Annotation of dietlibc/CHANGES, revision 1.73
1.73 ! fefe 1: 0.9:
1.65 fefe 2: fixed initgroups (oops)
1.66 fefe 3: Cleanup appending file open for stdio (thanks, James Antill)
4: imported vfprintf fix from James Antill.
1.67 fefe 5: add fnmatch
6: change index, rindex and bcmp as weak ELF symbols (were #defined before)
1.68 fefe 7: Olaf added character classes to *scanf.
1.69 fefe 8: Put object and library files into a separate directory (bin-$ARCH).
9: Add "cross" make target to make all architectures.
10: Add "diet" program and man page.
1.70 fefe 11: Add getgrnam, getgrgid
1.71 fefe 12: Add cfmakeraw (Thanks, Antonio Musumeci)
1.72 fefe 13: localtime won't call gettimeofday when the tzfile parser is used.
1.65 fefe 14:
1.64 fefe 15: 0.8:
1.34 fefe 16: vsnprintf(0,...) estimated the length of strings with a padding width
17: incorrectly.
1.35 fefe 18: new reboot() from Olaf.
1.36 fefe 19: added prototypes for getsockopt and setsockopt.
1.39 fefe 20: added experimental sem* and shm*
1.38 fefe 21: define EXIT_SUCCESS and EXIT_FAILURE
1.40 fefe 22: added setvbuf (only changes mode, not buffer)
1.41 fefe 23: added dietwarning.h (woohoo, linker warnings!) and used it to warn of
24: insecurity (system, sprintf) or bloat (printf and stdio).
1.42 fefe 25: added qsort() and bsearch(). Benchmark on 1000 random ints:
26: glibc: 679414 cycles qsort and 3352 cycles bsearch
27: diet libc: 300010 cycles qsort, 553 cycles bsearch
28: I have no idea why glibc is so slow.
1.43 fefe 29: added __isnan (references by libm.a from glibc).
30: added popen() and pclose(). Now the diet libc can link slsh (from
31: S-Lang!)
1.44 fefe 32: corrected usleep. It was more of an msleep until now.
1.45 fefe 33: corrected __dtostr (it looped when trying to format "0.0").
34: several fixes to stdio (thanks to S-Lang's excellent test suite!)
1.46 fefe 35: added ungetc (does it work?)
1.47 fefe 36: added inet_aton (yuck!)
1.50 fefe 37: added gethostbyname, gethostbyaddr, gethostbyname_r and gethostbyaddr_r
1.48 fefe 38: added h_errno and __h_errno_location
39: added inet_ntoa
1.49 fefe 40: added setpgrp and getpgrp, strsignal and execv (thanks, Olaf)
1.51 fefe 41: add bcopy, index, rindex macros
42: add isgraph and isxdigit
1.52 fefe 43: add strftime (ugh!!!)
1.53 olaf 44: Olaf fixed strftime, gmtime_r leap year bug.
1.54 fefe 45: Jeff Garzik contributed a spec file. I put it in contrib and hope
46: everbody knows what to do with it...
47: Olaf checked in a ton of libpthread stuff.
1.55 fefe 48: added mkstemp using /dev/urandom (harder to guess than getpid()).
1.56 fefe 49: add large file backwards compatibility support (so gzip will run on
50: kernel 2.2 and 2.0)
1.57 fefe 51: added shutdown
1.58 fefe 52: added nice
53: added pseudo-locale support for is* (they are now weak aliases,
54: isupper -> __isupper_ascii etc). So, for example, liblatin1 can
55: overwrite the default is* functions.
1.59 fefe 56: added liblatin1.a
1.60 fefe 57: added tzfile implementation (does it work for anyone else?)
1.61 fefe 58: added syslog(3) (thanks to Rene Müller!)
1.62 fefe 59: Jeff Garzik contributed getgr*, getpw* and getsp* in one tiny package.
60: Thanks!
1.63 fefe 61: merged duplicate code in execv and execvp into exec_lib.o
1.34 fefe 62:
1.21 fefe 63: 0.7.2:
64: added several prototypes to already implemented functions and
65: <sys/reboot.h>. Thanks for Abraham for pointing this out.
1.22 fefe 66: added prototypes for process group and dup syscalls. Thanks to
67: Abraham again. He also contributed an implementation for daemon(),
68: an ugly BSD hack. I decided to diversify the lib/ directory a
69: little and now created a libugly/ directory for stuff like system()
70: and daemon() ;-)
71: added libstdio and moved all that stdio crap from lib/ there. The
72: idea is to start separating now what we will separate for shared
73: libraries anyway.
1.26 fefe 74: optimized x86 unified-syscall by Olaf.
