Annotation of dietlibc/CHANGES, revision 1.1
1.1 ! cvs 1: 0.6.15:
! 2: clone support for i386 and alpha by Olaf.
! 3: Initial dynamic loading glue for i386, also by Olaf.
! 4: We don't have a dynamic loader yet, so this is not yet usable.
! 5: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
! 6: vsprintf, writev, all by Olaf.
! 7:
! 8: 0.6.14:
! 9: *scanf did not append the 0 byte for %s.
! 10: added creat and changed creat to call open directly instead of open64.
! 11: did for fread what 0.6.10 did for fwrite.
! 12: repaired buffered stdio a little (line buffering is still not
! 13: supported).
! 14: [All of these bugs have been reported by Guillaume. Thanks!]
! 15: added __pure__ to strchr and strrchr in <string.h>
! 16:
! 17: 0.6.13:
! 18: included COPYING to make clear that the diet libc is covered by the
! 19: GNU General Public License (and _not_ the LGPL). That means that
! 20: you need to obtain a license from me if you want to use the diet
! 21: libc in a proprietary program that you want to distribute.
! 22: The x86 unified syscall swapped arguments #4 and #5 and noone
! 23: noticed until now! Thanks to Guillaume Cottenceau for reporting
! 24: this!
! 25: The malloc had a subtle bug with small allocations that could cause
! 26: segfault. Guillaume reported it, Olaf fixed it.
! 27:
! 28: 0.6.12:
! 29: added assert, statfs and fstatfs
! 30: added include/net/if.h so fget compiles again.
! 31:
! 32: 0.6.11:
! 33: added putenv.
! 34:
! 35: 0.6.10:
! 36: fixed fwrite. It returned the number of bytes written, not the number
! 37: of records. Thanks to Albert D. Cahalan for pointing this out.
! 38: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
! 39: added brk and sbrk. harold@nb.com.sg made me do it.
! 40:
! 41: 0.6.9:
! 42: fixed execvp not to return on ENOENT
! 43: did for {open|read|close|seek|tell}dir what I did for stat before.
! 44: The glibc compatibility part is currently non-functional.
! 45: fixed white space handling in sscanf " (" would not match " (".
! 46: fixed printf to accept 'l' flag.
! 47: fixed getpwuid (ignored last line)
! 48:
! 49: 0.6.8:
! 50: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
! 51: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
! 52: added strlcpy and strlcat from OpenBSD.
! 53: added stat64, fstat64 and lstat64.
! 54: added endian.h and made ending of strlen.c endianness-aware
! 55: added string routines size tweak to dietfeatures.h. Most of the
! 56: string routines have been unrolled. This is up to three times as
! 57: fast but creates up to three times larger code. Now you can disable
! 58: the unrolling.
! 59: added sendfile.
! 60: I also added pread. It works on x86 and sparc but not on ppc and mips.
! 61: I have no idea what's going on. strace disagrees with the kernel.
! 62: Can anyone help? I asked the Linux kernel mailing list for help, too.
! 63: added a few aliases of the type __libc_open for open. nm on
! 64: libpthreads.so indicates that we will need them.
! 65:
! 66: 0.6.7:
! 67: the sources now compile without warnings with -Wall.
! 68: printf now prints "(null)" when %s is passed NULL. This can be
! 69: removed with WANT_NULL_PRINTF in dietfeatures.h
! 70: added vfprintf, execl, ttyname.
! 71: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
! 72: __dietfstat so they can use the normal kernel struct stat.
! 73: stat, lstat and fstat are now C wrappers that convert to the glibc
! 74: struct stat. So we can avoid that overhead for programs that use
! 75: dietlibc headers.
! 76:
! 77: 0.6.6:
! 78: changed the Makefiles so you can now set CFLAGS on the command line
! 79: added memccmp (analogous to memccpy) and strncmp.
! 80: Olaf fixed another bug in the sparc unified syscall.
! 81: Paul Clifford contributed a C version of his strlen.S that is much
! 82: more efficient than the previous strlen on all platforms! It also
! 83: looks like technology from Roswell. ;-}
! 84: added a "real" stdio. Well, almost. Please contribute!
! 85: added dietfeatures.h so you can remove features you don't need.
! 86: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
! 87: stdio now works with simple programs and minigzip from libz.
! 88: stdio uses some major trickery to avoid linking the stdio, stderr and
! 89: stdout (and reserving space for them) when they are not used.
