]> git.kernelconcepts.de Git - karo-tx-redboot.git/blob - packages/language/c/libm/v2_0/ChangeLog
unified MX27, MX25, MX37 trees
[karo-tx-redboot.git] / packages / language / c / libm / v2_0 / ChangeLog
1 2007-09-04  Stephen Finney  <shf@pfinc.com>
2 2007-09-04  Jonathan Larmour  <jifl@eCosCentric.com>
3
4         * src/misc/infconst.c: Declare infinity byte order as dependent on
5         double byte order, not integer.
6         Fixes bug #1000448.
7
8 2004-04-14  Andrew Lunn  <andrew.lunn@ascom.ch>
9
10         * src/double/ieee754-core/e_cosh.c (__ieee754_cosh): 
11         * src/double/ieee754-core/e_sinh.c (__ieee754_sinh): Make use of
12         the CYG_LIBM_LO() macro which does not generate a compiler warning.
13
14 2003-02-24  Jonathan Larmour  <jifl@eCosCentric.com>
15
16         * cdl/libm.cdl: Fix doc link.
17
18 2001-09-28  Jesper Skov  <jskov@redhat.com>
19
20         * src/double/portable-api/s_expm1.c (expm1): Init c before it's used.
21
22 2001-07-20  Jonathan Larmour  <jlarmour@redhat.com>
23
24         * src/double/portable-api/s_expm1.c: Reimport from newlib
25         * src/double/ieee754-core/e_pow.c: Ditto.
26         * src/mathincl/fdlibm.h: Define macros and types required for
27         newlib versions of libm files.
28
29 2001-04-25  Bart Veer  <bartv@redhat.com>
30
31         * cdl/libm.cdl:
32         Ensure -ffloat-store is also used for the updated synthetic target
33         when running on x86 hardware.
34
35 2000-09-11  Jonathan Larmour  <jlarmour@redhat.com>
36
37         * src/misc/infconst.c: Revert below two changes - the gcc problem
38         that provoked it is now fixed.
39
40 2000-09-01  Jonathan Larmour  <jlarmour@redhat.com>
41
42         * src/misc/infconst.c: Silence warning about number of brackets
43
44 2000-08-31  Jonathan Larmour  <jlarmour@redhat.com>
45
46         * include/math.h: Declare cyg_libm_infinity as an array type for correct
47         relocations
48         * src/misc/infconst.c: Define it likewise
49
50 2000-05-30  Jesper Skov  <jskov@redhat.com>
51
52         * include/sys/ieeefp.h: Undo below change.
53
54 2000-05-26  Jesper Skov  <jskov@redhat.com>
55
56         * include/sys/ieeefp.h: Support FPU double-LE layout in BE mode.
57         Fix paste bug.
58
59 2000-05-02  Jonathan Larmour  <jlarmour@redhat.co.uk>
60
61         * tests/vectors/vector_support.h:
62         Update startup configury to reflect changes in libc startup. 
63         Use default stacksize of CYGNUM_HAL_STACK_SIZE_TYPICAL rather than 32K
64
65 2000-03-28  John Dallaway  <jld@cygnus.co.uk>
66
67         * cdl/libm.cdl:
68
69         Adjust documentation URLs.
70
71 2000-03-16  Jonathan Larmour  <jlarmour@redhat.co.uk>
72
73         * cdl/libm.cdl (CYGPKG_LIBM_CFLAGS_ADD): Add -ffloat-store for i386
74         targets. The math library implementation relies on accurate IEEE FP.
75
76 2000-01-31 Simon FitzMaurice  <sdf@cygnus.co.uk>
77   * cdl/*.cdl:
78
79   Adjust help URLs in line with new doc layout.
80   
81 2000-01-28 Simon FitzMaurice  <sdf@cygnus.co.uk>
82   * cdl/*.cdl:
83
84   Adjust help URLs in line with new doc layout.
85   
86 2000-01-28  Jonathan Larmour  <jlarmour@redhat.co.uk>
87
88         * cdl/libm.cdl: Don't need to compile with -O1 any more
89
90 2000-01-19  Hugo Tyson  <hmt@cygnus.co.uk>
91
92         * cdl/*.cdl: Add descriptions to a number of options &c which were
93         lacking same, also tidied up other typos as noticed en passant.
94
95 1999-11-23  Hugo Tyson  <hmt@cygnus.co.uk>
96
97         * include/pkgconf/libm.h: Cooperate with permtest combo10 setting
98         an intermediate variable CYGNUM_LIBM_COMPAT_DEFAULT directly.
99
100 1999-11-03  John Dallaway  <jld@cygnus.co.uk>
101
102         * cdl/libm.cdl: Define tests.
103
104 1999-10-18  John Dallaway  <jld@cygnus.co.uk>
105
106         * cdl/libm.cdl:
107         Parent CYGPKG_LIBM_CFLAGS_ADD and CYGPKG_LIBM_CFLAGS_REMOVE
108         under a CDL component.
