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