1.27 fefe 75: added limits.h, removed __ARCHBITS from endian.h (use __WORDSIZE from
76: limits.h instead)
77: added remove.c and enough cruft to limits.h to make it compile lzo,
78: lzop and bzip2.
1.28 fefe 79: added mktime, asctime_r, asctime and ctime (needed for gzip).
80: Someone will need to implement a reader for the glibc localtime file
1.31 fefe 81: format some day (man tzfile, ugh!) :-(
1.29 fefe 82: added unlink prototype to unistd.h.
83: added gmtime and gmtime_r.
1.31 fefe 84: fixed sparc unified syscall error handling (please upgrade!).
1.30 fefe 85: removed readdir glibc compatibility cruft.
1.31 fefe 86: removed stat glibc compatibility cruft.
1.32 fefe 87: oops, my fread was broken for pipes.
1.33 fefe 88: Olaf found a bug in pipe() for sparc, too.
1.21 fefe 89:
1.9 fefe 90: 0.7.1:
1.11 fefe 91: optimized away the static arrays from __ltostr and __lltostr.
92: optimized away the static arrays (and more!) from strtoul and strtoull.
93: added system (by Olaf).
1.12 fefe 94: added tolower and toupper
1.13 fefe 95: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 96: added clone for mips (by Olaf).
97: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 98: added pause for all (by Olaf).
1.17 fefe 99: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 100: added isupper and islower.
101: added putc and putchar as macros.
1.17 fefe 102: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
103: insightful discussions and patches I partially used for the last three)
1.18 fefe 104: added getpriority and setpriority (also by Abraham)
1.19 fefe 105: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 106: oi, setjmp and longjmp were broken on other platforms, too!
107: Thanks to Sébastien Côté for reporting this.
1.9 fefe 108:
109: 0.7:
1.6 olaf 110: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 111: Initial dynamic loading glue for i386, also by Olaf.
112: We don't have a dynamic loader yet, so this is not yet usable.
113: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
114: vsprintf, writev, all by Olaf.
1.2 fefe 115: imported readdir fix from Guillaume. Mhh, can the kernel interface
116: really be this broken?
1.3 fefe 117: Found a kludgy way to detect whether the kernel includes define struct
118: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
119: not work for you, please tell me!
1.4 fefe 120: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 121: Fixed stdio line buffering. The code was already there but I forgot
122: to mark stdio and stdout as line buffered. This should really be
123: done dynamically using isatty, though.
1.8 fefe 124: Removed superfluous "if (1)" in execvp.c
1.1 cvs 125:
126: 0.6.14:
127: *scanf did not append the 0 byte for %s.
128: added creat and changed creat to call open directly instead of open64.
129: did for fread what 0.6.10 did for fwrite.
130: repaired buffered stdio a little (line buffering is still not
131: supported).
132: [All of these bugs have been reported by Guillaume. Thanks!]
133: added __pure__ to strchr and strrchr in <string.h>
134:
135: 0.6.13:
136: included COPYING to make clear that the diet libc is covered by the
137: GNU General Public License (and _not_ the LGPL). That means that
138: you need to obtain a license from me if you want to use the diet
139: libc in a proprietary program that you want to distribute.
140: The x86 unified syscall swapped arguments #4 and #5 and noone
141: noticed until now! Thanks to Guillaume Cottenceau for reporting
142: this!
143: The malloc had a subtle bug with small allocations that could cause
144: segfault. Guillaume reported it, Olaf fixed it.
145:
146: 0.6.12:
147: added assert, statfs and fstatfs
148: added include/net/if.h so fget compiles again.
149:
150: 0.6.11:
151: added putenv.
152:
153: 0.6.10:
154: fixed fwrite. It returned the number of bytes written, not the number
155: of records. Thanks to Albert D. Cahalan for pointing this out.
156: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
157: added brk and sbrk. harold@nb.com.sg made me do it.
158:
159: 0.6.9:
160: fixed execvp not to return on ENOENT
161: did for {open|read|close|seek|tell}dir what I did for stat before.
162: The glibc compatibility part is currently non-functional.
163: fixed white space handling in sscanf " (" would not match " (".
164: fixed printf to accept 'l' flag.
165: fixed getpwuid (ignored last line)
166:
167: 0.6.8:
168: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
169: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
170: added strlcpy and strlcat from OpenBSD.
171: added stat64, fstat64 and lstat64.
172: added endian.h and made ending of strlen.c endianness-aware
173: added string routines size tweak to dietfeatures.h. Most of the
174: string routines have been unrolled. This is up to three times as
175: fast but creates up to three times larger code. Now you can disable
176: the unrolling.
177: added sendfile.