109
110 1999-10-15  John Dallaway  <jld@cygnus.co.uk>
111
112         * cdl/libm.cdl:
113         Define CYGPKG_LIBM_CFLAGS_ADD and CYGPKG_LIBM_CFLAGS_REMOVE.
114
115 1999-10-07  John Dallaway  <jld@cygnus.co.uk>
116
117         * cdl/compat.cdl: Specify radio buttons using CDL interfaces.
118
119 1999-08-24  Jonathan Larmour  <jlarmour@cygnus.co.uk>
120
121         * src/PKGconf.mak: Optimize by default everywhere, except when compiling
122         e_pow.c, in which case use -fno-schedule-insns
123
124 1999-08-17  John Dallaway  <jld@cygnus.co.uk>
125
126         * cdl/compat.cdl:
127
128         Implement radio buttons using "FIXME radio" hack in
129         description field for now.
130
131 1999-08-14  John Dallaway  <jld@cygnus.co.uk>
132
133         * cdl/compat.cdl:
134         Change CYGNUM_LIBM_COMPATIBILITY flavor to 'booldata'.
135
136 1999-08-12  John Dallaway  <jld@cygnus.co.uk>
137
138         * cdl/libm.cdl, cdl/compat.cdl:
139         Take advantage of MACRONAME_ENUMVAL macros output by CDL.
140
141 1999-06-17  Jonathan Larmour  <jlarmour@cygnus.co.uk>
142
143         * src/PKGconf.mak: Disable optimization completely until CR 100802 is
144         fixed
145
146 1999-06-07  Jonathan Larmour  <jlarmour@cygnus.co.uk>
147
148         * src/mathincl/fdlibm.h: Fix for alias problem
149         Fix for PR 19698
150         * src/PKGconf.mak: workaround no longer needed
151
152 1999-06-04  Jonathan Larmour  <jlarmour@cygnus.co.uk>
153
154         * src/PKGconf.mak: 
155         Only use -fno-strict-aliasing if we have enough a recent toolchain
156         Default is that we have
157         
158 1999-06-03  Jonathan Larmour  <jlarmour@cygnus.co.uk>
159
160         * src/PKGconf.mak: 
161         Finally found the option to do the right workaround:
162         -fno-strict-aliasing
163
164 1999-06-02  Jonathan Larmour  <jlarmour@cygnus.co.uk>
165
166         * src/PKGconf.mak (SPECIAL_FLAGS): Don't use -fschedule-insns[2] on
167         mn10300
168
169 1999-05-27  Jonathan Larmour  <jlarmour@cygnus.co.uk>
170
171         * src/PKGconf.mak (SPECIAL_FLAGS): Add workaround flags for new
172         toolchains for all targets
173
174 1999-05-24  Jesper Skov  <jskov@cygnus.co.uk>
175
176         * src/PKGconf.mak: Same workaround for sh.
177
178 1999-05-19  Jonathan Larmour  <jlarmour@cygnus.co.uk>
179
180         * src/PKGconf.mak: Make below workaround be the other way around by
181         using -O1 and adding more explicit -O2 optimizations, just so we
182         don't lose very many optimizations. This is for sparclite
183
184 1999-05-19  Jonathan Larmour  <jlarmour@cygnus.co.uk>
185
186         * src/PKGconf.mak: Make below workaround be the other way around by
187         using -O1 and adding more explicit -O2 optimizations, just so we
188         don't lose very many optimizations. This is for ARM
189
190 1999-05-18  Jonathan Larmour  <jlarmour@cygnus.co.uk>
191
192         * src/PKGconf.mak: For the sparclite, prevent insn scheduling -
193         workaround for PR 19698
194
195 1999-05-18  Jonathan Larmour  <jlarmour@cygnus.co.uk>
196
197         * src/PKGconf.mak: For the ARM, change from -O1 to just prevent insn
198         scheduling - workaround for PR 19698
199
200 1999-04-15  John Dallaway  <jld@cygnus.co.uk>
201
202         * include/pkgconf/libm.h: Tidy display string capitalization.
203
204 1999-04-07  Jesper Skov  <jskov@cygnus.co.uk>
205         PR 19442
206         * tests/vectors/vector_support.h: Make an occasional chirp to keep
207         the testing infra interested.
208
209 1999-04-06  Gary Thomas  <gthomas@cygnus.co.uk>
210
211         * src/PKGconf.mak: More brute-force fix for ARM compiler
212         problems - PR 19772
213
214 1999-03-31  Jonathan Larmour  <jlarmour@cygnus.co.uk>
215
216         * include/pkgconf/libm.h: Remove unnecessary define
217         Fix for PR 19423
218
219 1999-03-29  Gary Thomas  <gthomas@cygnus.co.uk>
220
221         * src/PKGconf.mak: Add special compile options to work around ARM
222         compiler bugs.
