Annotation of dietlibc/CHANGES, revision 1.70
1.65 fefe 1: 0.8.1:
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
1.65 fefe 12:
1.64 fefe 13: 0.8:
1.34 fefe 14: vsnprintf(0,...) estimated the length of strings with a padding width
15: incorrectly.
1.35 fefe 16: new reboot() from Olaf.
1.36 fefe 17: added prototypes for getsockopt and setsockopt.
1.39 fefe 18: added experimental sem* and shm*
1.38 fefe 19: define EXIT_SUCCESS and EXIT_FAILURE
1.40 fefe 20: added setvbuf (only changes mode, not buffer)
1.41 fefe 21: added dietwarning.h (woohoo, linker warnings!) and used it to warn of
22: insecurity (system, sprintf) or bloat (printf and stdio).
1.42 fefe 23: added qsort() and bsearch(). Benchmark on 1000 random ints:
24: glibc: 679414 cycles qsort and 3352 cycles bsearch
25: diet libc: 300010 cycles qsort, 553 cycles bsearch
26: I have no idea why glibc is so slow.
1.43 fefe 27: added __isnan (references by libm.a from glibc).
28: added popen() and pclose(). Now the diet libc can link slsh (from
29: S-Lang!)
1.44 fefe 30: corrected usleep. It was more of an msleep until now.
1.45 fefe 31: corrected __dtostr (it looped when trying to format "0.0").
32: several fixes to stdio (thanks to S-Lang's excellent test suite!)
1.46 fefe 33: added ungetc (does it work?)
1.47 fefe 34: added inet_aton (yuck!)
1.50 fefe 35: added gethostbyname, gethostbyaddr, gethostbyname_r and gethostbyaddr_r
1.48 fefe 36: added h_errno and __h_errno_location
37: added inet_ntoa
1.49 fefe 38: added setpgrp and getpgrp, strsignal and execv (thanks, Olaf)
1.51 fefe 39: add bcopy, index, rindex macros
40: add isgraph and isxdigit
1.52 fefe 41: add strftime (ugh!!!)
1.53 olaf 42: Olaf fixed strftime, gmtime_r leap year bug.
1.54 fefe 43: Jeff Garzik contributed a spec file. I put it in contrib and hope
44: everbody knows what to do with it...
45: Olaf checked in a ton of libpthread stuff.
1.55 fefe 46: added mkstemp using /dev/urandom (harder to guess than getpid()).
1.56 fefe 47: add large file backwards compatibility support (so gzip will run on
48: kernel 2.2 and 2.0)
1.57 fefe 49: added shutdown
1.58 fefe 50: added nice
51: added pseudo-locale support for is* (they are now weak aliases,
52: isupper -> __isupper_ascii etc). So, for example, liblatin1 can
53: overwrite the default is* functions.
1.59 fefe 54: added liblatin1.a
1.60 fefe 55: added tzfile implementation (does it work for anyone else?)
1.61 fefe 56: added syslog(3) (thanks to Rene Müller!)
1.62 fefe 57: Jeff Garzik contributed getgr*, getpw* and getsp* in one tiny package.
58: Thanks!
1.63 fefe 59: merged duplicate code in execv and execvp into exec_lib.o
1.34 fefe 60:
1.21 fefe 61: 0.7.2:
62: added several prototypes to already implemented functions and
63: <sys/reboot.h>. Thanks for Abraham for pointing this out.
1.22 fefe 64: added prototypes for process group and dup syscalls. Thanks to
65: Abraham again. He also contributed an implementation for daemon(),
66: an ugly BSD hack. I decided to diversify the lib/ directory a
67: little and now created a libugly/ directory for stuff like system()
68: and daemon() ;-)
69: added libstdio and moved all that stdio crap from lib/ there. The
70: idea is to start separating now what we will separate for shared
71: libraries anyway.
1.26 fefe 72: optimized x86 unified-syscall by Olaf.
1.27 fefe 73: added limits.h, removed __ARCHBITS from endian.h (use __WORDSIZE from
74: limits.h instead)
75: added remove.c and enough cruft to limits.h to make it compile lzo,
76: lzop and bzip2.
1.28 fefe 77: added mktime, asctime_r, asctime and ctime (needed for gzip).
78: Someone will need to implement a reader for the glibc localtime file
1.31 fefe 79: format some day (man tzfile, ugh!) :-(
1.29 fefe 80: added unlink prototype to unistd.h.
81: added gmtime and gmtime_r.