178: I also added pread. It works on x86 and sparc but not on ppc and mips.
179: I have no idea what's going on. strace disagrees with the kernel.
180: Can anyone help? I asked the Linux kernel mailing list for help, too.
181: added a few aliases of the type __libc_open for open. nm on
182: libpthreads.so indicates that we will need them.
183:
184: 0.6.7:
185: the sources now compile without warnings with -Wall.
186: printf now prints "(null)" when %s is passed NULL. This can be
187: removed with WANT_NULL_PRINTF in dietfeatures.h
188: added vfprintf, execl, ttyname.
189: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
190: __dietfstat so they can use the normal kernel struct stat.
191: stat, lstat and fstat are now C wrappers that convert to the glibc
192: struct stat. So we can avoid that overhead for programs that use
193: dietlibc headers.
194:
195: 0.6.6:
196: changed the Makefiles so you can now set CFLAGS on the command line
197: added memccmp (analogous to memccpy) and strncmp.
198: Olaf fixed another bug in the sparc unified syscall.
199: Paul Clifford contributed a C version of his strlen.S that is much
200: more efficient than the previous strlen on all platforms! It also
201: looks like technology from Roswell. ;-}
202: added a "real" stdio. Well, almost. Please contribute!
203: added dietfeatures.h so you can remove features you don't need.
204: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
205: stdio now works with simple programs and minigzip from libz.
206: stdio uses some major trickery to avoid linking the stdio, stderr and
207: stdout (and reserving space for them) when they are not used.
208:
209: 0.6.5:
210: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
211: most of them bugs in the new header files, but also missing sigset
212: functions.
213: Paul Clifford contributed an assembly strlen.S for ARM and fixed
214: several bugs.
215: Fixed ppc/setjmp and mips/pipe.
216: Olaf contributed a new sparc unified syscall.
217:
218: 0.6.4:
219: declared is* static inline in getservent.
220: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
221: changed strlen to return 0 when passed NULL.
222: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
223: Scientist who actually implemented this without access to a MIPS
224: box, just from reading the architecture manual. And it worked out
225: of the box.
226: started a set of system includes, mainly so that I can use lcc and
227: my alpha-linux cross compiler (which is unable to cross-compile
228: glibc). diet libc can now be compiled without any include files
229: from a normal libc. The includes do declare more than diet libc
230: currently delivers and they are still far from complete for real
231: applications.
232: I will only add assembly versions that are smaller _and_ faster than
233: the C version. Larger routines are only accepted if they are called
234: very often and are substantially faster.
235: added i386 assembly strchr, which is smaller and faster than the old
236: version. It is, however, larger than the version contributed by
237: proton (thanks, anyway).
238: added i386 strlen (31% faster, 14% smaller)
239: "make t" will now create a map file called "mapfile".
240:
241: 0.6.3:
242: added sys_errlist, strerror and perror
243: added isblank
244: added atol
245:
246: 0.6.2:
247: mmap for ARM didn't compile. Thanks, Paul!
248:
249: 0.6.1:
250: split mmap into the architecture specific subdirectories.
251: getenv now copes with environ==NULL, thanks Paul Clifford.
252: Paul also contributed a smaller ARM startup code.
253:
254: 0.6:
255: strcat returned the wrong result. Thanks, Dietz Pröpper.
256: strtod now understands a negative exponent (oops, thanks Bertram Barth)
257: Port to arm-linux-gnu, but on the Netwinder I use for testing the
258: __dtostr does not work (I have no idea why!)
259: The MIPS port now uses (much smaller) non-PIC code. That means that
260: the applications you link against diet libc also have to be compiled
261: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
262: Thanks to Ralf Bächle for helping me with this!
263: I made subdirectories for the architectures and use VPATH to override
264: VPATH so that make finds the source file automatically. That should
265: simplify the sources greatly.
266: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
267: ARM and MIPS now also have setjmp and longjmp
268: "compile" and "load" are now make targets. Use them for djb programs.
269: added contrib/elftrunc.c which will remove unnecessary ELF headers.
270: Again, contributed from Olaf. Great work!
271:
272: 0.5.12:
273: printf also does signed numbers.
274: If you don't use atexit, dietlibc now does not link exit, only _exit.
275: *printf now correctly returns the number of bytes written.
276: Olaf contributed experimental sscanf and vsscanf implementations.
277: If passed NULL as buffer, snprintf will not write anything but still
278: return the number of bytes it would have written.
279: Initial MIPS port! (Oh, the agony!)
280: No setjmp and longjmp support yet! Please contribute!
281: I even implemented unified syscalls for MIPS. Still, MIPS code is
282: almost twice the size of SPARC code. If anyone knows why: please
283: tell me!
