Annotation of dietlibc/CHANGES, revision 1.28
1.21 fefe 1: 0.7.2:
2: added several prototypes to already implemented functions and
3: <sys/reboot.h>. Thanks for Abraham for pointing this out.
1.22 fefe 4: added prototypes for process group and dup syscalls. Thanks to
5: Abraham again. He also contributed an implementation for daemon(),
6: an ugly BSD hack. I decided to diversify the lib/ directory a
7: little and now created a libugly/ directory for stuff like system()
8: and daemon() ;-)
9: added libstdio and moved all that stdio crap from lib/ there. The
10: idea is to start separating now what we will separate for shared
11: libraries anyway.
1.26 fefe 12: optimized x86 unified-syscall by Olaf.
1.27 fefe 13: added limits.h, removed __ARCHBITS from endian.h (use __WORDSIZE from
14: limits.h instead)
15: added remove.c and enough cruft to limits.h to make it compile lzo,
16: lzop and bzip2.
1.28 ! fefe 17: added mktime, asctime_r, asctime and ctime (needed for gzip).
! 18: Someone will need to implement a reader for the glibc localtime file
! 19: format some day :-(
1.21 fefe 20:
1.9 fefe 21: 0.7.1:
1.11 fefe 22: optimized away the static arrays from __ltostr and __lltostr.
23: optimized away the static arrays (and more!) from strtoul and strtoull.
24: added system (by Olaf).
1.12 fefe 25: added tolower and toupper
1.13 fefe 26: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 27: added clone for mips (by Olaf).
28: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 29: added pause for all (by Olaf).
1.17 fefe 30: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 31: added isupper and islower.
32: added putc and putchar as macros.
1.17 fefe 33: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
34: insightful discussions and patches I partially used for the last three)
1.18 fefe 35: added getpriority and setpriority (also by Abraham)
1.19 fefe 36: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 37: oi, setjmp and longjmp were broken on other platforms, too!
38: Thanks to Sébastien Côté for reporting this.
1.9 fefe 39:
40: 0.7:
1.6 olaf 41: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 42: Initial dynamic loading glue for i386, also by Olaf.
43: We don't have a dynamic loader yet, so this is not yet usable.
44: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
45: vsprintf, writev, all by Olaf.
1.2 fefe 46: imported readdir fix from Guillaume. Mhh, can the kernel interface
47: really be this broken?
1.3 fefe 48: Found a kludgy way to detect whether the kernel includes define struct
49: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
50: not work for you, please tell me!
1.4 fefe 51: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 52: Fixed stdio line buffering. The code was already there but I forgot
53: to mark stdio and stdout as line buffered. This should really be
54: done dynamically using isatty, though.
1.8 fefe 55: Removed superfluous "if (1)" in execvp.c
1.1 cvs 56:
57: 0.6.14:
58: *scanf did not append the 0 byte for %s.
59: added creat and changed creat to call open directly instead of open64.
60: did for fread what 0.6.10 did for fwrite.
61: repaired buffered stdio a little (line buffering is still not
62: supported).
63: [All of these bugs have been reported by Guillaume. Thanks!]
64: added __pure__ to strchr and strrchr in <string.h>
65:
66: 0.6.13:
67: included COPYING to make clear that the diet libc is covered by the
68: GNU General Public License (and _not_ the LGPL). That means that
69: you need to obtain a license from me if you want to use the diet
70: libc in a proprietary program that you want to distribute.
71: The x86 unified syscall swapped arguments #4 and #5 and noone
72: noticed until now! Thanks to Guillaume Cottenceau for reporting
73: this!
74: The malloc had a subtle bug with small allocations that could cause
75: segfault. Guillaume reported it, Olaf fixed it.
76:
77: 0.6.12:
78: added assert, statfs and fstatfs
79: added include/net/if.h so fget compiles again.
80:
81: 0.6.11:
82: added putenv.
83:
84: 0.6.10:
85: fixed fwrite. It returned the number of bytes written, not the number
86: of records. Thanks to Albert D. Cahalan for pointing this out.
87: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
88: added brk and sbrk. harold@nb.com.sg made me do it.
89:
90: 0.6.9:
91: fixed execvp not to return on ENOENT
92: did for {open|read|close|seek|tell}dir what I did for stat before.
93: The glibc compatibility part is currently non-functional.
94: fixed white space handling in sscanf " (" would not match " (".
95: fixed printf to accept 'l' flag.
96: fixed getpwuid (ignored last line)
97:
98: 0.6.8:
99: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
100: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
101: added strlcpy and strlcat from OpenBSD.
102: added stat64, fstat64 and lstat64.