223
224 1999-02-09  Jonathan Larmour  <jlarmour@cygnus.co.uk>
225
226         * src/mathincl/fdlibm.h: 
227         Change definition of CYG_LIBM_HI/CYG_LIBM_LO to use inline
228         function
229         Fix for PR 18859
230
231 1999-01-28  Jonathan Larmour  <jlarmour@cygnus.co.uk>
232
233         * src/misc/signgam.cxx: 
234         * src/misc/compatmode.cxx: 
235         Include <pkgconf/kernel.h> if we use kernel header files
236
237 1999-01-21  Jonathan Larmour  <jlarmour@cygnus.co.uk>
238
239         * tests/vectors/vector_support.h (doTestVec): Makes more
240         sense to output vector number than array index
241
242         * src/double/portable-api/s_ceil.c: Fix comment s/+inf/-inf/
243
244 1999-01-20  Jonathan Larmour  <jlarmour@cygnus.co.uk>
245
246         * tests/vectors/vector_support.h (doTestVec): 
247         Add output to let you know which vectors failed
248
249 1999-01-12  Gary Thomas  <gthomas@cygnus.co.uk>
250
251         * src/mathincl/fdlibm.h: 
252         * include/sys/ieeefp.h: Add different structure layouts for
253         machines that have doubles that aren't pure big/little endian.
254
255 Fri Oct 23 19:27:04 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
256
257         * tests/vectors/vector_support.h (START_TEST): 
258         Ensure that dummy START_TEST (the one used when the test should
259         *not* be run) does have a reference to the test function. And
260         make sure that there is a default one if there's no C API
261
262         * src/mathincl/fdlibm.h: 
263         Add matherr() prototype if no namespace pollution
264
265         Fixes for PR 17972
266
267 1998-10-22  Jesper Skov  <jskov@cygnus.co.uk>
268         PR 17925
269
270         * include/pkgconf/libm.h: 
271         * src/misc/matherr.c:
272         Removed the CYGFUN_LIBM_matherr config. The function is already
273         weakly defined so shouldn't need to be configured out.
274
275 1998-10-19  Mark Galassi  <rosalia@cygnus.com>
276
277         * include/pkgconf/libm.h: updated the doc URL
278
279 Thu Oct 15 21:36:12 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
280
281         * src/misc/compatmode.cxx, src/misc/signgam.cxx:
282         Only include kernel headers if we want thread-safe operation
283         Required for PR 17229
284
285 Wed Oct 14 17:24:46 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
286
287         * include/pkgconf/libm.h: 
288         Make CYGPKG_LIBM_COMPATIBILITY_DEFAULT require CYGPKG_LIBC
289         rather than all of libm, and put a check (with #error) that
290         this is the case
291
292         * src/double/ieee754-api/w_scalb.c:
293         Only include errno.h if we aren't IEEE-only
294
295         * src/double/portable-api/s_ldexp.c, src/misc/standard.c:
296         Make the error checking in s_ldexp.c more like every other libm
297         sources by putting the actual error handling into standard.c.
298         The error handling is just like scalb
299
300         * tests/vectors/acos.c, tests/vectors/asin.c,
301           tests/vectors/atan.c, tests/vectors/atan2.c,
302           tests/vectors/ceil.c, tests/vectors/cos.c, tests/vectors/cosh.c,
303           tests/vectors/exp.c, tests/vectors/fabs.c,
304           tests/vectors/floor.c, tests/vectors/fmod.c,
305           tests/vectors/frexp.c, tests/vectors/ldexp.c,
306           tests/vectors/log.c, tests/vectors/log10.c,
307           tests/vectors/modf.c, tests/vectors/pow.c, tests/vectors/sin.c,
308           tests/vectors/sinh.c, tests/vectors/sqrt.c, tests/vectors/tan.c,
309           tests/vectors/tanh.c:
310         Remove unnecessary include of <errno.h>
311
312         * tests/vectors/vector_support.h:
313         Only include errno.h if we aren't IEEE-only.
314         If we don't have Cyg_ErrNo, just use a cyg_int32 instead, to
315         make sure that the fields are still correct for the input
316         test vector.
317         In doTestVec(), before checking errno, check if we're
318         in IEEE mode (either compiled in or set at runtime) because if
319         so, and errno is set in the errno field of the input test
320         vector, then we have to skip this, since we can't validate the
321         output.
322         In all modes, if errno is set in the input test vector, don't
323         bother checking the actual values
324
325         All of the above changes are for PR 17503
326         
327         * tests/vectors/pow.c:
328         Remove requirement of being in POSIX mode, as the change to
329         vector_support.h should make this work better now.
330
331
332 Sun Sep 27 20:46:09 1998  David Moore  <dsm@keema.cygnus.co.uk>
333
334         * include/pkgconf/libm.h:
335         PR 17515: Corrected notcdl doc string.
336
337 1998-09-26  Bart Veer  <bartv@cygnus.co.uk>
338
339         * include/pkgconf/libm.h:
340         PR 17503: there are uses of errno in a couple of places in libm,
341         and errno will only be available if the C library is present.
