Annotation of dietlibc/CHANGES, revision 1.52
1.34 fefe 1: 0.7.3:
2: vsnprintf(0,...) estimated the length of strings with a padding width
3: incorrectly.
1.35 fefe 4: new reboot() from Olaf.
1.36 fefe 5: added prototypes for getsockopt and setsockopt.
1.39 fefe 6: added experimental sem* and shm*
1.38 fefe 7: define EXIT_SUCCESS and EXIT_FAILURE
1.40 fefe 8: added setvbuf (only changes mode, not buffer)
1.41 fefe 9: added dietwarning.h (woohoo, linker warnings!) and used it to warn of
10: insecurity (system, sprintf) or bloat (printf and stdio).
1.42 fefe 11: added qsort() and bsearch(). Benchmark on 1000 random ints:
12: glibc: 679414 cycles qsort and 3352 cycles bsearch
13: diet libc: 300010 cycles qsort, 553 cycles bsearch
14: I have no idea why glibc is so slow.
1.43 fefe 15: added __isnan (references by libm.a from glibc).
16: added popen() and pclose(). Now the diet libc can link slsh (from
17: S-Lang!)
1.44 fefe 18: corrected usleep. It was more of an msleep until now.
1.45 fefe 19: corrected __dtostr (it looped when trying to format "0.0").
20: several fixes to stdio (thanks to S-Lang's excellent test suite!)
1.46 fefe 21: added ungetc (does it work?)
1.47 fefe 22: added inet_aton (yuck!)
1.50 fefe 23: added gethostbyname, gethostbyaddr, gethostbyname_r and gethostbyaddr_r
1.48 fefe 24: added h_errno and __h_errno_location
25: added inet_ntoa
1.49 fefe 26: added setpgrp and getpgrp, strsignal and execv (thanks, Olaf)
1.51 fefe 27: add bcopy, index, rindex macros
28: add isgraph and isxdigit
1.52 ! fefe 29: add strftime (ugh!!!)
1.34 fefe 30:
1.21 fefe 31: 0.7.2:
32: added several prototypes to already implemented functions and
33: <sys/reboot.h>. Thanks for Abraham for pointing this out.
1.22 fefe 34: added prototypes for process group and dup syscalls. Thanks to
35: Abraham again. He also contributed an implementation for daemon(),
36: an ugly BSD hack. I decided to diversify the lib/ directory a
37: little and now created a libugly/ directory for stuff like system()
38: and daemon() ;-)
39: added libstdio and moved all that stdio crap from lib/ there. The
40: idea is to start separating now what we will separate for shared
41: libraries anyway.
1.26 fefe 42: optimized x86 unified-syscall by Olaf.
1.27 fefe 43: added limits.h, removed __ARCHBITS from endian.h (use __WORDSIZE from
44: limits.h instead)
45: added remove.c and enough cruft to limits.h to make it compile lzo,
46: lzop and bzip2.
1.28 fefe 47: added mktime, asctime_r, asctime and ctime (needed for gzip).
48: Someone will need to implement a reader for the glibc localtime file
1.31 fefe 49: format some day (man tzfile, ugh!) :-(
1.29 fefe 50: added unlink prototype to unistd.h.
51: added gmtime and gmtime_r.
1.31 fefe 52: fixed sparc unified syscall error handling (please upgrade!).
1.30 fefe 53: removed readdir glibc compatibility cruft.
1.31 fefe 54: removed stat glibc compatibility cruft.
1.32 fefe 55: oops, my fread was broken for pipes.
1.33 fefe 56: Olaf found a bug in pipe() for sparc, too.
1.21 fefe 57:
1.9 fefe 58: 0.7.1:
1.11 fefe 59: optimized away the static arrays from __ltostr and __lltostr.
60: optimized away the static arrays (and more!) from strtoul and strtoull.
61: added system (by Olaf).
1.12 fefe 62: added tolower and toupper
1.13 fefe 63: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 64: added clone for mips (by Olaf).
65: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 66: added pause for all (by Olaf).
1.17 fefe 67: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 68: added isupper and islower.
69: added putc and putchar as macros.
