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