1.31 fefe 82: fixed sparc unified syscall error handling (please upgrade!).
1.30 fefe 83: removed readdir glibc compatibility cruft.
1.31 fefe 84: removed stat glibc compatibility cruft.
1.32 fefe 85: oops, my fread was broken for pipes.
1.33 fefe 86: Olaf found a bug in pipe() for sparc, too.
1.21 fefe 87:
1.9 fefe 88: 0.7.1:
1.11 fefe 89: optimized away the static arrays from __ltostr and __lltostr.
90: optimized away the static arrays (and more!) from strtoul and strtoull.
91: added system (by Olaf).
1.12 fefe 92: added tolower and toupper
1.13 fefe 93: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 94: added clone for mips (by Olaf).
95: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 96: added pause for all (by Olaf).
1.17 fefe 97: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 98: added isupper and islower.
99: added putc and putchar as macros.
1.17 fefe 100: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
101: insightful discussions and patches I partially used for the last three)
1.18 fefe 102: added getpriority and setpriority (also by Abraham)
1.19 fefe 103: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 104: oi, setjmp and longjmp were broken on other platforms, too!
105: Thanks to Sébastien Côté for reporting this.
1.9 fefe 106:
107: 0.7:
1.6 olaf 108: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 109: Initial dynamic loading glue for i386, also by Olaf.
110: We don't have a dynamic loader yet, so this is not yet usable.
111: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
112: vsprintf, writev, all by Olaf.
1.2 fefe 113: imported readdir fix from Guillaume. Mhh, can the kernel interface
114: really be this broken?
1.3 fefe 115: Found a kludgy way to detect whether the kernel includes define struct
116: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
117: not work for you, please tell me!
1.4 fefe 118: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 119: Fixed stdio line buffering. The code was already there but I forgot
120: to mark stdio and stdout as line buffered. This should really be
121: done dynamically using isatty, though.
1.8 fefe 122: Removed superfluous "if (1)" in execvp.c
1.1 cvs 123:
124: 0.6.14:
125: *scanf did not append the 0 byte for %s.
126: added creat and changed creat to call open directly instead of open64.
127: did for fread what 0.6.10 did for fwrite.
128: repaired buffered stdio a little (line buffering is still not
129: supported).
130: [All of these bugs have been reported by Guillaume. Thanks!]
131: added __pure__ to strchr and strrchr in <string.h>
132:
133: 0.6.13:
134: included COPYING to make clear that the diet libc is covered by the
135: GNU General Public License (and _not_ the LGPL). That means that
136: you need to obtain a license from me if you want to use the diet
137: libc in a proprietary program that you want to distribute.
138: The x86 unified syscall swapped arguments #4 and #5 and noone
139: noticed until now! Thanks to Guillaume Cottenceau for reporting
140: this!
141: The malloc had a subtle bug with small allocations that could cause
142: segfault. Guillaume reported it, Olaf fixed it.
143:
144: 0.6.12:
145: added assert, statfs and fstatfs
146: added include/net/if.h so fget compiles again.
147:
148: 0.6.11:
149: added putenv.
150:
151: 0.6.10:
152: fixed fwrite. It returned the number of bytes written, not the number
153: of records. Thanks to Albert D. Cahalan for pointing this out.
154: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
155: added brk and sbrk. harold@nb.com.sg made me do it.
156:
157: 0.6.9:
158: fixed execvp not to return on ENOENT
159: did for {open|read|close|seek|tell}dir what I did for stat before.
160: The glibc compatibility part is currently non-functional.
161: fixed white space handling in sscanf " (" would not match " (".
162: fixed printf to accept 'l' flag.
163: fixed getpwuid (ignored last line)
164:
165: 0.6.8:
166: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
167: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
168: added strlcpy and strlcat from OpenBSD.
169: added stat64, fstat64 and lstat64.
170: added endian.h and made ending of strlen.c endianness-aware
171: added string routines size tweak to dietfeatures.h. Most of the
172: string routines have been unrolled. This is up to three times as
173: fast but creates up to three times larger code. Now you can disable
174: the unrolling.
175: added sendfile.
176: I also added pread. It works on x86 and sparc but not on ppc and mips.
177: I have no idea what's going on. strace disagrees with the kernel.
178: Can anyone help? I asked the Linux kernel mailing list for help, too.
179: added a few aliases of the type __libc_open for open. nm on
180: libpthreads.so indicates that we will need them.
181:
182: 0.6.7:
183: the sources now compile without warnings with -Wall.