1.17 fefe 70: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
71: insightful discussions and patches I partially used for the last three)
1.18 fefe 72: added getpriority and setpriority (also by Abraham)
1.19 fefe 73: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 74: oi, setjmp and longjmp were broken on other platforms, too!
75: Thanks to Sébastien Côté for reporting this.
1.9 fefe 76:
77: 0.7:
1.6 olaf 78: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 79: Initial dynamic loading glue for i386, also by Olaf.
80: We don't have a dynamic loader yet, so this is not yet usable.
81: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
82: vsprintf, writev, all by Olaf.
1.2 fefe 83: imported readdir fix from Guillaume. Mhh, can the kernel interface
84: really be this broken?
1.3 fefe 85: Found a kludgy way to detect whether the kernel includes define struct
86: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
87: not work for you, please tell me!
1.4 fefe 88: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 89: Fixed stdio line buffering. The code was already there but I forgot
90: to mark stdio and stdout as line buffered. This should really be
91: done dynamically using isatty, though.
1.8 fefe 92: Removed superfluous "if (1)" in execvp.c
1.1 cvs 93:
94: 0.6.14:
95: *scanf did not append the 0 byte for %s.
96: added creat and changed creat to call open directly instead of open64.
97: did for fread what 0.6.10 did for fwrite.
98: repaired buffered stdio a little (line buffering is still not
99: supported).
100: [All of these bugs have been reported by Guillaume. Thanks!]
101: added __pure__ to strchr and strrchr in <string.h>
102:
103: 0.6.13:
104: included COPYING to make clear that the diet libc is covered by the
105: GNU General Public License (and _not_ the LGPL). That means that
106: you need to obtain a license from me if you want to use the diet
107: libc in a proprietary program that you want to distribute.
108: The x86 unified syscall swapped arguments #4 and #5 and noone
109: noticed until now! Thanks to Guillaume Cottenceau for reporting
110: this!
111: The malloc had a subtle bug with small allocations that could cause
112: segfault. Guillaume reported it, Olaf fixed it.
113:
114: 0.6.12:
115: added assert, statfs and fstatfs
116: added include/net/if.h so fget compiles again.
117:
118: 0.6.11:
119: added putenv.
120:
121: 0.6.10:
122: fixed fwrite. It returned the number of bytes written, not the number
123: of records. Thanks to Albert D. Cahalan for pointing this out.
124: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
125: added brk and sbrk. harold@nb.com.sg made me do it.
126:
127: 0.6.9:
128: fixed execvp not to return on ENOENT
129: did for {open|read|close|seek|tell}dir what I did for stat before.
130: The glibc compatibility part is currently non-functional.
131: fixed white space handling in sscanf " (" would not match " (".
132: fixed printf to accept 'l' flag.
133: fixed getpwuid (ignored last line)
134:
135: 0.6.8:
136: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
137: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
138: added strlcpy and strlcat from OpenBSD.
139: added stat64, fstat64 and lstat64.
140: added endian.h and made ending of strlen.c endianness-aware
141: added string routines size tweak to dietfeatures.h. Most of the
142: string routines have been unrolled. This is up to three times as
143: fast but creates up to three times larger code. Now you can disable
144: the unrolling.
145: added sendfile.
146: I also added pread. It works on x86 and sparc but not on ppc and mips.
147: I have no idea what's going on. strace disagrees with the kernel.
148: Can anyone help? I asked the Linux kernel mailing list for help, too.
149: added a few aliases of the type __libc_open for open. nm on
150: libpthreads.so indicates that we will need them.
151:
152: 0.6.7:
153: the sources now compile without warnings with -Wall.
154: printf now prints "(null)" when %s is passed NULL. This can be
155: removed with WANT_NULL_PRINTF in dietfeatures.h
156: added vfprintf, execl, ttyname.
157: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
158: __dietfstat so they can use the normal kernel struct stat.
159: stat, lstat and fstat are now C wrappers that convert to the glibc
160: struct stat. So we can avoid that overhead for programs that use
161: dietlibc headers.
162:
163: 0.6.6:
164: changed the Makefiles so you can now set CFLAGS on the command line
165: added memccmp (analogous to memccpy) and strncmp.
166: Olaf fixed another bug in the sparc unified syscall.