103: added endian.h and made ending of strlen.c endianness-aware
104: added string routines size tweak to dietfeatures.h. Most of the
105: string routines have been unrolled. This is up to three times as
106: fast but creates up to three times larger code. Now you can disable
107: the unrolling.
108: added sendfile.
109: I also added pread. It works on x86 and sparc but not on ppc and mips.
110: I have no idea what's going on. strace disagrees with the kernel.
111: Can anyone help? I asked the Linux kernel mailing list for help, too.
112: added a few aliases of the type __libc_open for open. nm on
113: libpthreads.so indicates that we will need them.
114:
115: 0.6.7:
116: the sources now compile without warnings with -Wall.
117: printf now prints "(null)" when %s is passed NULL. This can be
118: removed with WANT_NULL_PRINTF in dietfeatures.h
119: added vfprintf, execl, ttyname.
120: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
121: __dietfstat so they can use the normal kernel struct stat.
122: stat, lstat and fstat are now C wrappers that convert to the glibc
123: struct stat. So we can avoid that overhead for programs that use
124: dietlibc headers.
125:
126: 0.6.6:
127: changed the Makefiles so you can now set CFLAGS on the command line
128: added memccmp (analogous to memccpy) and strncmp.
129: Olaf fixed another bug in the sparc unified syscall.
130: Paul Clifford contributed a C version of his strlen.S that is much
131: more efficient than the previous strlen on all platforms! It also
132: looks like technology from Roswell. ;-}
133: added a "real" stdio. Well, almost. Please contribute!
134: added dietfeatures.h so you can remove features you don't need.
135: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
136: stdio now works with simple programs and minigzip from libz.
137: stdio uses some major trickery to avoid linking the stdio, stderr and
138: stdout (and reserving space for them) when they are not used.
139:
140: 0.6.5:
141: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
142: most of them bugs in the new header files, but also missing sigset
143: functions.
144: Paul Clifford contributed an assembly strlen.S for ARM and fixed
145: several bugs.
146: Fixed ppc/setjmp and mips/pipe.
147: Olaf contributed a new sparc unified syscall.
148:
149: 0.6.4:
150: declared is* static inline in getservent.
151: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
152: changed strlen to return 0 when passed NULL.
153: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
154: Scientist who actually implemented this without access to a MIPS
155: box, just from reading the architecture manual. And it worked out
156: of the box.
157: started a set of system includes, mainly so that I can use lcc and
158: my alpha-linux cross compiler (which is unable to cross-compile
159: glibc). diet libc can now be compiled without any include files
160: from a normal libc. The includes do declare more than diet libc
161: currently delivers and they are still far from complete for real
162: applications.
163: I will only add assembly versions that are smaller _and_ faster than
164: the C version. Larger routines are only accepted if they are called
165: very often and are substantially faster.
166: added i386 assembly strchr, which is smaller and faster than the old
167: version. It is, however, larger than the version contributed by
168: proton (thanks, anyway).
169: added i386 strlen (31% faster, 14% smaller)
170: "make t" will now create a map file called "mapfile".
171:
172: 0.6.3:
173: added sys_errlist, strerror and perror
174: added isblank
175: added atol
176:
177: 0.6.2:
178: mmap for ARM didn't compile. Thanks, Paul!
179:
180: 0.6.1:
181: split mmap into the architecture specific subdirectories.
182: getenv now copes with environ==NULL, thanks Paul Clifford.
183: Paul also contributed a smaller ARM startup code.
184:
185: 0.6:
186: strcat returned the wrong result. Thanks, Dietz Pröpper.
187: strtod now understands a negative exponent (oops, thanks Bertram Barth)
188: Port to arm-linux-gnu, but on the Netwinder I use for testing the
189: __dtostr does not work (I have no idea why!)
190: The MIPS port now uses (much smaller) non-PIC code. That means that
191: the applications you link against diet libc also have to be compiled
192: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
193: Thanks to Ralf Bächle for helping me with this!
194: I made subdirectories for the architectures and use VPATH to override
195: VPATH so that make finds the source file automatically. That should
196: simplify the sources greatly.
197: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
198: ARM and MIPS now also have setjmp and longjmp
199: "compile" and "load" are now make targets. Use them for djb programs.
200: added contrib/elftrunc.c which will remove unnecessary ELF headers.
201: Again, contributed from Olaf. Great work!
202:
203: 0.5.12:
204: printf also does signed numbers.
205: If you don't use atexit, dietlibc now does not link exit, only _exit.
206: *printf now correctly returns the number of bytes written.
207: Olaf contributed experimental sscanf and vsscanf implementations.
208: If passed NULL as buffer, snprintf will not write anything but still
209: return the number of bytes it would have written.
210: Initial MIPS port! (Oh, the agony!)
211: No setjmp and longjmp support yet! Please contribute!