342         Therefore libm requires libc.
343         
344
345 Fri Sep 25 19:00:35 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
346
347         * src/misc/compatmode.cxx: 
348         Make sure CYGNUM_LIBM_COMPATMODE_TRACE_LEVEL is defined before
349         we attempt tracing, otherwise set trace level to 0 i.e. never
350
351         * src/misc/signgam.cxx: 
352         Make sure CYGNUM_LIBM_SIGNGAM_TRACE_LEVEL is defined before
353         we attempt tracing, otherwise set trace level to 0 i.e. never
354         Fix for PR 17478
355
356         * include/pkgconf/libm.h: 
357         Define trace levels only if parent CYGPKG_LIBM_TRACE is defined
358         for consistency with Configuration Tool
359
360 Fri Sep 25 18:25:02 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
361
362         * include/pkgconf/libm.h: 
363         Only check valid compatibility setting if we aren't IEEE-only
364         mode
365         Fix for PR 17481
366
367 Fri Sep 25 18:13:24 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
368
369         * src/mathincl/fdlibm.h: 
370         * src/double/ieee754-core/e_scalb.c: 
371         * src/double/ieee754-api/w_scalb.c: 
372         Use CYGFUN_LIBM_SVID3_scalb for whether 2nd arg of scalb is
373         double or int, rather than _SCALB_INT.
374         Fix for PR 17479
375
376 Wed Sep 23 08:07:48 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
377
378         * tests/vectors/pow.c (test): 
379         Make pow test conditional on being in POSIX compatibility mode.
380         This is the only mode that can set errno correctly, in
381         relation to what the tests expect.
382         Fix for PR 16933
383
384 1998-09-20  Mark Galassi  <rosalia@cygnus.com>
385
386         * include/pkgconf/libm.h: updated CDL doc strings.
387
388 Thu Sep 17 17:10:47 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
389
390         * tests/vectors/vector_support.h (checkErrorAcceptable): 
391         Move check for signed-ness after all of the special cases
392         Fix for PR 16933
393
394 Tue Sep 15 19:26:32 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
395
396         * include/pkgconf/libm.h: 
397         Updated descriptions and a few other minor things after review
398
399 1998-09-11  Bart Veer  <bartv@cygnus.co.uk>
400
401         * include/pkgconf/libm.h:
402         Sorting out dependencies for error package.
403
404 Fri Sep  4 14:20:04 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
405
406         * include/pkgconf/libm.h: 
407         Fix thread-safe modes "requires" lines not to be associated
408         with the dummy parent package, but with the options themselves.
409         Also we don't need "requires CYGPKG_KERNEL" if we already require
410         a particular kernel option anyway.
411
412 Thu Sep  3 09:50:20 1998  Tim Goodwin  <tgoodwin@cygnus.co.uk>
413
414         * tests/vectors/vector_support.h
415         Change to cyg_thread_create() interface.
416
417 Wed Sep  2 14:41:54 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
418
419         * include/pkgconf/libm.h: 
420         Add default define for CYGPKG_LIBM_COMPATIBILITY_DEFAULT,
421         and shorten some display texts
422
423 Wed Sep  2 13:15:43 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
424
425         * src/double/ieee754-core/e_pow.c: 
426         Fix bug for when x is negative and y isn't an integer. A CYGNUS
427         LOCAL change got it wrong when casting a signed integer to
428         unsigned before right shifting.
429         
430         Reformat for 76 columns
431
432         * src/double/ieee754-api/w_pow.c: 
433         Reformat for 76 columns
434         
435
436 Wed Sep  2 12:30:11 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
437
438         * include/pkgconf/libm.h: 
439         Default tracing to off
440         Reformat for 76 columns
441
442         * tests/vectors/vector_support.h: 
443         Align '{'s and "verbs" on the same line. Tidy to 76 columns
444
445 Mon Aug 31 13:38:25 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
446
447         * include/pkgconf/libm.h: 
448         Require infra tracing support to allow tracing
449
450 1998-08-28  Bart Veer  <bartv@cygnus.co.uk>
451
452         * include/pkgconf/libm.h:
453         * src/misc/compatmode.cxx, src/misc/signgam.cxx:
454         * tests/vectors/vector_support.h
455
456         Updated for new kernel configuration option symbol names
457
458 Fri Aug 28 09:07:07 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
459
460         * tests/vectors/vector_support.h: 
461         Change START_TEST macro, adjust config options and add
462         cyg_package_start() to get tests working correctly with new entry
463         point mechanism, including cyg_iso_c_start() if available
464
465 Wed Aug 26 20:25:36 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
466
467         * include/pkgconf/libm.h: 
468         Remove #define CYGPKG_LIBM is that is meant to live in
469         <pkgconf/system.h> after all
470
471
472 Wed Aug 26 19:38:54 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
473
474         * include/pkgconf/libm.h: 
475         More fixes to CDL stuff to make it work better, and look better on
476         the screen
477
478 Tue Aug 25 20:23:25 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
479
480         * include/pkgconf/libm.h: 
481         Make some fixes after running the ConfigTool
482
483 Tue Aug 25 16:59:11 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
484
485         * include/pkgconf/libm.h: 
486         Turns out the nesting wasn't quite right. Correct it, and replace
487         the "not" prefix on CFG_DATA, which slipped in to the previous change.