! 90:
! 91: 0.6.5:
! 92: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
! 93: most of them bugs in the new header files, but also missing sigset
! 94: functions.
! 95: Paul Clifford contributed an assembly strlen.S for ARM and fixed
! 96: several bugs.
! 97: Fixed ppc/setjmp and mips/pipe.
! 98: Olaf contributed a new sparc unified syscall.
! 99:
! 100: 0.6.4:
! 101: declared is* static inline in getservent.
! 102: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
! 103: changed strlen to return 0 when passed NULL.
! 104: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
! 105: Scientist who actually implemented this without access to a MIPS
! 106: box, just from reading the architecture manual. And it worked out
! 107: of the box.
! 108: started a set of system includes, mainly so that I can use lcc and
! 109: my alpha-linux cross compiler (which is unable to cross-compile
! 110: glibc). diet libc can now be compiled without any include files
! 111: from a normal libc. The includes do declare more than diet libc
! 112: currently delivers and they are still far from complete for real
! 113: applications.
! 114: I will only add assembly versions that are smaller _and_ faster than
! 115: the C version. Larger routines are only accepted if they are called
! 116: very often and are substantially faster.
! 117: added i386 assembly strchr, which is smaller and faster than the old
! 118: version. It is, however, larger than the version contributed by
! 119: proton (thanks, anyway).
! 120: added i386 strlen (31% faster, 14% smaller)
! 121: "make t" will now create a map file called "mapfile".
! 122:
! 123: 0.6.3:
! 124: added sys_errlist, strerror and perror
! 125: added isblank
! 126: added atol
! 127:
! 128: 0.6.2:
! 129: mmap for ARM didn't compile. Thanks, Paul!
! 130:
! 131: 0.6.1:
! 132: split mmap into the architecture specific subdirectories.
! 133: getenv now copes with environ==NULL, thanks Paul Clifford.
! 134: Paul also contributed a smaller ARM startup code.
! 135:
! 136: 0.6:
! 137: strcat returned the wrong result. Thanks, Dietz Pröpper.
! 138: strtod now understands a negative exponent (oops, thanks Bertram Barth)
! 139: Port to arm-linux-gnu, but on the Netwinder I use for testing the
! 140: __dtostr does not work (I have no idea why!)
! 141: The MIPS port now uses (much smaller) non-PIC code. That means that
! 142: the applications you link against diet libc also have to be compiled
! 143: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
! 144: Thanks to Ralf Bächle for helping me with this!
! 145: I made subdirectories for the architectures and use VPATH to override
! 146: VPATH so that make finds the source file automatically. That should
! 147: simplify the sources greatly.
! 148: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
! 149: ARM and MIPS now also have setjmp and longjmp
! 150: "compile" and "load" are now make targets. Use them for djb programs.
! 151: added contrib/elftrunc.c which will remove unnecessary ELF headers.
! 152: Again, contributed from Olaf. Great work!
! 153:
! 154: 0.5.12:
! 155: printf also does signed numbers.
! 156: If you don't use atexit, dietlibc now does not link exit, only _exit.
! 157: *printf now correctly returns the number of bytes written.
! 158: Olaf contributed experimental sscanf and vsscanf implementations.
! 159: If passed NULL as buffer, snprintf will not write anything but still
! 160: return the number of bytes it would have written.
! 161: Initial MIPS port! (Oh, the agony!)
! 162: No setjmp and longjmp support yet! Please contribute!
! 163: I even implemented unified syscalls for MIPS. Still, MIPS code is
! 164: almost twice the size of SPARC code. If anyone knows why: please
! 165: tell me!
! 166:
! 167: 0.5.11:
! 168: I implemented new unified syscalls for x86, sparc and ppc, this time based
! 169: on .s files and not .c files, and I moved the syscalls into
! 170: subdirectores to clean the dietlibc sources up a little.
! 171: A binary consisting of printf("%s is %d\n","olaf",23) is now
! 172: 2864 bytes on sparc
! 173: 2488 bytes on intel
! 174: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
! 175: to make a few more applications work.
! 176: Olaf Dreesen also implemented some unified syscall stuff for x86.
! 177: He discovered that with our unified syscall interface it actually
! 178: costs <10 bytes total to make all system calls thread safe!
! 179: I implemented atexit() (can register up to 4 callbacks).
! 180: I implemented strtod and __dtostr (the opposite). Now we can add
! 181: floating point support to vsnprintf and sscanf!