212: I even implemented unified syscalls for MIPS. Still, MIPS code is
213: almost twice the size of SPARC code. If anyone knows why: please
214: tell me!
215:
216: 0.5.11:
217: I implemented new unified syscalls for x86, sparc and ppc, this time based
218: on .s files and not .c files, and I moved the syscalls into
219: subdirectores to clean the dietlibc sources up a little.
220: A binary consisting of printf("%s is %d\n","olaf",23) is now
221: 2864 bytes on sparc
222: 2488 bytes on intel
223: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
224: to make a few more applications work.
225: Olaf Dreesen also implemented some unified syscall stuff for x86.
226: He discovered that with our unified syscall interface it actually
227: costs <10 bytes total to make all system calls thread safe!
228: I implemented atexit() (can register up to 4 callbacks).
229: I implemented strtod and __dtostr (the opposite). Now we can add
230: floating point support to vsnprintf and sscanf!
231:
232: 0.5.10:
233: I actually saw that I can not only merge the errno handling code of
234: the system calls on x86, I can also merge the rest (including the
235: arguments) except for setting the system call number. All those
236: system calls are now just a jump to a unified system call handler.
237: I got rid of x86openclose again. The savings are substantial:
238: chown from embutils went from 7664 to 7184 bytes!
239: If I'd move the system call wrappers to assembly language, I could
240: even reduce the alignment (does not matter since it's just a jump
241: anyway) and get rid of the "ret" behind each jump (one byte per
242: system call!)
243:
244: 0.5.9:
245: added memchr, strpbrk, strstr, strtol, isspace (hehe)
246: fixed strdup
247: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
248: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
249: (no signed integers, only strings and unsigned integers (but
250: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
251: that code must be wonderful, I don't understand it at all *bg*)
252: Olaf's printf does understand padding like in "%08d" and "%8d" and
253: automatically pads pointers with '0'.
254: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
255: Thanks to proton for inspiration on this.
256: The new x86 start code also does not reference exit any more, thus saving
257: 64 bytes for executables that don't call exit explicitly.
258: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
259: Moved errno and environ to start.S, saving no byte binary size but
260: speeding up compilation and slightly shortening dietlibc.a ;-)
261: For x86: joined open and close into one assembler file, sharing the
262: errno handling. This is a feasibility test and it actually saves a
263: few bytes. I think I will reimplement the _syscall[1-6] macros on
264: all platforms now to share the errno handling code for them.
265:
266: 0.5.8:
267: fixed strchr to be able to look for 0.
268:
269: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
270: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
271: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
272: contributed by Olaf Dreesen)
273:
274: added execvp, getcwd
275:
276: fixed __xmknod
277:
278: Note: can it be that the ftw interface really is so broken that I
279: cannot implement it without having to implement some searching data
280: structure? I included an experimental ftw implementation that will
281: not follow symlinks.
282:
283: 0.5.7:
284: oops, if_nametoindex was broken!
285:
286: 0.5.6:
287: ported to ppc-linux.
288: fixed i386 sigsetjmp (I mistyped the function name)
289: included sigjmp.c
290:
291: 0.5.5:
292: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
293: actually, it wasn't fork that wasn't working for sparc, it was pipe.
294: I fixed it now.
295: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
296: fixed readdir to use getdents and not the intel inline asm
297:
298: 0.5.4:
299: ported to sparclinux (sparc32 only). Beware: does not work yet.
300: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
301:
302: 0.5.3:
303: added tcsetattr and getenv to compile e3
304:
305: 0.5.2:
306: added vhangup, tcgetattr, isatty and memcmp for fgetty
307: (http://www.fefe.de/fgetty/)
308: added localtime from uC-libc.
309:
310: 0.5.1:
311: stat and friends actually work now. Yuck! Another case of
312: translation between kernel and userland. Why can't the kernel
313: people and the libc people simply agree on a standard?
314: removed many unnecessary includes to speed up compilation.
315: added -fomit-frame-pointer and i386 compilation target to reduce code size.
316: inlined socketcall to reduce code size.
317: "load" and "compile" are examples for djb code.
318:
319: 0.5:
320: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
321: each object file is now treated with "strip -x -R .note -R .comment"
322: split each system call into a separate object file
323: added ntohs, htons, alarm, if_indextoname and if_nametoindex
324:
325: 0.4.1:
326: oops, the strchr implementation was wrong. Thanks Jens Laas!
327:
328: 0.4:
329: copied opendir and friends from uC-libc.
330: added getservent and getservby* in gerservent.c
331:
332: 0.3:
333: remove readdir system call
334: added getdents system call
335: added getpwnam/getpwuid implementation in getpwnam.c
336: added a few string functions (in str*.c)
337:
LinuxTV legacy CVS <linuxtv.org/cvs>