488
489 Tue Aug 25 16:53:03 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
490
491         * include/pkgconf/libm.h: 
492         Ensure there is a default compatibility mode and fix the
493         ifdef nesting for the compat modes
494
495 Tue Aug 25 15:24:51 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
496
497         * include/pkgconf/libm.h: 
498         Tidy up CDL stuff prior to checking if it all actually works!
499
500 Tue Aug 25 05:17:55 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
501
502         * include/pkgconf/libm.h: 
503         Do a first pass of CDLification of configuration data
504
505         * include/pkgconf/libm.h, src/misc/matherr.c: 
506         Rename CYGFUN_LIBC_matherr to CYGFUN_LIBM_matherr
507
508 Wed Aug 19 19:35:16 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
509
510         * tests/vectors/acos.c, tests/vectors/cosh.c, tests/vectors/ldexp.c,
511           tests/vectors/sinh.c, tests/vectors/asin.c, tests/vectors/exp.c,
512           tests/vectors/log.c, tests/vectors/sqrt.c, tests/vectors/atan.c,
513           tests/vectors/fabs.c, tests/vectors/log10.c, tests/vectors/tan.c,
514           tests/vectors/atan2.c, tests/vectors/floor.c, tests/vectors/modf.c,
515           tests/vectors/tanh.c, tests/vectors/ceil.c, tests/vectors/fmod.c,
516           tests/vectors/pow.c, tests/vectors/cos.c, tests/vectors/frexp.c,
517           tests/vectors/sin.c
518
519         * tests/vectors/vector_support.h (doTestVec): 
520         Remove "register" on variable
521
522 Wed Aug 19 18:59:47 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
523
524         * src/double/ieee754-core/e_gamma.c,
525           src/double/ieee754-core/e_lgamma.c,
526           src/double/ieee754-api/w_gamma.c,
527           src/double/ieee754-api/w_lgamma.c:
528         Remove direct reference to extern int signgam and include <math.h>
529         instead
530
531 Mon Jul 27 23:47:52 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
532
533         * include/math.h, include/pkgconf/libm.h,
534           src/double/ieee754-api/w_acos.c, src/double/ieee754-api/w_acosh.c,
535           src/double/ieee754-api/w_asin.c, src/double/ieee754-api/w_atan2.c,
536           src/double/ieee754-api/w_atanh.c, src/double/ieee754-api/w_cosh.c,
537           src/double/ieee754-api/w_exp.c, src/double/ieee754-api/w_fmod.c,
538           src/double/ieee754-api/w_gamma.c,
539           src/double/ieee754-api/w_gamma_r.c,
540           src/double/ieee754-api/w_hypot.c, src/double/ieee754-api/w_j0.c,
541           src/double/ieee754-api/w_j1.c, src/double/ieee754-api/w_jn.c,
542           src/double/ieee754-api/w_lgamma.c,
543           src/double/ieee754-api/w_lgamma_r.c,
544           src/double/ieee754-api/w_log.c,
545           src/double/ieee754-api/w_log10.c, src/double/ieee754-api/w_pow.c,
546           src/double/ieee754-api/w_remainder.c,
547           src/double/ieee754-api/w_scalb.c, src/double/ieee754-api/w_sinh.c,
548           src/double/ieee754-api/w_sqrt.c, src/double/ieee754-core/e_acos.c,
549           src/double/ieee754-core/e_acosh.c,
550           src/double/ieee754-core/e_asin.c,
551           src/double/ieee754-core/e_atan2.c,
552           src/double/ieee754-core/e_atanh.c,
553           src/double/ieee754-core/e_cosh.c,
554           src/double/ieee754-core/e_exp.c, src/double/ieee754-core/e_fmod.c,
555           src/double/ieee754-core/e_gamma.c,
556           src/double/ieee754-core/e_gamma_r.c,
557           src/double/ieee754-core/e_hypot.c, src/double/ieee754-core/e_j0.c,
558           src/double/ieee754-core/e_j1.c, src/double/ieee754-core/e_jn.c,
559           src/double/ieee754-core/e_lgamma.c,
560           src/double/ieee754-core/e_lgamma_r.c,
561           src/double/ieee754-core/e_log.c, src/double/ieee754-core/e_log10.c,
562           src/double/ieee754-core/e_pow.c,
563           src/double/ieee754-core/e_rem_pio2.c,
564           src/double/ieee754-core/e_remainder.c,
565           src/double/ieee754-core/e_scalb.c,
566           src/double/ieee754-core/e_sinh.c,
567           src/double/ieee754-core/e_sqrt.c, src/double/internal/k_cos.c,
568           src/double/internal/k_rem_pio2.c, src/double/internal/k_sin.c,
569           src/double/internal/k_tan.c, src/double/portable-api/s_asinh.c,
570           src/double/portable-api/s_atan.c, src/double/portable-api/s_cbrt.c,
