Trying to get RV64 glibc binaries able to work.

README.md

@@ -23,7 +23,7 @@ use cases. It is not intended to be a general purpose processor arch or for the
 more acceptable for there to be variations or breaks in binary compatibility between implementations (granted, it is
 preferable if breaks in compatibility can be avoided if possible).
 
-My personal use-case is currently mostly in the context of using BSR1 for real-time motor control (Update, 2021-07: BJX2 has mostly taken over this role, as single-wide cores can be fit onto an XC7S25).
+My personal use-case is currently mostly in the context of using BSR1 for real-time motor control (Update, 2021-07: BJX2 has mostly taken over this role, as single-wide cores can be fit onto an XC7S25). ( Update, 2024-09: I have mostly dropped these use-cases in favor of using RISC-V instead. I am currently focusing on BJX2 for more computationally oriented tasks, rather than microcontroller tasks. )
 
 As of 2018-05-20, This is still at a fairly early stage of development and is not yet ready for general use, however some
 initial results look promising, and I am making better progress than I was with BJX1 (if anything, because BSR1 is a bit simpler
@@ -58,6 +58,10 @@ Update 2021-07: The core ISA of BJX2 has mostly stabilized, and significant desi
 bgbcc22 (BGBCC): C compiler, partly reused from my BJX1 project, but modified to add support for BSR1 and BJX2.
 * TODO: Make it more obvious how to use said compiler.
 
+I will add an optional exception to the MIT license for BGBCC:
+* One may, at their discretion, choose to disregard the middle clause of the MIT license, allowing derived copies to be understood as-if they were CC0 or Public Domain.
+
+
 vmbase: Holds an emulator for the BSR1 ISA.
 
 jx2vm: Emulator for the BJX2 ISA.

bgbcc22/Makefile.inc

@@ -28,6 +28,9 @@ SRCS=\
 	mm/qoilz_enc.c \
 	mm/lcif_dec.c \
 	mm/lcif_enc.c \
+	mm/tkupic_dec.c \
+	mm/tkupic_enc.c \
+	mm/bufpng.c \
 	\
 	ccxl/ccxl_cgif.c \
 	ccxl/ccxl_compile.c \

bgbcc22/bgbcc_multi.c

@@ -101,6 +101,10 @@
 #include "mm/lcif_dec.c"
 #include "mm/lcif_enc.c"
 
+#include "mm/tkupic_dec.c"
+#include "mm/tkupic_enc.c"
+#include "mm/bufpng.c"
+
 
 #include "ccxl/ccxl_cgif.c"
 #include "ccxl/ccxl_compile.c"

bgbcc22/bgbmeta.c

@@ -1598,6 +1598,24 @@ byte *BGBCC_LoadConvResource(byte *buf, int sz, fourcc lang,
 		*rsz=sz1;
 		return(obuf);
 	}
+
+	if(	!bgbcc_stricmp(cnv, "upic") )
+	{
+		ibuf=BGBCC_Img_DecodeImage(buf, &xs, &ys);
+		if(!ibuf)
+			return(NULL);
+
+		fl=1<<8;
+
+		fl|=100-((qlvl&7)*12);
+
+		obuf=malloc(xs*ys*2);
+		sz1=TKuPI_EncodeImageBufferTemp(obuf, ibuf, xs, ys, fl);
+
+		*rfcc=BGBCC_FMT_UPIC;
+		*rsz=sz1;
+		return(obuf);
+	}
 
 	return(NULL);
 }

bgbcc22/ccxl/ccxl_stack.c

@@ -4465,7 +4465,8 @@ ccxl_status BGBCC_CCXL_StackLoadSlotAddr(BGBCC_TransState *ctx, char *name)
 	BGBCC_CCXL_Error(ctx, "Undefined Member %s -> %s\n",
 		st->decl->qname, name);
 
-	BGBCC_CCXL_PushRegister(ctx, dreg);
+//	BGBCC_CCXL_PushRegister(ctx, dreg);
+	BGBCC_CCXL_StackPushConstInt(ctx, 0);
 
 	BGBCC_CCXL_TagError(ctx,
 		CCXL_TERR_STATUS(CCXL_STATUS_ERR_BADOPARGS));

bgbcc22/docs/2020-06-14_WAD2A.txt

@@ -3,11 +3,15 @@ WAD2(A) is a container format holding a series of 'lumps'.
 * It is closely related to the Quake WAD2 format.
 
  Header {
- 	FOURCC	magic;		//'WAD2'
- 	U32		numlumps;	//Number of lumps in the directory.
- 	U32		diroffs;	//Offset of directory, in bytes
- 	U32		typeoffs;	//Offset of types table
- }
+ 	FOURCC	magic;		//00: 'WAD2'
+ 	U32		numlumps;	//04: Number of lumps in the directory.
+ 	U32		diroffs;	//08: Offset of directory, in bytes
+ 	U32		typeoffs;	//0C: Offset of types table
+  	U32		resv1;		//10: Reserved
+ 	U32		resv2;		//14: Reserved
+ 	FOURCC	intent1;	//18: Intention 1
+ 	FOURCC	intent2;	//1C: Intention 2
+}
 
 The header appears at the start of a file, and encodes the location of the directory.
 

bgbcc22/docs/2020-10-15_WAD4.txt

@@ -23,8 +23,8 @@ A WAD4 Image will begin with a WAD4 File Header:
  	U32		hashoffs;	//0C: Offset of Directory Hash.
  	U32		bmpoffs;	//10: Offset of Allocation Bitmap.
  	U32		bmpsize;	//14: Bitmap Size
- 	U32		resv4;		//18: Reserved
- 	U32		resv5;		//1C: Reserved
+ 	FOURCC	intent1;	//18: Intention 1
+ 	FOURCC	intent2;	//1C: Intention 2
  	//...
  }
 
@@ -42,6 +42,10 @@ Header Fields:
 * bmpsize:
 ** Gives the size of the bitmap in bytes.
 ** The number of cells in the image may be determined by the bitmap size.
+* intent1/intent2:
+** Gives an 'intention' for the file.
+** These may be used in the creation of derived formats.
+** If zero, image is seen as a collection of files.
 
 
 === Directory ===

bgbcc22/docs/2023-11-15_WDEF0.txt

@@ -42,6 +42,7 @@ Where, fmt may be one of:
 * qoi: QOI Format
 * qoli: QOLI (QOI with inline LZ compression)
 * lcif: LCIF Format
+* upic: UPIC Format
 
 BMPA is a modified BMP differing mostly in that the fields and image data will be aligned. The headers will have a native 32-bit alignment, whereas the image data will have a 64-bit alignment. The magic will be changed from 'BM' to ' BMP', but apart from the alignment change the format is basically the same.
 
@@ -56,7 +57,7 @@ For BMP16 and BMP16A, the MSB of each pixel will encode Alpha:
 
 For QOI and QOLI, a 'q' parameter will speficy a lossy quality level (0=lossless, 1=best, 7=worst). The default will be lossless.
 
-For LCIF, the 'q' parameter will speficy a lossy quality level (0=best, 7=worst). The default will be best.
+For LCIF and UPIC, the 'q' parameter will speficy a lossy quality level (0=best, 7=worst). The default will be best.
 
 Both QOI and LCIF will include an full Alpha channel.