! 182:
! 183: 0.5.10:
! 184: I actually saw that I can not only merge the errno handling code of
! 185: the system calls on x86, I can also merge the rest (including the
! 186: arguments) except for setting the system call number. All those
! 187: system calls are now just a jump to a unified system call handler.
! 188: I got rid of x86openclose again. The savings are substantial:
! 189: chown from embutils went from 7664 to 7184 bytes!
! 190: If I'd move the system call wrappers to assembly language, I could
! 191: even reduce the alignment (does not matter since it's just a jump
! 192: anyway) and get rid of the "ret" behind each jump (one byte per
! 193: system call!)
! 194:
! 195: 0.5.9:
! 196: added memchr, strpbrk, strstr, strtol, isspace (hehe)
! 197: fixed strdup
! 198: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
! 199: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
! 200: (no signed integers, only strings and unsigned integers (but
! 201: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
! 202: that code must be wonderful, I don't understand it at all *bg*)
! 203: Olaf's printf does understand padding like in "%08d" and "%8d" and
! 204: automatically pads pointers with '0'.
! 205: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
! 206: Thanks to proton for inspiration on this.
! 207: The new x86 start code also does not reference exit any more, thus saving
! 208: 64 bytes for executables that don't call exit explicitly.
! 209: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
! 210: Moved errno and environ to start.S, saving no byte binary size but
! 211: speeding up compilation and slightly shortening dietlibc.a ;-)
! 212: For x86: joined open and close into one assembler file, sharing the
! 213: errno handling. This is a feasibility test and it actually saves a
! 214: few bytes. I think I will reimplement the _syscall[1-6] macros on
! 215: all platforms now to share the errno handling code for them.
! 216:
! 217: 0.5.8:
! 218: fixed strchr to be able to look for 0.
! 219:
! 220: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
! 221: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
! 222: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
! 223: contributed by Olaf Dreesen)
! 224:
! 225: added execvp, getcwd
! 226:
! 227: fixed __xmknod
! 228:
! 229: Note: can it be that the ftw interface really is so broken that I
! 230: cannot implement it without having to implement some searching data
! 231: structure? I included an experimental ftw implementation that will
! 232: not follow symlinks.
! 233:
! 234: 0.5.7:
! 235: oops, if_nametoindex was broken!
! 236:
! 237: 0.5.6:
! 238: ported to ppc-linux.
! 239: fixed i386 sigsetjmp (I mistyped the function name)
! 240: included sigjmp.c
! 241:
! 242: 0.5.5:
! 243: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
! 244: actually, it wasn't fork that wasn't working for sparc, it was pipe.
! 245: I fixed it now.
! 246: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
! 247: fixed readdir to use getdents and not the intel inline asm
! 248:
! 249: 0.5.4:
! 250: ported to sparclinux (sparc32 only). Beware: does not work yet.
! 251: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
! 252:
! 253: 0.5.3:
! 254: added tcsetattr and getenv to compile e3
! 255:
! 256: 0.5.2:
! 257: added vhangup, tcgetattr, isatty and memcmp for fgetty
! 258: (http://www.fefe.de/fgetty/)
! 259: added localtime from uC-libc.
! 260:
! 261: 0.5.1:
! 262: stat and friends actually work now. Yuck! Another case of
! 263: translation between kernel and userland. Why can't the kernel
! 264: people and the libc people simply agree on a standard?
! 265: removed many unnecessary includes to speed up compilation.
! 266: added -fomit-frame-pointer and i386 compilation target to reduce code size.
! 267: inlined socketcall to reduce code size.
! 268: "load" and "compile" are examples for djb code.
! 269:
! 270: 0.5:
! 271: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
! 272: each object file is now treated with "strip -x -R .note -R .comment"
! 273: split each system call into a separate object file
! 274: added ntohs, htons, alarm, if_indextoname and if_nametoindex
! 275:
! 276: 0.4.1:
! 277: oops, the strchr implementation was wrong. Thanks Jens Laas!
! 278:
! 279: 0.4:
! 280: copied opendir and friends from uC-libc.
! 281: added getservent and getservby* in gerservent.c
! 282:
! 283: 0.3:
! 284: remove readdir system call
! 285: added getdents system call
! 286: added getpwnam/getpwuid implementation in getpwnam.c
! 287: added a few string functions (in str*.c)
! 288:
LinuxTV legacy CVS <linuxtv.org/cvs>