571           src/double/portable-api/s_ceil.c,
572           src/double/portable-api/s_copysign.c,
573           src/double/portable-api/s_cos.c, src/double/portable-api/s_erf.c,
574           src/double/portable-api/s_expm1.c,
575           src/double/portable-api/s_fabs.c,
576           src/double/portable-api/s_finite.c,
577           src/double/portable-api/s_floor.c,
578           src/double/portable-api/s_frexp.c,
579           src/double/portable-api/s_ilogb.c,
580           src/double/portable-api/s_isnan.c,
581           src/double/portable-api/s_ldexp.c,
582           src/double/portable-api/s_log1p.c,
583           src/double/portable-api/s_logb.c,
584           src/double/portable-api/s_modf.c,
585           src/double/portable-api/s_nextafter.c,
586           src/double/portable-api/s_rint.c,
587           src/double/portable-api/s_scalbn.c,
588           src/double/portable-api/s_significand.c,
589           src/double/portable-api/s_sin.c, src/double/portable-api/s_tan.c,
590           src/double/portable-api/s_tanh.c, src/mathincl/fdlibm.h,
591           src/misc/standard.c:
592
593         Untabify
594
595 Mon Jul 27 22:52:24 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
596
597         * tests/vectors/vector_support.h (doTestVec): 
598         Remove "register" when declaring variable i
599
600         * host/acosgen.c, host/cosgen.c, host/fmodgen.c, host/modfgen.c,
601           host/tangen.c, host/asingen.c, host/coshgen.c, host/frexpgen.c,
602           host/powgen.c, host/tanhgen.c, host/atan2gen.c, host/expgen.c,
603           host/ldexpgen.c, host/singen.c, host/atangen.c, host/fabsgen.c,
604           host/log10gen.c, host/sinhgen.c, host/ceilgen.c, host/floorgen.c,
605           host/loggen.c, host/sqrtgen.c:
606         Remove "register" when declaring variable "int x"
607
608 Mon Jul 27 22:45:01 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
609
610         * tests/vectors/vector_support.h (checkErrorAcceptable): 
611         Add extra check for tolerance acceptance so it behaves correctly with
612         underflow and overflow
613
614         Also reverse sense of checkErrorAcceptable() so it now returns
615         false on success and true on failure, as suggested by hmt. Change
616         this in doTestVec() (which is where it is called) too.
617
618         * tests/vectors/vector_support.h: 
619         Remove #if 0'd shift_double() function
620
621 Mon Jul 27 21:55:04 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
622
623         * src/mathincl/fdlibm.h: 
624         Rename CYGONCE_LIBM_FDLIBM_H to CYGONCE_LIBM_MATHINCL_FDLIBM_H as per
625         code review comments
626
627 Fri Jul 24 22:20:42 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
628
629         * include/math.h, include/pkgconf/libm.h,
630           src/double/ieee754-api/w_acos.c, src/double/ieee754-api/w_acosh.c,
631           src/double/ieee754-api/w_asin.c, src/double/ieee754-api/w_atan2.c,
632           src/double/ieee754-api/w_atanh.c, src/double/ieee754-api/w_cosh.c,
633           src/double/ieee754-api/w_exp.c, src/double/ieee754-api/w_fmod.c,
634           src/double/ieee754-api/w_gamma.c, src/double/ieee754-api/w_gamma_r.c,
635           src/double/ieee754-api/w_hypot.c, src/double/ieee754-api/w_j0.c,
636           src/double/ieee754-api/w_j1.c, src/double/ieee754-api/w_jn.c,
637           src/double/ieee754-api/w_lgamma.c,
638           src/double/ieee754-api/w_lgamma_r.c, src/double/ieee754-api/w_log.c,
639           src/double/ieee754-api/w_log10.c, src/double/ieee754-api/w_pow.c,
640           src/double/ieee754-api/w_remainder.c,
641           src/double/ieee754-api/w_scalb.c, src/double/ieee754-api/w_sinh.c,
642           src/double/ieee754-api/w_sqrt.c, src/mathincl/fdlibm.h,
643           src/misc/compatmode.cxx, src/misc/matherr.c, src/misc/signgam.cxx,
644           src/misc/standard.c:
645
646         Rename libm config option names as per the libm coding standard
647         review. Chiefly ensuring that CYGXXX_ prefixes are correct
648
649         * src/misc/matherr.c, include/pkgconf/libm.h:
650         Add new CYGFUN_LIBC_matherr config option to prevent inclusion of
651         default matherr
652
653 Mon Jul 20 18:07:31 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
654
655         * src/misc/infconst.c: 
