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