Annotation of dietlibc/CHANGES, revision 1.24
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.
1.21 fefe 12:
1.9 fefe 13: 0.7.1:
1.11 fefe 14: optimized away the static arrays from __ltostr and __lltostr.
15: optimized away the static arrays (and more!) from strtoul and strtoull.
16: added system (by Olaf).
1.12 fefe 17: added tolower and toupper
1.13 fefe 18: added getopt from Olaf (~600 bytes on x86!)
1.14 fefe 19: added clone for mips (by Olaf).
20: fixed longjmp and setjmp on sparc not to require glibc include files.
1.15 olaf 21: added pause for all (by Olaf).
1.17 fefe 22: added __attribute__ ((__const__)) to ctype.h
1.16 fefe 23: added isupper and islower.
24: added putc and putchar as macros.
1.17 fefe 25: added strcasecmp and strncasecmp (thanks to Abraham vd Merwe for
26: insightful discussions and patches I partially used for the last three)
1.18 fefe 27: added getpriority and setpriority (also by Abraham)
1.19 fefe 28: oops, I broke strtoul in a recent optimization attempt.
1.20 fefe 29: oi, setjmp and longjmp were broken on other platforms, too!
30: Thanks to Sébastien Côté for reporting this.
1.9 fefe 31:
32: 0.7:
1.6 olaf 33: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 34: Initial dynamic loading glue for i386, also by Olaf.
35: We don't have a dynamic loader yet, so this is not yet usable.
36: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
37: vsprintf, writev, all by Olaf.
1.2 fefe 38: imported readdir fix from Guillaume. Mhh, can the kernel interface
39: really be this broken?
1.3 fefe 40: Found a kludgy way to detect whether the kernel includes define struct
41: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
42: not work for you, please tell me!
1.4 fefe 43: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 44: Fixed stdio line buffering. The code was already there but I forgot
45: to mark stdio and stdout as line buffered. This should really be
46: done dynamically using isatty, though.
1.8 fefe 47: Removed superfluous "if (1)" in execvp.c
1.1 cvs 48:
49: 0.6.14:
50: *scanf did not append the 0 byte for %s.
51: added creat and changed creat to call open directly instead of open64.
52: did for fread what 0.6.10 did for fwrite.
53: repaired buffered stdio a little (line buffering is still not
54: supported).
55: [All of these bugs have been reported by Guillaume. Thanks!]
56: added __pure__ to strchr and strrchr in <string.h>
57:
58: 0.6.13:
59: included COPYING to make clear that the diet libc is covered by the
60: GNU General Public License (and _not_ the LGPL). That means that
61: you need to obtain a license from me if you want to use the diet
62: libc in a proprietary program that you want to distribute.
63: The x86 unified syscall swapped arguments #4 and #5 and noone
64: noticed until now! Thanks to Guillaume Cottenceau for reporting
65: this!
66: The malloc had a subtle bug with small allocations that could cause
67: segfault. Guillaume reported it, Olaf fixed it.
68:
69: 0.6.12:
70: added assert, statfs and fstatfs
71: added include/net/if.h so fget compiles again.
72:
73: 0.6.11:
74: added putenv.
75:
76: 0.6.10:
77: fixed fwrite. It returned the number of bytes written, not the number
78: of records. Thanks to Albert D. Cahalan for pointing this out.
79: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
80: added brk and sbrk. harold@nb.com.sg made me do it.
81:
82: 0.6.9:
83: fixed execvp not to return on ENOENT
84: did for {open|read|close|seek|tell}dir what I did for stat before.
85: The glibc compatibility part is currently non-functional.
86: fixed white space handling in sscanf " (" would not match " (".
87: fixed printf to accept 'l' flag.
88: fixed getpwuid (ignored last line)
89:
90: 0.6.8:
91: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
92: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
93: added strlcpy and strlcat from OpenBSD.
94: added stat64, fstat64 and lstat64.
95: added endian.h and made ending of strlen.c endianness-aware
96: added string routines size tweak to dietfeatures.h. Most of the
97: string routines have been unrolled. This is up to three times as
98: fast but creates up to three times larger code. Now you can disable
99: the unrolling.
100: added sendfile.
101: I also added pread. It works on x86 and sparc but not on ppc and mips.
102: I have no idea what's going on. strace disagrees with the kernel.
103: Can anyone help? I asked the Linux kernel mailing list for help, too.
104: added a few aliases of the type __libc_open for open. nm on
105: libpthreads.so indicates that we will need them.