184: printf now prints "(null)" when %s is passed NULL. This can be
185: removed with WANT_NULL_PRINTF in dietfeatures.h
186: added vfprintf, execl, ttyname.
187: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
188: __dietfstat so they can use the normal kernel struct stat.
189: stat, lstat and fstat are now C wrappers that convert to the glibc
190: struct stat. So we can avoid that overhead for programs that use
191: dietlibc headers.
192:
193: 0.6.6:
194: changed the Makefiles so you can now set CFLAGS on the command line
195: added memccmp (analogous to memccpy) and strncmp.
196: Olaf fixed another bug in the sparc unified syscall.
197: Paul Clifford contributed a C version of his strlen.S that is much
198: more efficient than the previous strlen on all platforms! It also
199: looks like technology from Roswell. ;-}
200: added a "real" stdio. Well, almost. Please contribute!
201: added dietfeatures.h so you can remove features you don't need.
202: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
203: stdio now works with simple programs and minigzip from libz.
204: stdio uses some major trickery to avoid linking the stdio, stderr and
205: stdout (and reserving space for them) when they are not used.
206:
207: 0.6.5:
208: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
209: most of them bugs in the new header files, but also missing sigset
210: functions.
211: Paul Clifford contributed an assembly strlen.S for ARM and fixed
212: several bugs.
213: Fixed ppc/setjmp and mips/pipe.
214: Olaf contributed a new sparc unified syscall.
215:
216: 0.6.4:
217: declared is* static inline in getservent.
218: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
219: changed strlen to return 0 when passed NULL.
220: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
221: Scientist who actually implemented this without access to a MIPS
222: box, just from reading the architecture manual. And it worked out
223: of the box.
224: started a set of system includes, mainly so that I can use lcc and
225: my alpha-linux cross compiler (which is unable to cross-compile
226: glibc). diet libc can now be compiled without any include files
227: from a normal libc. The includes do declare more than diet libc
228: currently delivers and they are still far from complete for real
229: applications.
230: I will only add assembly versions that are smaller _and_ faster than
231: the C version. Larger routines are only accepted if they are called
232: very often and are substantially faster.
233: added i386 assembly strchr, which is smaller and faster than the old
234: version. It is, however, larger than the version contributed by
235: proton (thanks, anyway).
236: added i386 strlen (31% faster, 14% smaller)
237: "make t" will now create a map file called "mapfile".
238:
239: 0.6.3:
240: added sys_errlist, strerror and perror
241: added isblank
242: added atol
243:
244: 0.6.2:
245: mmap for ARM didn't compile. Thanks, Paul!
246:
247: 0.6.1:
248: split mmap into the architecture specific subdirectories.
249: getenv now copes with environ==NULL, thanks Paul Clifford.
250: Paul also contributed a smaller ARM startup code.
251:
252: 0.6:
253: strcat returned the wrong result. Thanks, Dietz Pröpper.
254: strtod now understands a negative exponent (oops, thanks Bertram Barth)
255: Port to arm-linux-gnu, but on the Netwinder I use for testing the
256: __dtostr does not work (I have no idea why!)
257: The MIPS port now uses (much smaller) non-PIC code. That means that
258: the applications you link against diet libc also have to be compiled
259: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
260: Thanks to Ralf Bächle for helping me with this!
261: I made subdirectories for the architectures and use VPATH to override
262: VPATH so that make finds the source file automatically. That should
263: simplify the sources greatly.
264: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
265: ARM and MIPS now also have setjmp and longjmp
266: "compile" and "load" are now make targets. Use them for djb programs.
267: added contrib/elftrunc.c which will remove unnecessary ELF headers.
268: Again, contributed from Olaf. Great work!
269:
270: 0.5.12:
271: printf also does signed numbers.
272: If you don't use atexit, dietlibc now does not link exit, only _exit.
273: *printf now correctly returns the number of bytes written.
274: Olaf contributed experimental sscanf and vsscanf implementations.
275: If passed NULL as buffer, snprintf will not write anything but still
276: return the number of bytes it would have written.
277: Initial MIPS port! (Oh, the agony!)
278: No setjmp and longjmp support yet! Please contribute!
279: I even implemented unified syscalls for MIPS. Still, MIPS code is
280: almost twice the size of SPARC code. If anyone knows why: please
281: tell me!
282:
283: 0.5.11:
284: I implemented new unified syscalls for x86, sparc and ppc, this time based
285: on .s files and not .c files, and I moved the syscalls into
286: subdirectores to clean the dietlibc sources up a little.
