manipulating these ID values. The reserved value ``TSK_NULL`` (-1) defines
+missing data.
+@endrst
+*/
+#ifdef _TSK_BIG_TABLES
+/* Allow tables to have more than 2^31 rows. This is an EXPERIMENTAL feature
+ * and is not supported in any way. This typedef is only included for
+ * future-proofing purposes, so that we can be sure that we don't make any
+ * design decisions that are incompatible with big tables by building the
+ * library in 64 bit mode in CI. See the discussion here for more background:
+
+ * https://github.com/tskit-dev/tskit/issues/343
+ *
+ * If you need big tables, please open an issue on GitHub to discuss, or comment
+ * on the thread above.
+ */
+typedef int64_t tsk_id_t;
+#define TSK_MAX_ID INT64_MAX
+#define TSK_ID_STORAGE_TYPE KAS_INT64
+#else
+typedef int32_t tsk_id_t;
+#define TSK_MAX_ID INT32_MAX
+#define TSK_ID_STORAGE_TYPE KAS_INT32
+#endif
+
+/**
+@brief Tskit sizes.
+
+@rst
+The ``tsk_size_t`` type is an unsigned integer used for any size or count value.
+@endrst
+*/
+typedef uint64_t tsk_size_t;
+#define TSK_MAX_SIZE UINT64_MAX
+#define TSK_SIZE_STORAGE_TYPE KAS_UINT64
+
+/**
+@brief Container for bitwise flags.
+
+@rst
+Bitwise flags are used in tskit as a column type and also as a way to
+specify options to API functions.
+@endrst
+*/
+typedef uint32_t tsk_flags_t;
+#define TSK_FLAGS_STORAGE_TYPE KAS_UINT32
+
 // clang-format off
 /**
 @defgroup API_VERSION_GROUP API version macros.
@@ -129,7 +182,7 @@ to the API or ABI are introduced, i.e., the addition of a new function.
 The library patch version. Incremented when any changes not relevant to the
 to the API or ABI are introduced, i.e., internal refactors of bugfixes.
 */
-#define TSK_VERSION_PATCH   12
+#define TSK_VERSION_PATCH   14
 /** @} */
 
 /* Node flags */
@@ -144,7 +197,7 @@ to the API or ABI are introduced, i.e., internal refactors of bugfixes.
 #define TSK_FILE_FORMAT_NAME          "tskit.trees"
 #define TSK_FILE_FORMAT_NAME_LENGTH   11
 #define TSK_FILE_FORMAT_VERSION_MAJOR 12
-#define TSK_FILE_FORMAT_VERSION_MINOR 4
+#define TSK_FILE_FORMAT_VERSION_MINOR 5
 
 /**
 @defgroup GENERAL_ERROR_GROUP General errors.
@@ -208,6 +261,12 @@ not found in the file.
 */
 #define TSK_ERR_BOTH_COLUMNS_REQUIRED                               -104
 
+/**
+An unsupported type was provided for a column in the file.
+*/
+#define TSK_ERR_BAD_COLUMN_TYPE                                     -105
+
+
 /* Out of bounds errors */
 #define TSK_ERR_BAD_OFFSET                                          -200
 #define TSK_ERR_OUT_OF_BOUNDS                                       -201
@@ -291,6 +350,7 @@ not found in the file.
 /* Mutation mapping errors */
 #define TSK_ERR_GENOTYPES_ALL_MISSING                              -1000
 #define TSK_ERR_BAD_GENOTYPE                                       -1001
+#define TSK_ERR_BAD_ANCESTRAL_STATE                                -1002
 
 /* Genotype decoding errors */
 #define TSK_ERR_MUST_IMPUTE_NON_SAMPLES                            -1100
@@ -402,14 +462,24 @@ extern int tsk_blkalloc_init(tsk_blkalloc_t *self, size_t chunk_size);
 extern void *tsk_blkalloc_get(tsk_blkalloc_t *self, size_t size);
 extern void tsk_blkalloc_free(tsk_blkalloc_t *self);
 
-size_t tsk_search_sorted(const double *array, size_t size, double value);
+tsk_size_t tsk_search_sorted(const double *array, tsk_size_t size, double value);
 double tsk_round(double x, unsigned int ndigits);
 
 bool tsk_is_unknown_time(double val);
 
 #define TSK_UUID_SIZE 36
 int tsk_generate_uuid(char *dest, int flags);
 
+/* TODO most of these can probably be macros so they compile out as no-ops.
+ * Lets do the 64 bit tsk_size_t switch first though. */
+void *tsk_malloc(tsk_size_t size);
+void *tsk_realloc(void *ptr, tsk_size_t size);
+void *tsk_calloc(tsk_size_t n, size_t size);
+void *tsk_memset(void *ptr, int fill, tsk_size_t size);
+void *tsk_memcpy(void *dest, const void *src, tsk_size_t size);
+void *tsk_memmove(void *dest, const void *src, tsk_size_t size);
+int tsk_memcmp(const void *s1, const void *s2, tsk_size_t size);
+
 #ifdef __cplusplus
 }
 #endif