106:
107: 0.6.7:
108: the sources now compile without warnings with -Wall.
109: printf now prints "(null)" when %s is passed NULL. This can be
110: removed with WANT_NULL_PRINTF in dietfeatures.h
111: added vfprintf, execl, ttyname.
112: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
113: __dietfstat so they can use the normal kernel struct stat.
114: stat, lstat and fstat are now C wrappers that convert to the glibc
115: struct stat. So we can avoid that overhead for programs that use
116: dietlibc headers.
117:
118: 0.6.6:
119: changed the Makefiles so you can now set CFLAGS on the command line
120: added memccmp (analogous to memccpy) and strncmp.
121: Olaf fixed another bug in the sparc unified syscall.
122: Paul Clifford contributed a C version of his strlen.S that is much
123: more efficient than the previous strlen on all platforms! It also
124: looks like technology from Roswell. ;-}
125: added a "real" stdio. Well, almost. Please contribute!
126: added dietfeatures.h so you can remove features you don't need.
127: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
128: stdio now works with simple programs and minigzip from libz.
129: stdio uses some major trickery to avoid linking the stdio, stderr and
130: stdout (and reserving space for them) when they are not used.
131:
132: 0.6.5:
133: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
134: most of them bugs in the new header files, but also missing sigset
135: functions.
136: Paul Clifford contributed an assembly strlen.S for ARM and fixed
137: several bugs.
138: Fixed ppc/setjmp and mips/pipe.
139: Olaf contributed a new sparc unified syscall.
140:
141: 0.6.4:
142: declared is* static inline in getservent.
143: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
144: changed strlen to return 0 when passed NULL.
145: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
146: Scientist who actually implemented this without access to a MIPS
147: box, just from reading the architecture manual. And it worked out
148: of the box.
149: started a set of system includes, mainly so that I can use lcc and
150: my alpha-linux cross compiler (which is unable to cross-compile
151: glibc). diet libc can now be compiled without any include files
152: from a normal libc. The includes do declare more than diet libc
153: currently delivers and they are still far from complete for real
154: applications.
155: I will only add assembly versions that are smaller _and_ faster than
156: the C version. Larger routines are only accepted if they are called
157: very often and are substantially faster.
158: added i386 assembly strchr, which is smaller and faster than the old
159: version. It is, however, larger than the version contributed by
160: proton (thanks, anyway).
161: added i386 strlen (31% faster, 14% smaller)
162: "make t" will now create a map file called "mapfile".
163:
164: 0.6.3:
165: added sys_errlist, strerror and perror
166: added isblank
167: added atol
168:
169: 0.6.2:
170: mmap for ARM didn't compile. Thanks, Paul!
171:
172: 0.6.1:
173: split mmap into the architecture specific subdirectories.
174: getenv now copes with environ==NULL, thanks Paul Clifford.
175: Paul also contributed a smaller ARM startup code.
176:
177: 0.6:
178: strcat returned the wrong result. Thanks, Dietz Pröpper.
179: strtod now understands a negative exponent (oops, thanks Bertram Barth)
180: Port to arm-linux-gnu, but on the Netwinder I use for testing the
181: __dtostr does not work (I have no idea why!)
182: The MIPS port now uses (much smaller) non-PIC code. That means that
183: the applications you link against diet libc also have to be compiled
184: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
185: Thanks to Ralf Bächle for helping me with this!
186: I made subdirectories for the architectures and use VPATH to override
187: VPATH so that make finds the source file automatically. That should
188: simplify the sources greatly.
189: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
190: ARM and MIPS now also have setjmp and longjmp
191: "compile" and "load" are now make targets. Use them for djb programs.
192: added contrib/elftrunc.c which will remove unnecessary ELF headers.
193: Again, contributed from Olaf. Great work!
194:
195: 0.5.12:
196: printf also does signed numbers.
197: If you don't use atexit, dietlibc now does not link exit, only _exit.
198: *printf now correctly returns the number of bytes written.
199: Olaf contributed experimental sscanf and vsscanf implementations.
200: If passed NULL as buffer, snprintf will not write anything but still
201: return the number of bytes it would have written.
202: Initial MIPS port! (Oh, the agony!)
203: No setjmp and longjmp support yet! Please contribute!
204: I even implemented unified syscalls for MIPS. Still, MIPS code is
205: almost twice the size of SPARC code. If anyone knows why: please
206: tell me!
207:
208: 0.5.11:
209: I implemented new unified syscalls for x86, sparc and ppc, this time based
210: on .s files and not .c files, and I moved the syscalls into
211: subdirectores to clean the dietlibc sources up a little.