287: A binary consisting of printf("%s is %d\n","olaf",23) is now
288: 2864 bytes on sparc
289: 2488 bytes on intel
290: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
291: to make a few more applications work.
292: Olaf Dreesen also implemented some unified syscall stuff for x86.
293: He discovered that with our unified syscall interface it actually
294: costs <10 bytes total to make all system calls thread safe!
295: I implemented atexit() (can register up to 4 callbacks).
296: I implemented strtod and __dtostr (the opposite). Now we can add
297: floating point support to vsnprintf and sscanf!
298:
299: 0.5.10:
300: I actually saw that I can not only merge the errno handling code of
301: the system calls on x86, I can also merge the rest (including the
302: arguments) except for setting the system call number. All those
303: system calls are now just a jump to a unified system call handler.
304: I got rid of x86openclose again. The savings are substantial:
305: chown from embutils went from 7664 to 7184 bytes!
306: If I'd move the system call wrappers to assembly language, I could
307: even reduce the alignment (does not matter since it's just a jump
308: anyway) and get rid of the "ret" behind each jump (one byte per
309: system call!)
310:
311: 0.5.9:
312: added memchr, strpbrk, strstr, strtol, isspace (hehe)
313: fixed strdup
314: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
315: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
316: (no signed integers, only strings and unsigned integers (but
317: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
318: that code must be wonderful, I don't understand it at all *bg*)
319: Olaf's printf does understand padding like in "%08d" and "%8d" and
320: automatically pads pointers with '0'.
321: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
322: Thanks to proton for inspiration on this.
323: The new x86 start code also does not reference exit any more, thus saving
324: 64 bytes for executables that don't call exit explicitly.
325: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
326: Moved errno and environ to start.S, saving no byte binary size but
327: speeding up compilation and slightly shortening dietlibc.a ;-)
328: For x86: joined open and close into one assembler file, sharing the
329: errno handling. This is a feasibility test and it actually saves a
330: few bytes. I think I will reimplement the _syscall[1-6] macros on
331: all platforms now to share the errno handling code for them.
332:
333: 0.5.8:
334: fixed strchr to be able to look for 0.
335:
336: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
337: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
338: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
339: contributed by Olaf Dreesen)
340:
341: added execvp, getcwd
342:
343: fixed __xmknod
344:
345: Note: can it be that the ftw interface really is so broken that I
346: cannot implement it without having to implement some searching data
347: structure? I included an experimental ftw implementation that will
348: not follow symlinks.
349:
350: 0.5.7:
351: oops, if_nametoindex was broken!
352:
353: 0.5.6:
354: ported to ppc-linux.
355: fixed i386 sigsetjmp (I mistyped the function name)
356: included sigjmp.c
357:
358: 0.5.5:
359: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
360: actually, it wasn't fork that wasn't working for sparc, it was pipe.
361: I fixed it now.
362: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
363: fixed readdir to use getdents and not the intel inline asm
364:
365: 0.5.4:
366: ported to sparclinux (sparc32 only). Beware: does not work yet.
367: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
368:
369: 0.5.3:
370: added tcsetattr and getenv to compile e3
371:
372: 0.5.2:
373: added vhangup, tcgetattr, isatty and memcmp for fgetty
374: (http://www.fefe.de/fgetty/)
375: added localtime from uC-libc.
376:
377: 0.5.1:
378: stat and friends actually work now. Yuck! Another case of
379: translation between kernel and userland. Why can't the kernel
380: people and the libc people simply agree on a standard?
381: removed many unnecessary includes to speed up compilation.
382: added -fomit-frame-pointer and i386 compilation target to reduce code size.
383: inlined socketcall to reduce code size.
384: "load" and "compile" are examples for djb code.
385:
386: 0.5:
387: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
388: each object file is now treated with "strip -x -R .note -R .comment"
389: split each system call into a separate object file
390: added ntohs, htons, alarm, if_indextoname and if_nametoindex
391:
392: 0.4.1:
393: oops, the strchr implementation was wrong. Thanks Jens Laas!
394:
395: 0.4:
396: copied opendir and friends from uC-libc.
397: added getservent and getservby* in gerservent.c
398:
399: 0.3:
400: remove readdir system call
401: added getdents system call
402: added getpwnam/getpwuid implementation in getpwnam.c
403: added a few string functions (in str*.c)
404:
LinuxTV legacy CVS <linuxtv.org/cvs>