167: Paul Clifford contributed a C version of his strlen.S that is much
168: more efficient than the previous strlen on all platforms! It also
169: looks like technology from Roswell. ;-}
170: added a "real" stdio. Well, almost. Please contribute!
171: added dietfeatures.h so you can remove features you don't need.
172: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
173: stdio now works with simple programs and minigzip from libz.
174: stdio uses some major trickery to avoid linking the stdio, stderr and
175: stdout (and reserving space for them) when they are not used.
176:
177: 0.6.5:
178: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
179: most of them bugs in the new header files, but also missing sigset
180: functions.
181: Paul Clifford contributed an assembly strlen.S for ARM and fixed
182: several bugs.
183: Fixed ppc/setjmp and mips/pipe.
184: Olaf contributed a new sparc unified syscall.
185:
186: 0.6.4:
187: declared is* static inline in getservent.
188: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
189: changed strlen to return 0 when passed NULL.
190: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
191: Scientist who actually implemented this without access to a MIPS
192: box, just from reading the architecture manual. And it worked out
193: of the box.
194: started a set of system includes, mainly so that I can use lcc and
195: my alpha-linux cross compiler (which is unable to cross-compile
196: glibc). diet libc can now be compiled without any include files
197: from a normal libc. The includes do declare more than diet libc
198: currently delivers and they are still far from complete for real
199: applications.
200: I will only add assembly versions that are smaller _and_ faster than
201: the C version. Larger routines are only accepted if they are called
202: very often and are substantially faster.
203: added i386 assembly strchr, which is smaller and faster than the old
204: version. It is, however, larger than the version contributed by
205: proton (thanks, anyway).
206: added i386 strlen (31% faster, 14% smaller)
207: "make t" will now create a map file called "mapfile".
208:
209: 0.6.3:
210: added sys_errlist, strerror and perror
211: added isblank
212: added atol
213:
214: 0.6.2:
215: mmap for ARM didn't compile. Thanks, Paul!
216:
217: 0.6.1:
218: split mmap into the architecture specific subdirectories.
219: getenv now copes with environ==NULL, thanks Paul Clifford.
220: Paul also contributed a smaller ARM startup code.
221:
222: 0.6:
223: strcat returned the wrong result. Thanks, Dietz Pröpper.
224: strtod now understands a negative exponent (oops, thanks Bertram Barth)
225: Port to arm-linux-gnu, but on the Netwinder I use for testing the
226: __dtostr does not work (I have no idea why!)
227: The MIPS port now uses (much smaller) non-PIC code. That means that
228: the applications you link against diet libc also have to be compiled
229: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
230: Thanks to Ralf Bächle for helping me with this!
231: I made subdirectories for the architectures and use VPATH to override
232: VPATH so that make finds the source file automatically. That should
233: simplify the sources greatly.
234: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
235: ARM and MIPS now also have setjmp and longjmp
236: "compile" and "load" are now make targets. Use them for djb programs.
237: added contrib/elftrunc.c which will remove unnecessary ELF headers.
238: Again, contributed from Olaf. Great work!
239:
240: 0.5.12:
241: printf also does signed numbers.
242: If you don't use atexit, dietlibc now does not link exit, only _exit.
243: *printf now correctly returns the number of bytes written.
244: Olaf contributed experimental sscanf and vsscanf implementations.
245: If passed NULL as buffer, snprintf will not write anything but still
246: return the number of bytes it would have written.
247: Initial MIPS port! (Oh, the agony!)
248: No setjmp and longjmp support yet! Please contribute!
249: I even implemented unified syscalls for MIPS. Still, MIPS code is
250: almost twice the size of SPARC code. If anyone knows why: please
251: tell me!
252:
253: 0.5.11:
254: I implemented new unified syscalls for x86, sparc and ppc, this time based
255: on .s files and not .c files, and I moved the syscalls into
256: subdirectores to clean the dietlibc sources up a little.
257: A binary consisting of printf("%s is %d\n","olaf",23) is now
258: 2864 bytes on sparc
259: 2488 bytes on intel
260: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
261: to make a few more applications work.
262: Olaf Dreesen also implemented some unified syscall stuff for x86.
