Annotation of dietlibc/CHANGES, revision 1.78
1.74 fefe 1: 0.10:
2: update getpagesize.
3: add personality.
1.75 fefe 4: add "%*s" and "%*.*s" support to vsnprintf.
1.76 fefe 5: add putchar (apparently arm-linux-gcc will substitute printf("\n")
6: with putchar('\n') behind your back...?!
1.77 fefe 7: import include/scsi/ from glibc.
1.78 ! fefe 8: add iopl.
1.74 fefe 9:
1.73 fefe 10: 0.9:
1.65 fefe 11: fixed initgroups (oops)
1.66 fefe 12: Cleanup appending file open for stdio (thanks, James Antill)
13: imported vfprintf fix from James Antill.
1.67 fefe 14: add fnmatch
15: change index, rindex and bcmp as weak ELF symbols (were #defined before)
1.68 fefe 16: Olaf added character classes to *scanf.
1.69 fefe 17: Put object and library files into a separate directory (bin-$ARCH).
18: Add "cross" make target to make all architectures.
19: Add "diet" program and man page.
1.70 fefe 20: Add getgrnam, getgrgid
1.71 fefe 21: Add cfmakeraw (Thanks, Antonio Musumeci)
1.72 fefe 22: localtime won't call gettimeofday when the tzfile parser is used.
1.65 fefe 23:
1.64 fefe 24: 0.8:
1.34 fefe 25: vsnprintf(0,...) estimated the length of strings with a padding width
26: incorrectly.
1.35 fefe 27: new reboot() from Olaf.
1.36 fefe 28: added prototypes for getsockopt and setsockopt.
1.39 fefe 29: added experimental sem* and shm*
1.38 fefe 30: define EXIT_SUCCESS and EXIT_FAILURE
1.40 fefe 31: added setvbuf (only changes mode, not buffer)
1.41 fefe 32: added dietwarning.h (woohoo, linker warnings!) and used it to warn of
33: insecurity (system, sprintf) or bloat (printf and stdio).
1.42 fefe 34: added qsort() and bsearch(). Benchmark on 1000 random ints:
35: glibc: 679414 cycles qsort and 3352 cycles bsearch
36: diet libc: 300010 cycles qsort, 553 cycles bsearch
37: I have no idea why glibc is so slow.
1.43 fefe 38: added __isnan (references by libm.a from glibc).
39: added popen() and pclose(). Now the diet libc can link slsh (from
40: S-Lang!)
1.44 fefe 41: corrected usleep. It was more of an msleep until now.
1.45 fefe 42: corrected __dtostr (it looped when trying to format "0.0").
43: several fixes to stdio (thanks to S-Lang's excellent test suite!)
1.46 fefe 44: added ungetc (does it work?)
1.47 fefe 45: added inet_aton (yuck!)
1.50 fefe 46: added gethostbyname, gethostbyaddr, gethostbyname_r and gethostbyaddr_r
1.48 fefe 47: added h_errno and __h_errno_location
48: added inet_ntoa
1.49 fefe 49: added setpgrp and getpgrp, strsignal and execv (thanks, Olaf)
1.51 fefe 50: add bcopy, index, rindex macros
51: add isgraph and isxdigit
1.52 fefe 52: add strftime (ugh!!!)
1.53 olaf 53: Olaf fixed strftime, gmtime_r leap year bug.
1.54 fefe 54: Jeff Garzik contributed a spec file. I put it in contrib and hope
55: everbody knows what to do with it...
56: Olaf checked in a ton of libpthread stuff.
1.55 fefe 57: added mkstemp using /dev/urandom (harder to guess than getpid()).
1.56 fefe 58: add large file backwards compatibility support (so gzip will run on
59: kernel 2.2 and 2.0)
1.57 fefe 60: added shutdown
1.58 fefe 61: added nice
62: added pseudo-locale support for is* (they are now weak aliases,
63: isupper -> __isupper_ascii etc). So, for example, liblatin1 can
64: overwrite the default is* functions.
1.59 fefe 65: added liblatin1.a
1.60 fefe 66: added tzfile implementation (does it work for anyone else?)
1.61 fefe 67: added syslog(3) (thanks to Rene Müller!)
1.62 fefe 68: Jeff Garzik contributed getgr*, getpw* and getsp* in one tiny package.
69: Thanks!
1.63 fefe 70: merged duplicate code in execv and execvp into exec_lib.o
1.34 fefe 71:
1.21 fefe 72: 0.7.2:
73: added several prototypes to already implemented functions and
74: <sys/reboot.h>. Thanks for Abraham for pointing this out.