656         Oops. Don't let cyg_libm_infinity be static!
657
658 Thu Jul 16 20:07:26 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
659
660         * include/math.h, include/sys/ieeefp.h, src/misc/infconst.c:
661         Remove union Cyg_libm_int_double from math.h, which was used to
662         define infinity, and replace with use of
663         Cyg_libm_ieee_double_shape_type from ieeefp.h
664
665         That type (which is a union) is now reordered so that it can be
666         initialised in terms of two int's
667
668 Mon Jul 13 21:55:06 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
669
670         * src/double/ieee754-api/w_acos.c, src/double/ieee754-api/w_acosh.c,
671           src/double/ieee754-api/w_asin.c, src/double/ieee754-api/w_atan2.c,
672           src/double/ieee754-api/w_atanh.c, src/double/ieee754-api/w_cosh.c,
673           src/double/ieee754-api/w_exp.c, src/double/ieee754-api/w_fmod.c,
674           src/double/ieee754-api/w_gamma.c, src/double/ieee754-api/w_gamma_r.c,
675           src/double/ieee754-api/w_hypot.c, src/double/ieee754-api/w_j0.c,
676           src/double/ieee754-api/w_j1.c, src/double/ieee754-api/w_jn.c,
677           src/double/ieee754-api/w_lgamma.c,
678           src/double/ieee754-api/w_lgamma_r.c, src/double/ieee754-api/w_log.c,
679           src/double/ieee754-api/w_log10.c, src/double/ieee754-api/w_pow.c,
680           src/double/ieee754-api/w_remainder.c,
681           src/double/ieee754-api/w_scalb.c, src/double/ieee754-api/w_sinh.c,
682           src/double/ieee754-api/w_sqrt.c, src/double/ieee754-core/e_acos.c,
683           src/double/ieee754-core/e_acosh.c, src/double/ieee754-core/e_asin.c,
684           src/double/ieee754-core/e_atan2.c, src/double/ieee754-core/e_atanh.c,
685           src/double/ieee754-core/e_cosh.c, src/double/ieee754-core/e_exp.c,
686           src/double/ieee754-core/e_fmod.c, src/double/ieee754-core/e_gamma.c,
687           src/double/ieee754-core/e_gamma_r.c,
688           src/double/ieee754-core/e_hypot.c, src/double/ieee754-core/e_j0.c,
689           src/double/ieee754-core/e_j1.c, src/double/ieee754-core/e_jn.c,
690           src/double/ieee754-core/e_lgamma.c,
691           src/double/ieee754-core/e_lgamma_r.c,
692           src/double/ieee754-core/e_log.c, src/double/ieee754-core/e_log10.c,
693           src/double/ieee754-core/e_pow.c,
694           src/double/ieee754-core/e_rem_pio2.c,
695           src/double/ieee754-core/e_remainder.c,
696           src/double/ieee754-core/e_scalb.c, src/double/ieee754-core/e_sinh.c,
697           src/double/ieee754-core/e_sqrt.c, src/double/internal/k_cos.c, 
698           src/double/internal/k_rem_pio2.c, src/double/internal/k_sin.c,
699           src/double/internal/k_tan.c, src/double/portable-api/s_asinh.c,
700           src/double/portable-api/s_atan.c, src/double/portable-api/s_cbrt.c,
701           src/double/portable-api/s_ceil.c,
702           src/double/portable-api/s_copysign.c,
703           src/double/portable-api/s_cos.c, src/double/portable-api/s_erf.c,
704           src/double/portable-api/s_expm1.c, src/double/portable-api/s_fabs.c,
705           src/double/portable-api/s_finite.c,
706           src/double/portable-api/s_floor.c, src/double/portable-api/s_frexp.c,
707           src/double/portable-api/s_ilogb.c, src/double/portable-api/s_isnan.c,
708           src/double/portable-api/s_ldexp.c, src/double/portable-api/s_log1p.c,
709           src/double/portable-api/s_logb.c, src/double/portable-api/s_modf.c,
710           src/double/portable-api/s_nextafter.c,
711           src/double/portable-api/s_rint.c, src/double/portable-api/s_scalbn.c,
712           src/double/portable-api/s_significand.c,
713           src/double/portable-api/s_sin.c, src/double/portable-api/s_tan.c,
714           src/double/portable-api/s_tanh.c, src/misc/standard.c:
715
716         Make #ifdef __STDC__ bits compulsory and remove the associated #else
717         clause. We know that gcc is ISO C compatible!
