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