1.22 fefe 75: added prototypes for process group and dup syscalls. Thanks to
76: Abraham again. He also contributed an implementation for daemon(),
77: an ugly BSD hack. I decided to diversify the lib/ directory a
78: little and now created a libugly/ directory for stuff like system()
79: and daemon() ;-)
80: added libstdio and moved all that stdio crap from lib/ there. The
81: idea is to start separating now what we will separate for shared
82: libraries anyway.
1.26 fefe 83: optimized x86 unified-syscall by Olaf.
1.27 fefe 84: added limits.h, removed __ARCHBITS from endian.h (use __WORDSIZE from
85: limits.h instead)
86: added remove.c and enough cruft to limits.h to make it compile lzo,
87: lzop and bzip2.
1.28 fefe 88: added mktime, asctime_r, asctime and ctime (needed for gzip).
89: Someone will need to implement a reader for the glibc localtime file
1.31 fefe 90: format some day (man tzfile, ugh!) :-(
1.29 fefe 91: added unlink prototype to unistd.h.
92: added gmtime and gmtime_r.
1.31 fefe 93: fixed sparc unified syscall error handling (please upgrade!).
1.30 fefe 94: removed readdir glibc compatibility cruft.
1.31 fefe 95: removed stat glibc compatibility cruft.
1.32 fefe 96: oops, my fread was broken for pipes.
1.33 fefe 97: Olaf found a bug in pipe() for sparc, too.
1.21 fefe 98:
1.9 fefe 99: 0.7.1:
1.11 fefe 100: optimized away the static arrays from __ltostr and __lltostr.
101: optimized away the static arrays (and more!) from strtoul and strtoull.
102: added system (by Olaf).
1.12 fefe 103: added tolower and toupper
1.13 fefe 104: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 105: added clone for mips (by Olaf).
106: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 107: added pause for all (by Olaf).
1.17 fefe 108: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 109: added isupper and islower.
110: added putc and putchar as macros.
1.17 fefe 111: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
112: insightful discussions and patches I partially used for the last three)
1.18 fefe 113: added getpriority and setpriority (also by Abraham)
1.19 fefe 114: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 115: oi, setjmp and longjmp were broken on other platforms, too!
116: Thanks to Sébastien Côté for reporting this.
1.9 fefe 117:
118: 0.7:
1.6 olaf 119: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 120: Initial dynamic loading glue for i386, also by Olaf.
121: We don't have a dynamic loader yet, so this is not yet usable.
122: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
123: vsprintf, writev, all by Olaf.
1.2 fefe 124: imported readdir fix from Guillaume. Mhh, can the kernel interface
125: really be this broken?
1.3 fefe 126: Found a kludgy way to detect whether the kernel includes define struct
127: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
128: not work for you, please tell me!
1.4 fefe 129: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 130: Fixed stdio line buffering. The code was already there but I forgot
131: to mark stdio and stdout as line buffered. This should really be
132: done dynamically using isatty, though.
1.8 fefe 133: Removed superfluous "if (1)" in execvp.c
1.1 cvs 134:
135: 0.6.14:
136: *scanf did not append the 0 byte for %s.
137: added creat and changed creat to call open directly instead of open64.
138: did for fread what 0.6.10 did for fwrite.
139: repaired buffered stdio a little (line buffering is still not
140: supported).
141: [All of these bugs have been reported by Guillaume. Thanks!]
142: added __pure__ to strchr and strrchr in <string.h>
143:
144: 0.6.13:
145: included COPYING to make clear that the diet libc is covered by the
146: GNU General Public License (and _not_ the LGPL). That means that
147: you need to obtain a license from me if you want to use the diet
148: libc in a proprietary program that you want to distribute.
149: The x86 unified syscall swapped arguments #4 and #5 and noone
150: noticed until now! Thanks to Guillaume Cottenceau for reporting
151: this!
152: The malloc had a subtle bug with small allocations that could cause
153: segfault. Guillaume reported it, Olaf fixed it.
154:
155: 0.6.12:
156: added assert, statfs and fstatfs
157: added include/net/if.h so fget compiles again.
158:
159: 0.6.11:
160: added putenv.
161:
162: 0.6.10:
163: fixed fwrite. It returned the number of bytes written, not the number
164: of records. Thanks to Albert D. Cahalan for pointing this out.
165: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
166: added brk and sbrk. harold@nb.com.sg made me do it.
167:
168: 0.6.9:
169: fixed execvp not to return on ENOENT
170: did for {open|read|close|seek|tell}dir what I did for stat before.
171: The glibc compatibility part is currently non-functional.
172: fixed white space handling in sscanf " (" would not match " (".
173: fixed printf to accept 'l' flag.
174: fixed getpwuid (ignored last line)
175:
176: 0.6.8:
177: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
178: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
179: added strlcpy and strlcat from OpenBSD.
180: added stat64, fstat64 and lstat64.
