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