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