181: added endian.h and made ending of strlen.c endianness-aware
182: added string routines size tweak to dietfeatures.h. Most of the
183: string routines have been unrolled. This is up to three times as
184: fast but creates up to three times larger code. Now you can disable
185: the unrolling.
186: added sendfile.
187: I also added pread. It works on x86 and sparc but not on ppc and mips.
188: I have no idea what's going on. strace disagrees with the kernel.
189: Can anyone help? I asked the Linux kernel mailing list for help, too.
190: added a few aliases of the type __libc_open for open. nm on
191: libpthreads.so indicates that we will need them.
192:
193: 0.6.7:
194: the sources now compile without warnings with -Wall.
195: printf now prints "(null)" when %s is passed NULL. This can be
196: removed with WANT_NULL_PRINTF in dietfeatures.h
197: added vfprintf, execl, ttyname.
198: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
199: __dietfstat so they can use the normal kernel struct stat.
200: stat, lstat and fstat are now C wrappers that convert to the glibc
201: struct stat. So we can avoid that overhead for programs that use
202: dietlibc headers.
203:
204: 0.6.6:
205: changed the Makefiles so you can now set CFLAGS on the command line
206: added memccmp (analogous to memccpy) and strncmp.
207: Olaf fixed another bug in the sparc unified syscall.
208: Paul Clifford contributed a C version of his strlen.S that is much
209: more efficient than the previous strlen on all platforms! It also
210: looks like technology from Roswell. ;-}
211: added a "real" stdio. Well, almost. Please contribute!
212: added dietfeatures.h so you can remove features you don't need.
213: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
214: stdio now works with simple programs and minigzip from libz.
215: stdio uses some major trickery to avoid linking the stdio, stderr and
216: stdout (and reserving space for them) when they are not used.
217:
218: 0.6.5:
219: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
220: most of them bugs in the new header files, but also missing sigset
221: functions.
222: Paul Clifford contributed an assembly strlen.S for ARM and fixed
223: several bugs.
224: Fixed ppc/setjmp and mips/pipe.
225: Olaf contributed a new sparc unified syscall.
226:
227: 0.6.4:
228: declared is* static inline in getservent.
229: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
230: changed strlen to return 0 when passed NULL.
231: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
232: Scientist who actually implemented this without access to a MIPS
233: box, just from reading the architecture manual. And it worked out
234: of the box.
235: started a set of system includes, mainly so that I can use lcc and
236: my alpha-linux cross compiler (which is unable to cross-compile
237: glibc). diet libc can now be compiled without any include files
238: from a normal libc. The includes do declare more than diet libc
239: currently delivers and they are still far from complete for real
240: applications.
241: I will only add assembly versions that are smaller _and_ faster than
242: the C version. Larger routines are only accepted if they are called
243: very often and are substantially faster.
244: added i386 assembly strchr, which is smaller and faster than the old
245: version. It is, however, larger than the version contributed by
246: proton (thanks, anyway).
247: added i386 strlen (31% faster, 14% smaller)
248: "make t" will now create a map file called "mapfile".
249:
250: 0.6.3:
251: added sys_errlist, strerror and perror
252: added isblank
253: added atol
254:
255: 0.6.2:
256: mmap for ARM didn't compile. Thanks, Paul!
257:
258: 0.6.1:
259: split mmap into the architecture specific subdirectories.
260: getenv now copes with environ==NULL, thanks Paul Clifford.
261: Paul also contributed a smaller ARM startup code.
262:
263: 0.6:
264: strcat returned the wrong result. Thanks, Dietz Pröpper.
265: strtod now understands a negative exponent (oops, thanks Bertram Barth)
266: Port to arm-linux-gnu, but on the Netwinder I use for testing the
267: __dtostr does not work (I have no idea why!)
268: The MIPS port now uses (much smaller) non-PIC code. That means that
269: the applications you link against diet libc also have to be compiled
270: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
271: Thanks to Ralf Bächle for helping me with this!
272: I made subdirectories for the architectures and use VPATH to override
273: VPATH so that make finds the source file automatically. That should
274: simplify the sources greatly.
275: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
276: ARM and MIPS now also have setjmp and longjmp
277: "compile" and "load" are now make targets. Use them for djb programs.
278: added contrib/elftrunc.c which will remove unnecessary ELF headers.
279: Again, contributed from Olaf. Great work!
280:
281: 0.5.12:
282: printf also does signed numbers.
283: If you don't use atexit, dietlibc now does not link exit, only _exit.
284: *printf now correctly returns the number of bytes written.
285: Olaf contributed experimental sscanf and vsscanf implementations.
286: If passed NULL as buffer, snprintf will not write anything but still
287: return the number of bytes it would have written.
288: Initial MIPS port! (Oh, the agony!)
289: No setjmp and longjmp support yet! Please contribute!
