Annotation of dietlibc/CHANGES, revision 1.23
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.
1.23 ! fefe 12: test
1.21 fefe 13:
1.9 fefe 14: 0.7.1:
1.11 fefe 15: optimized away the static arrays from __ltostr and __lltostr.
16: optimized away the static arrays (and more!) from strtoul and strtoull.
17: added system (by Olaf).
1.12 fefe 18: added tolower and toupper
1.13 fefe 19: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 20: added clone for mips (by Olaf).
21: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 22: added pause for all (by Olaf).
1.17 fefe 23: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 24: added isupper and islower.
25: added putc and putchar as macros.
1.17 fefe 26: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
27: insightful discussions and patches I partially used for the last three)
1.18 fefe 28: added getpriority and setpriority (also by Abraham)
1.19 fefe 29: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 30: oi, setjmp and longjmp were broken on other platforms, too!
31: Thanks to Sébastien Côté for reporting this.
1.9 fefe 32:
33: 0.7:
1.6 olaf 34: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 35: Initial dynamic loading glue for i386, also by Olaf.
36: We don't have a dynamic loader yet, so this is not yet usable.
37: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
38: vsprintf, writev, all by Olaf.
1.2 fefe 39: imported readdir fix from Guillaume. Mhh, can the kernel interface
40: really be this broken?
1.3 fefe 41: Found a kludgy way to detect whether the kernel includes define struct
42: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
43: not work for you, please tell me!
1.4 fefe 44: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 45: Fixed stdio line buffering. The code was already there but I forgot
46: to mark stdio and stdout as line buffered. This should really be
47: done dynamically using isatty, though.
1.8 fefe 48: Removed superfluous "if (1)" in execvp.c
1.1 cvs 49:
50: 0.6.14:
51: *scanf did not append the 0 byte for %s.
52: added creat and changed creat to call open directly instead of open64.
53: did for fread what 0.6.10 did for fwrite.
54: repaired buffered stdio a little (line buffering is still not
55: supported).
56: [All of these bugs have been reported by Guillaume. Thanks!]
57: added __pure__ to strchr and strrchr in <string.h>
58:
59: 0.6.13:
60: included COPYING to make clear that the diet libc is covered by the
61: GNU General Public License (and _not_ the LGPL). That means that
62: you need to obtain a license from me if you want to use the diet
63: libc in a proprietary program that you want to distribute.
64: The x86 unified syscall swapped arguments #4 and #5 and noone
65: noticed until now! Thanks to Guillaume Cottenceau for reporting
66: this!
67: The malloc had a subtle bug with small allocations that could cause
68: segfault. Guillaume reported it, Olaf fixed it.
69:
70: 0.6.12:
71: added assert, statfs and fstatfs
72: added include/net/if.h so fget compiles again.
73:
74: 0.6.11:
75: added putenv.
76:
77: 0.6.10:
78: fixed fwrite. It returned the number of bytes written, not the number
79: of records. Thanks to Albert D. Cahalan for pointing this out.
80: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
81: added brk and sbrk. harold@nb.com.sg made me do it.
82:
83: 0.6.9:
84: fixed execvp not to return on ENOENT
85: did for {open|read|close|seek|tell}dir what I did for stat before.
86: The glibc compatibility part is currently non-functional.
87: fixed white space handling in sscanf " (" would not match " (".
88: fixed printf to accept 'l' flag.
89: fixed getpwuid (ignored last line)
90:
91: 0.6.8:
92: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
93: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
94: added strlcpy and strlcat from OpenBSD.
95: added stat64, fstat64 and lstat64.
96: added endian.h and made ending of strlen.c endianness-aware
97: added string routines size tweak to dietfeatures.h. Most of the
98: string routines have been unrolled. This is up to three times as
99: fast but creates up to three times larger code. Now you can disable
100: the unrolling.
101: added sendfile.
102: I also added pread. It works on x86 and sparc but not on ppc and mips.
103: I have no idea what's going on. strace disagrees with the kernel.
104: Can anyone help? I asked the Linux kernel mailing list for help, too.
105: added a few aliases of the type __libc_open for open. nm on
106: libpthreads.so indicates that we will need them.
107:
108: 0.6.7:
109: the sources now compile without warnings with -Wall.
110: printf now prints "(null)" when %s is passed NULL. This can be
111: removed with WANT_NULL_PRINTF in dietfeatures.h
112: added vfprintf, execl, ttyname.
