openSUSE Build Service

Overview Repositories Revisions Requests Users Attributes Meta

File libdfp.changes of Package libdfp (Revision 4d5d954c618d036dccdbde52b8c34aba)

Currently displaying revision 4d5d954c618d036dccdbde52b8c34aba , Show latest

-------------------------------------------------------------------
Mon Sep 14 13:04:22 UTC 2020 - matz@suse.com

- Upgrade to version 1.0.15 [jsc#SLE-13839, jsc#SLE-14885, jsc#IBM-704,
  jsc#SLE-15622, jsc#SLE-14722, jsc#IBM-703, jsc#ECO-2376]
  * Add support for arch13/z15 s390 CPUs
  * Integrate libdecnumber into build
  * Fix spurious integer overflow warnings
  * sin/cos: more correctly compute the requested value
  * Fix and test ldexp
  * Add native expd128, logd128
  * Add fminmagN function as TS18661 recommends
- Enable build for x86_64
- Fix _smp_mflags rpm macro typo

-------------------------------------------------------------------
Fri Jan 17 14:00:57 UTC 2020 - matz@suse.com

- Upgrade to version 1.0.14 [jsc#SLE-7641 aka jsc#SLE-7401]
  Since 1.0.13:
  * enable POWER9 build
  * add support for arch12/z14 s390 CPUs
  * add hardware implementations on s390 for:
    islessgreaterd32|64|128, isunorderedd32|64|128, islessequald32|64|128,
    islessd32|64|128, isgreaterequald32|64|128, isgreaterd32|64|128,
    fdimd32|64|128, fmaxd32|64|128, fmind32|64|128, signbitd32|64|128,
    llquantexpd32|64|128, samequantumd32|64|128, quantumd32|64|128,
    issignalingd32|64|128, isinfd32|64|128, isnand32|64|128,
    isnormald32|64|128, isfinited32|64|128, fpclassifyd32|64|128,
    copysignd32|64|128, fabsd32|64|128, llrintd32|64|128,
    lrintd32|64|128, llround32|64|128, lroundd32|64|128,
    roundevend32|64|128, roundd32|64|128, truncd32|64|128,
    floord32|64|128, ceild32|64|128, rintd32|64|128,
    nearbyintd[32|64|128], quantized32|64|128
  * add more tests or fix them for above

Since 1.0.12:
  * Implement ti <-> Decimal conversions
  * Add mantissa extraction functions
  * Optimize 128b conversions
  * Add support for newer s390 cpus

Since 1.0.11:
  * Print correct number of digits testing ulps
  * Ensure hardware dfp is disabled with sw only
  * Fix edge case with strtod*
  * libdfp: powerpc: Fix fabsd128 implementation
  * Fix the usage of mfcr extended mnemonic

Since 1.0.10:
  * Extend the release and build process and fix various
    compile problems

Since 1.0.9.1:
  * Added BID support for x86_64 and x86
  * Fixes to avoid external PLT calls to internal supplied symbols
  * Add optimized fabsd[32|64|128], and copysignd[32|64|128] for
    PowerPC with DFP hardware support

-------------------------------------------------------------------
Mon Dec  4 08:56:36 UTC 2017 - rguenther@suse.com

- Adjust according to packaging conventions, add libdfp-devel-static
  for static library.

-------------------------------------------------------------------
Wed Sep  3 01:41:37 CEST 2014 - ro@suse.de

- sanitize release line in specfile

-------------------------------------------------------------------
Wed Mar  5 15:07:25 UTC 2014 - matz@suse.com

- Upgrade to version 1.0.9.1.  [bnc #866739]
  Support for ppc64le and power8.
  Fix for isinfd[32|64|128] (return +1 for inf, -1 for -inf).
  Fix finited.
  New functions llogbd[32|64|128], roundevend[32|64|128],
  quantumd[32|64|128], new macro issignaling.  All based on
  draft TS 18661.
  Optimized ceild[32|64|128], truncd[32|64|128], floord[32|64|128],
  nearbyintd[32|64|128], roundd[32|64|128], samequantumd[32|64|128]
  for PowerPC with DFP hardware.

-------------------------------------------------------------------
Fri Dec  7 15:17:36 UTC 2012 - matz@suse.com

- Upgrade to version 1.0.8.  [fate #314087]
  Minor bug fixes all over, add c++ header decimal/decimal,
  add dfp/float.h defining _Decimal[32|64|128] if necessary.

-------------------------------------------------------------------
Wed Aug 17 09:12:31 CEST 2011 - dmueller@suse.de

- add baselibs.conf to sources

-------------------------------------------------------------------
Fri Jul  1 15:01:28 CEST 2011 - matz@suse.de

- Upgrade to version 1.0.7.  [fate #311678]
  Added isfinite() and deprecated finite(), per POSIX.
  Fixed locale base printf.
  Fixed printf_dfp() to work for several hundred use cases
    which previously produced partially correct output.
  Fixed strtod*() for all test suite cases.
  Fixed classification function (fpclassifyd*()).
  Fixed regression where libdecnumber sub objects weren't
    built with -fpic and caused segv failure on s390.
  Fixed segv error in printf_dfp() due to negative array index access.
  Exported missing __fabsd* and __fpclassifyd* function prototypes.
  Fixed header include order problem.

-------------------------------------------------------------------
Thu Oct  7 13:50:51 CEST 2010 - matz@suse.de

- Update to version 1.0.3, fixing several bugs, among
  them wrong strtod32/64/128 behaviour.  [bnc #635803, #635809]

-------------------------------------------------------------------
Tue Feb 23 14:38:08 CET 2010 - matz@suse.de

- Package the s390(x) hw-accelerated libdfp into libdir/dfp.
  Compile libdecnumber with -D__STDC_DEC_FP__=200704L to define
  the correct internal symbols.
  [bnc #582176]

-------------------------------------------------------------------
Wed Feb 10 14:10:29 CET 2010 - matz@suse.de

- Remove the requires on libdfp-32bit.  [bnc #578613]

-------------------------------------------------------------------
Fri Feb  5 14:55:25 CET 2010 - matz@suse.de

- Enable build for s390, otherwise 31bit libs would be missing.
  [bnc #572581 again]

-------------------------------------------------------------------
Thu Jan 21 15:30:44 CET 2010 - matz@suse.de

- Update to 1.0.1, build soft-fp plus power6/power7/z9-ec variants,
  add baselibs.conf.  [bnc #572581]

-------------------------------------------------------------------
Fri Nov 20 15:13:25 CET 2009 - matz@suse.de

- New package.  [fate #304044, #307062]

Places

File libdfp.changes of Package libdfp (Revision 4d5d954c618d036dccdbde52b8c34aba)

Places