290: I even implemented unified syscalls for MIPS. Still, MIPS code is
291: almost twice the size of SPARC code. If anyone knows why: please
292: tell me!
293:
294: 0.5.11:
295: I implemented new unified syscalls for x86, sparc and ppc, this time based
296: on .s files and not .c files, and I moved the syscalls into
297: subdirectores to clean the dietlibc sources up a little.
298: A binary consisting of printf("%s is %d\n","olaf",23) is now
299: 2864 bytes on sparc
300: 2488 bytes on intel
301: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
302: to make a few more applications work.
303: Olaf Dreesen also implemented some unified syscall stuff for x86.
304: He discovered that with our unified syscall interface it actually
305: costs <10 bytes total to make all system calls thread safe!
306: I implemented atexit() (can register up to 4 callbacks).
307: I implemented strtod and __dtostr (the opposite). Now we can add
308: floating point support to vsnprintf and sscanf!
309:
310: 0.5.10:
311: I actually saw that I can not only merge the errno handling code of
312: the system calls on x86, I can also merge the rest (including the
313: arguments) except for setting the system call number. All those
314: system calls are now just a jump to a unified system call handler.
315: I got rid of x86openclose again. The savings are substantial:
316: chown from embutils went from 7664 to 7184 bytes!
317: If I'd move the system call wrappers to assembly language, I could
318: even reduce the alignment (does not matter since it's just a jump
319: anyway) and get rid of the "ret" behind each jump (one byte per
320: system call!)
321:
322: 0.5.9:
323: added memchr, strpbrk, strstr, strtol, isspace (hehe)
324: fixed strdup
325: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
326: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
327: (no signed integers, only strings and unsigned integers (but
328: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
329: that code must be wonderful, I don't understand it at all *bg*)
330: Olaf's printf does understand padding like in "%08d" and "%8d" and
331: automatically pads pointers with '0'.
332: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
333: Thanks to proton for inspiration on this.
334: The new x86 start code also does not reference exit any more, thus saving
335: 64 bytes for executables that don't call exit explicitly.
336: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
337: Moved errno and environ to start.S, saving no byte binary size but
338: speeding up compilation and slightly shortening dietlibc.a ;-)
339: For x86: joined open and close into one assembler file, sharing the
340: errno handling. This is a feasibility test and it actually saves a
341: few bytes. I think I will reimplement the _syscall[1-6] macros on
342: all platforms now to share the errno handling code for them.
343:
344: 0.5.8:
345: fixed strchr to be able to look for 0.
346:
347: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
348: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
349: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
350: contributed by Olaf Dreesen)
351:
352: added execvp, getcwd
353:
354: fixed __xmknod
355:
356: Note: can it be that the ftw interface really is so broken that I
357: cannot implement it without having to implement some searching data
358: structure? I included an experimental ftw implementation that will
359: not follow symlinks.
360:
361: 0.5.7:
362: oops, if_nametoindex was broken!
363:
364: 0.5.6:
365: ported to ppc-linux.
366: fixed i386 sigsetjmp (I mistyped the function name)
367: included sigjmp.c
368:
369: 0.5.5:
370: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
371: actually, it wasn't fork that wasn't working for sparc, it was pipe.
372: I fixed it now.
373: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
374: fixed readdir to use getdents and not the intel inline asm
375:
376: 0.5.4:
377: ported to sparclinux (sparc32 only). Beware: does not work yet.
378: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
379:
380: 0.5.3:
381: added tcsetattr and getenv to compile e3
382:
383: 0.5.2:
384: added vhangup, tcgetattr, isatty and memcmp for fgetty
385: (http://www.fefe.de/fgetty/)
386: added localtime from uC-libc.
387:
388: 0.5.1:
389: stat and friends actually work now. Yuck! Another case of
390: translation between kernel and userland. Why can't the kernel
391: people and the libc people simply agree on a standard?
392: removed many unnecessary includes to speed up compilation.
393: added -fomit-frame-pointer and i386 compilation target to reduce code size.
394: inlined socketcall to reduce code size.
395: "load" and "compile" are examples for djb code.
396:
397: 0.5:
398: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
399: each object file is now treated with "strip -x -R .note -R .comment"
400: split each system call into a separate object file
401: added ntohs, htons, alarm, if_indextoname and if_nametoindex
402:
403: 0.4.1:
404: oops, the strchr implementation was wrong. Thanks Jens Laas!
405:
406: 0.4:
407: copied opendir and friends from uC-libc.
408: added getservent and getservby* in gerservent.c
409:
410: 0.3:
411: remove readdir system call
412: added getdents system call
413: added getpwnam/getpwuid implementation in getpwnam.c
414: added a few string functions (in str*.c)
415:
LinuxTV legacy CVS <linuxtv.org/cvs>