113: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
114: __dietfstat so they can use the normal kernel struct stat.
115: stat, lstat and fstat are now C wrappers that convert to the glibc
116: struct stat. So we can avoid that overhead for programs that use
117: dietlibc headers.
118:
119: 0.6.6:
120: changed the Makefiles so you can now set CFLAGS on the command line
121: added memccmp (analogous to memccpy) and strncmp.
122: Olaf fixed another bug in the sparc unified syscall.
123: Paul Clifford contributed a C version of his strlen.S that is much
124: more efficient than the previous strlen on all platforms! It also
125: looks like technology from Roswell. ;-}
126: added a "real" stdio. Well, almost. Please contribute!
127: added dietfeatures.h so you can remove features you don't need.
128: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
129: stdio now works with simple programs and minigzip from libz.
130: stdio uses some major trickery to avoid linking the stdio, stderr and
131: stdout (and reserving space for them) when they are not used.
132:
133: 0.6.5:
134: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
135: most of them bugs in the new header files, but also missing sigset
136: functions.
137: Paul Clifford contributed an assembly strlen.S for ARM and fixed
138: several bugs.
139: Fixed ppc/setjmp and mips/pipe.
140: Olaf contributed a new sparc unified syscall.
141:
142: 0.6.4:
143: declared is* static inline in getservent.
144: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
145: changed strlen to return 0 when passed NULL.
146: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
147: Scientist who actually implemented this without access to a MIPS
148: box, just from reading the architecture manual. And it worked out
149: of the box.
150: started a set of system includes, mainly so that I can use lcc and
151: my alpha-linux cross compiler (which is unable to cross-compile
152: glibc). diet libc can now be compiled without any include files
153: from a normal libc. The includes do declare more than diet libc
154: currently delivers and they are still far from complete for real
155: applications.
156: I will only add assembly versions that are smaller _and_ faster than
157: the C version. Larger routines are only accepted if they are called
158: very often and are substantially faster.
159: added i386 assembly strchr, which is smaller and faster than the old
160: version. It is, however, larger than the version contributed by
161: proton (thanks, anyway).
162: added i386 strlen (31% faster, 14% smaller)
163: "make t" will now create a map file called "mapfile".
164:
165: 0.6.3:
166: added sys_errlist, strerror and perror
167: added isblank
168: added atol
169:
170: 0.6.2:
171: mmap for ARM didn't compile. Thanks, Paul!
172:
173: 0.6.1:
174: split mmap into the architecture specific subdirectories.
175: getenv now copes with environ==NULL, thanks Paul Clifford.
176: Paul also contributed a smaller ARM startup code.
177:
178: 0.6:
179: strcat returned the wrong result. Thanks, Dietz Pröpper.
180: strtod now understands a negative exponent (oops, thanks Bertram Barth)
181: Port to arm-linux-gnu, but on the Netwinder I use for testing the
182: __dtostr does not work (I have no idea why!)
183: The MIPS port now uses (much smaller) non-PIC code. That means that
184: the applications you link against diet libc also have to be compiled
185: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
186: Thanks to Ralf Bächle for helping me with this!
187: I made subdirectories for the architectures and use VPATH to override
188: VPATH so that make finds the source file automatically. That should
189: simplify the sources greatly.
190: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
191: ARM and MIPS now also have setjmp and longjmp
192: "compile" and "load" are now make targets. Use them for djb programs.
193: added contrib/elftrunc.c which will remove unnecessary ELF headers.
194: Again, contributed from Olaf. Great work!
195:
196: 0.5.12:
197: printf also does signed numbers.
198: If you don't use atexit, dietlibc now does not link exit, only _exit.
199: *printf now correctly returns the number of bytes written.
200: Olaf contributed experimental sscanf and vsscanf implementations.
201: If passed NULL as buffer, snprintf will not write anything but still
202: return the number of bytes it would have written.
203: Initial MIPS port! (Oh, the agony!)
204: No setjmp and longjmp support yet! Please contribute!
205: I even implemented unified syscalls for MIPS. Still, MIPS code is
206: almost twice the size of SPARC code. If anyone knows why: please
207: tell me!
208:
209: 0.5.11:
210: I implemented new unified syscalls for x86, sparc and ppc, this time based
211: on .s files and not .c files, and I moved the syscalls into
212: subdirectores to clean the dietlibc sources up a little.