212: A binary consisting of printf("%s is %d\n","olaf",23) is now
213: 2864 bytes on sparc
214: 2488 bytes on intel
215: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
216: to make a few more applications work.
217: Olaf Dreesen also implemented some unified syscall stuff for x86.
218: He discovered that with our unified syscall interface it actually
219: costs <10 bytes total to make all system calls thread safe!
220: I implemented atexit() (can register up to 4 callbacks).
221: I implemented strtod and __dtostr (the opposite). Now we can add
222: floating point support to vsnprintf and sscanf!
223:
224: 0.5.10:
225: I actually saw that I can not only merge the errno handling code of
226: the system calls on x86, I can also merge the rest (including the
227: arguments) except for setting the system call number. All those
228: system calls are now just a jump to a unified system call handler.
229: I got rid of x86openclose again. The savings are substantial:
230: chown from embutils went from 7664 to 7184 bytes!
231: If I'd move the system call wrappers to assembly language, I could
232: even reduce the alignment (does not matter since it's just a jump
233: anyway) and get rid of the "ret" behind each jump (one byte per
234: system call!)
235:
236: 0.5.9:
237: added memchr, strpbrk, strstr, strtol, isspace (hehe)
238: fixed strdup
239: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
240: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
241: (no signed integers, only strings and unsigned integers (but
242: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
243: that code must be wonderful, I don't understand it at all *bg*)
244: Olaf's printf does understand padding like in "%08d" and "%8d" and
245: automatically pads pointers with '0'.
246: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
247: Thanks to proton for inspiration on this.
248: The new x86 start code also does not reference exit any more, thus saving
249: 64 bytes for executables that don't call exit explicitly.
250: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
251: Moved errno and environ to start.S, saving no byte binary size but
252: speeding up compilation and slightly shortening dietlibc.a ;-)
253: For x86: joined open and close into one assembler file, sharing the
254: errno handling. This is a feasibility test and it actually saves a
255: few bytes. I think I will reimplement the _syscall[1-6] macros on
256: all platforms now to share the errno handling code for them.
257:
258: 0.5.8:
259: fixed strchr to be able to look for 0.
260:
261: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
262: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
263: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
264: contributed by Olaf Dreesen)
265:
266: added execvp, getcwd
267:
268: fixed __xmknod
269:
270: Note: can it be that the ftw interface really is so broken that I
271: cannot implement it without having to implement some searching data
272: structure? I included an experimental ftw implementation that will
273: not follow symlinks.
274:
275: 0.5.7:
276: oops, if_nametoindex was broken!
277:
278: 0.5.6:
279: ported to ppc-linux.
280: fixed i386 sigsetjmp (I mistyped the function name)
281: included sigjmp.c
282:
283: 0.5.5:
284: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
285: actually, it wasn't fork that wasn't working for sparc, it was pipe.
286: I fixed it now.
287: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
288: fixed readdir to use getdents and not the intel inline asm
289:
290: 0.5.4:
291: ported to sparclinux (sparc32 only). Beware: does not work yet.
292: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
293:
294: 0.5.3:
295: added tcsetattr and getenv to compile e3
296:
297: 0.5.2:
298: added vhangup, tcgetattr, isatty and memcmp for fgetty
299: (http://www.fefe.de/fgetty/)
300: added localtime from uC-libc.
301:
302: 0.5.1:
303: stat and friends actually work now. Yuck! Another case of
304: translation between kernel and userland. Why can't the kernel
305: people and the libc people simply agree on a standard?
306: removed many unnecessary includes to speed up compilation.
307: added -fomit-frame-pointer and i386 compilation target to reduce code size.
308: inlined socketcall to reduce code size.
309: "load" and "compile" are examples for djb code.
310:
311: 0.5:
312: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
313: each object file is now treated with "strip -x -R .note -R .comment"
314: split each system call into a separate object file
315: added ntohs, htons, alarm, if_indextoname and if_nametoindex
316:
317: 0.4.1:
318: oops, the strchr implementation was wrong. Thanks Jens Laas!
319:
320: 0.4:
321: copied opendir and friends from uC-libc.
322: added getservent and getservby* in gerservent.c
323:
324: 0.3:
325: remove readdir system call
326: added getdents system call
327: added getpwnam/getpwuid implementation in getpwnam.c
328: added a few string functions (in str*.c)
329:
LinuxTV legacy CVS <linuxtv.org/cvs>