718
719 Tue Jul  7 16:05:48 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
720
721         * src/misc/compatmode.cxx:
722         Fix conditionalisation to prevent unused variable warning for tracing
723
724 Thu Jun 25 04:58:21 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
725
726         * tests/vectors/acos.c, tests/vectors/acos.h, tests/vectors/asin.c,
727           tests/vectors/asin.h, tests/vectors/atan.c, tests/vectors/atan.h,
728           tests/vectors/atan2.c, tests/vectors/atan2.h, tests/vectors/ceil.c,
729           tests/vectors/ceil.h, tests/vectors/cos.c, tests/vectors/cos.h,
730           tests/vectors/cosh.c, tests/vectors/cosh.h, tests/vectors/exp.c,
731           tests/vectors/exp.h, tests/vectors/fabs.c, tests/vectors/fabs.h,
732           tests/vectors/floor.c, tests/vectors/floor.h, tests/vectors/fmod.c,
733           tests/vectors/fmod.h, tests/vectors/frexp.c, tests/vectors/frexp.h,
734           tests/vectors/ldexp.c, tests/vectors/ldexp.h, tests/vectors/log.c,
735           tests/vectors/log.h, tests/vectors/log10.c, tests/vectors/log10.h,
736           tests/vectors/modf.c, tests/vectors/modf.h, tests/vectors/pow.c,
737           tests/vectors/pow.h, tests/vectors/sin.c, tests/vectors/sin.h,
738           tests/vectors/sinh.c, tests/vectors/sinh.h, tests/vectors/sqrt.c,
739           tests/vectors/sqrt.h, tests/vectors/tan.c, tests/vectors/tan.h,
740           tests/vectors/tanh.c, tests/vectors/tanh.h,
741           tests/vectors/vector_support.h:
742         Add math library tests for functions specified in the ISO standard.
743         These are based on precomputed test vectors
744
745         * tests/PKGconf.mak:
746         Create with above files being built
747
748         * host/acosgen.c, host/asingen.c, host/atan2gen.c, host/atangen.c,
749           host/ceilgen.c, host/cosgen.c, host/coshgen.c, host/expgen.c,
750           host/fabsgen.c, host/floorgen.c, host/fmodgen.c, host/frexpgen.c,
751           host/ldexpgen.c, host/log10gen.c, host/loggen.c, host/modfgen.c,
752           host/powgen.c, host/singen.c, host/sinhgen.c, host/sqrtgen.c,
753           host/tangen.c, host/tanhgen.c:
754         Host-side (and currently linux-specific) sources to automatically
755         generate test vector data used by the above tests.
756
757         * include/pkgconf/libm.h:
758         Ensure the kernel C API is available if we want thread safety
759
760 Wed Jun 24 15:42:31 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
761
762         * src/misc/standard.c:
763         Fix typo in ifdef
764
765 Wed Jun 24 13:04:30 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
766
767         * src/misc/standard.c:
768         Fix #defines for IEEE mode, revealed by permutation testing
769
770 Mon Jun 22 14:42:50 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
771
772         * src/misc/infconst.c:
773         Fix compiler warning
774
775 Sun Jun 21 04:42:18 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
776
777         * include/sys/ieeefp.h:
778         Add this to allow bitfield break down of IEEE floating point doubles
779
780 Wed Jun 17 08:55:21 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
781
782         * include/pkgconf/libm.h:
783         Make the use of stderr for messages _not_ the default
784         Make thread safety the default
785
786         * src/misc/standard.c:
787         Change alternative for fprintf(stderr,...) _not_ being used from
788         "CYG_EMPTY_STATEMENT" to just "0" to placate compiler.
789
790 Mon Jun 15 19:59:36 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
791
792         * src/misc/compatmode.cxx:
793         Fix typo in compatibility mode name
794
795 Fri Jun  5 16:31:26 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
796
797         * src/misc/signgam.cxx, src/misc/compatmode.cxx:
798         Fix some minor typos when thread safety is turned on
799
800         * src/misc/matherr.c
801         Correct syntax of weak attribute
802
803 Fri Jun  5 07:07:03 1998  Jonathan Larmour  <jlarmour@cygnus.co.uk>
804
805         * Initial check-in of all math library sources. Based on fdlibm-5.2.
806
807
808 //===========================================================================
809 //####ECOSGPLCOPYRIGHTBEGIN####
810 // -------------------------------------------
811 // This file is part of eCos, the Embedded Configurable Operating System.
812 // Copyright (C) 1998, 1999, 2000, 2001, 2002 Red Hat, Inc.
813 //
814 // eCos is free software; you can redistribute it and/or modify it under
815 // the terms of the GNU General Public License as published by the Free
816 // Software Foundation; either version 2 or (at your option) any later version.
817 //
818 // eCos is distributed in the hope that it will be useful, but WITHOUT ANY
819 // WARRANTY; without even the implied warranty of MERCHANTABILITY or
820 // FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
821 // for more details.
822 //
823 // You should have received a copy of the GNU General Public License along
824 // with eCos; if not, write to the Free Software Foundation, Inc.,
825 // 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
826 //
827 // As a special exception, if other files instantiate templates or use macros
828 // or inline functions from this file, or you compile this file and link it
829 // with other works to produce a work based on this file, this file does not
830 // by itself cause the resulting work to be covered by the GNU General Public
831 // License. However the source code for this file must still be made available
832 // in accordance with section (3) of the GNU General Public License.
833 //
834 // This exception does not invalidate any other reasons why a work based on
835 // this file might be covered by the GNU General Public License.
836 //
837 // Alternative licenses for eCos may be arranged by contacting Red Hat, Inc.
838 // at http://sources.redhat.com/ecos/ecos-license/
839 // -------------------------------------------
840 //####ECOSGPLCOPYRIGHTEND####
841 //===========================================================================