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