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