213: A binary consisting of printf("%s is %d\n","olaf",23) is now
214: 2864 bytes on sparc
215: 2488 bytes on intel
216: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
217: to make a few more applications work.
218: Olaf Dreesen also implemented some unified syscall stuff for x86.
219: He discovered that with our unified syscall interface it actually
220: costs <10 bytes total to make all system calls thread safe!
221: I implemented atexit() (can register up to 4 callbacks).
222: I implemented strtod and __dtostr (the opposite). Now we can add
223: floating point support to vsnprintf and sscanf!
224:
225: 0.5.10:
226: I actually saw that I can not only merge the errno handling code of
227: the system calls on x86, I can also merge the rest (including the
228: arguments) except for setting the system call number. All those
229: system calls are now just a jump to a unified system call handler.
230: I got rid of x86openclose again. The savings are substantial:
231: chown from embutils went from 7664 to 7184 bytes!
232: If I'd move the system call wrappers to assembly language, I could
233: even reduce the alignment (does not matter since it's just a jump
234: anyway) and get rid of the "ret" behind each jump (one byte per
235: system call!)
236:
237: 0.5.9:
238: added memchr, strpbrk, strstr, strtol, isspace (hehe)
239: fixed strdup
240: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
241: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
242: (no signed integers, only strings and unsigned integers (but
243: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
244: that code must be wonderful, I don't understand it at all *bg*)
245: Olaf's printf does understand padding like in "%08d" and "%8d" and
246: automatically pads pointers with '0'.
247: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
248: Thanks to proton for inspiration on this.
249: The new x86 start code also does not reference exit any more, thus saving
250: 64 bytes for executables that don't call exit explicitly.
251: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
252: Moved errno and environ to start.S, saving no byte binary size but
253: speeding up compilation and slightly shortening dietlibc.a ;-)
254: For x86: joined open and close into one assembler file, sharing the
255: errno handling. This is a feasibility test and it actually saves a
256: few bytes. I think I will reimplement the _syscall[1-6] macros on
257: all platforms now to share the errno handling code for them.
258:
259: 0.5.8:
260: fixed strchr to be able to look for 0.
261:
262: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
263: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
264: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
265: contributed by Olaf Dreesen)
266:
267: added execvp, getcwd
268:
269: fixed __xmknod
270:
271: Note: can it be that the ftw interface really is so broken that I
272: cannot implement it without having to implement some searching data
273: structure? I included an experimental ftw implementation that will
274: not follow symlinks.
275:
276: 0.5.7:
277: oops, if_nametoindex was broken!
278:
279: 0.5.6:
280: ported to ppc-linux.
281: fixed i386 sigsetjmp (I mistyped the function name)
282: included sigjmp.c
283:
284: 0.5.5:
285: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
286: actually, it wasn't fork that wasn't working for sparc, it was pipe.
287: I fixed it now.
288: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
289: fixed readdir to use getdents and not the intel inline asm
290:
291: 0.5.4:
292: ported to sparclinux (sparc32 only). Beware: does not work yet.
293: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
294:
295: 0.5.3:
296: added tcsetattr and getenv to compile e3
297:
298: 0.5.2:
299: added vhangup, tcgetattr, isatty and memcmp for fgetty
300: (http://www.fefe.de/fgetty/)
301: added localtime from uC-libc.
302:
303: 0.5.1:
304: stat and friends actually work now. Yuck! Another case of
305: translation between kernel and userland. Why can't the kernel
306: people and the libc people simply agree on a standard?
307: removed many unnecessary includes to speed up compilation.
308: added -fomit-frame-pointer and i386 compilation target to reduce code size.
309: inlined socketcall to reduce code size.
310: "load" and "compile" are examples for djb code.
311:
312: 0.5:
313: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
314: each object file is now treated with "strip -x -R .note -R .comment"
315: split each system call into a separate object file
316: added ntohs, htons, alarm, if_indextoname and if_nametoindex
317:
318: 0.4.1:
319: oops, the strchr implementation was wrong. Thanks Jens Laas!
320:
321: 0.4:
322: copied opendir and friends from uC-libc.
323: added getservent and getservby* in gerservent.c
324:
325: 0.3:
326: remove readdir system call
327: added getdents system call
328: added getpwnam/getpwuid implementation in getpwnam.c
329: added a few string functions (in str*.c)
330:
LinuxTV legacy CVS <linuxtv.org/cvs>