263: He discovered that with our unified syscall interface it actually
264: costs <10 bytes total to make all system calls thread safe!
265: I implemented atexit() (can register up to 4 callbacks).
266: I implemented strtod and __dtostr (the opposite). Now we can add
267: floating point support to vsnprintf and sscanf!
268:
269: 0.5.10:
270: I actually saw that I can not only merge the errno handling code of
271: the system calls on x86, I can also merge the rest (including the
272: arguments) except for setting the system call number. All those
273: system calls are now just a jump to a unified system call handler.
274: I got rid of x86openclose again. The savings are substantial:
275: chown from embutils went from 7664 to 7184 bytes!
276: If I'd move the system call wrappers to assembly language, I could
277: even reduce the alignment (does not matter since it's just a jump
278: anyway) and get rid of the "ret" behind each jump (one byte per
279: system call!)
280:
281: 0.5.9:
282: added memchr, strpbrk, strstr, strtol, isspace (hehe)
283: fixed strdup
284: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
285: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
286: (no signed integers, only strings and unsigned integers (but
287: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
288: that code must be wonderful, I don't understand it at all *bg*)
289: Olaf's printf does understand padding like in "%08d" and "%8d" and
290: automatically pads pointers with '0'.
291: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
292: Thanks to proton for inspiration on this.
293: The new x86 start code also does not reference exit any more, thus saving
294: 64 bytes for executables that don't call exit explicitly.
295: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
296: Moved errno and environ to start.S, saving no byte binary size but
297: speeding up compilation and slightly shortening dietlibc.a ;-)
298: For x86: joined open and close into one assembler file, sharing the
299: errno handling. This is a feasibility test and it actually saves a
300: few bytes. I think I will reimplement the _syscall[1-6] macros on
301: all platforms now to share the errno handling code for them.
302:
303: 0.5.8:
304: fixed strchr to be able to look for 0.
305:
306: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
307: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
308: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
309: contributed by Olaf Dreesen)
310:
311: added execvp, getcwd
312:
313: fixed __xmknod
314:
315: Note: can it be that the ftw interface really is so broken that I
316: cannot implement it without having to implement some searching data
317: structure? I included an experimental ftw implementation that will
318: not follow symlinks.
319:
320: 0.5.7:
321: oops, if_nametoindex was broken!
322:
323: 0.5.6:
324: ported to ppc-linux.
325: fixed i386 sigsetjmp (I mistyped the function name)
326: included sigjmp.c
327:
328: 0.5.5:
329: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
330: actually, it wasn't fork that wasn't working for sparc, it was pipe.
331: I fixed it now.
332: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
333: fixed readdir to use getdents and not the intel inline asm
334:
335: 0.5.4:
336: ported to sparclinux (sparc32 only). Beware: does not work yet.
337: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
338:
339: 0.5.3:
340: added tcsetattr and getenv to compile e3
341:
342: 0.5.2:
343: added vhangup, tcgetattr, isatty and memcmp for fgetty
344: (http://www.fefe.de/fgetty/)
345: added localtime from uC-libc.
346:
347: 0.5.1:
348: stat and friends actually work now. Yuck! Another case of
349: translation between kernel and userland. Why can't the kernel
350: people and the libc people simply agree on a standard?
351: removed many unnecessary includes to speed up compilation.
352: added -fomit-frame-pointer and i386 compilation target to reduce code size.
353: inlined socketcall to reduce code size.
354: "load" and "compile" are examples for djb code.
355:
356: 0.5:
357: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
358: each object file is now treated with "strip -x -R .note -R .comment"
359: split each system call into a separate object file
360: added ntohs, htons, alarm, if_indextoname and if_nametoindex
361:
362: 0.4.1:
363: oops, the strchr implementation was wrong. Thanks Jens Laas!
364:
365: 0.4:
366: copied opendir and friends from uC-libc.
367: added getservent and getservby* in gerservent.c
368:
369: 0.3:
370: remove readdir system call
371: added getdents system call
372: added getpwnam/getpwuid implementation in getpwnam.c
373: added a few string functions (in str*.c)
374:
LinuxTV legacy CVS <linuxtv.org/cvs>