284:
285: 0.5.11:
286: I implemented new unified syscalls for x86, sparc and ppc, this time based
287: on .s files and not .c files, and I moved the syscalls into
288: subdirectores to clean the dietlibc sources up a little.
289: A binary consisting of printf("%s is %d\n","olaf",23) is now
290: 2864 bytes on sparc
291: 2488 bytes on intel
292: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
293: to make a few more applications work.
294: Olaf Dreesen also implemented some unified syscall stuff for x86.
295: He discovered that with our unified syscall interface it actually
296: costs <10 bytes total to make all system calls thread safe!
297: I implemented atexit() (can register up to 4 callbacks).
298: I implemented strtod and __dtostr (the opposite). Now we can add
299: floating point support to vsnprintf and sscanf!
300:
301: 0.5.10:
302: I actually saw that I can not only merge the errno handling code of
303: the system calls on x86, I can also merge the rest (including the
304: arguments) except for setting the system call number. All those
305: system calls are now just a jump to a unified system call handler.
306: I got rid of x86openclose again. The savings are substantial:
307: chown from embutils went from 7664 to 7184 bytes!
308: If I'd move the system call wrappers to assembly language, I could
309: even reduce the alignment (does not matter since it's just a jump
310: anyway) and get rid of the "ret" behind each jump (one byte per
311: system call!)
312:
313: 0.5.9:
314: added memchr, strpbrk, strstr, strtol, isspace (hehe)
315: fixed strdup
316: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
317: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
318: (no signed integers, only strings and unsigned integers (but
319: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
320: that code must be wonderful, I don't understand it at all *bg*)
321: Olaf's printf does understand padding like in "%08d" and "%8d" and
322: automatically pads pointers with '0'.
323: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
324: Thanks to proton for inspiration on this.
325: The new x86 start code also does not reference exit any more, thus saving
326: 64 bytes for executables that don't call exit explicitly.
327: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
328: Moved errno and environ to start.S, saving no byte binary size but
329: speeding up compilation and slightly shortening dietlibc.a ;-)
330: For x86: joined open and close into one assembler file, sharing the
331: errno handling. This is a feasibility test and it actually saves a
332: few bytes. I think I will reimplement the _syscall[1-6] macros on
333: all platforms now to share the errno handling code for them.
334:
335: 0.5.8:
336: fixed strchr to be able to look for 0.
337:
338: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
339: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
340: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
341: contributed by Olaf Dreesen)
342:
343: added execvp, getcwd
344:
345: fixed __xmknod
346:
347: Note: can it be that the ftw interface really is so broken that I
348: cannot implement it without having to implement some searching data
349: structure? I included an experimental ftw implementation that will
350: not follow symlinks.
351:
352: 0.5.7:
353: oops, if_nametoindex was broken!
354:
355: 0.5.6:
356: ported to ppc-linux.
357: fixed i386 sigsetjmp (I mistyped the function name)
358: included sigjmp.c
359:
360: 0.5.5:
361: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
362: actually, it wasn't fork that wasn't working for sparc, it was pipe.
363: I fixed it now.
364: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
365: fixed readdir to use getdents and not the intel inline asm
366:
367: 0.5.4:
368: ported to sparclinux (sparc32 only). Beware: does not work yet.
369: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
370:
371: 0.5.3:
372: added tcsetattr and getenv to compile e3
373:
374: 0.5.2:
375: added vhangup, tcgetattr, isatty and memcmp for fgetty
376: (http://www.fefe.de/fgetty/)
377: added localtime from uC-libc.
378:
379: 0.5.1:
380: stat and friends actually work now. Yuck! Another case of
381: translation between kernel and userland. Why can't the kernel
382: people and the libc people simply agree on a standard?
383: removed many unnecessary includes to speed up compilation.
384: added -fomit-frame-pointer and i386 compilation target to reduce code size.
385: inlined socketcall to reduce code size.
386: "load" and "compile" are examples for djb code.
387:
388: 0.5:
389: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
390: each object file is now treated with "strip -x -R .note -R .comment"
391: split each system call into a separate object file
392: added ntohs, htons, alarm, if_indextoname and if_nametoindex
393:
394: 0.4.1:
395: oops, the strchr implementation was wrong. Thanks Jens Laas!
396:
397: 0.4:
398: copied opendir and friends from uC-libc.
399: added getservent and getservby* in gerservent.c
400:
401: 0.3:
402: remove readdir system call
403: added getdents system call
404: added getpwnam/getpwuid implementation in getpwnam.c
405: added a few string functions (in str*.c)
406:
LinuxTV legacy CVS <linuxtv.org/cvs>