Debian Patches

Status for lapack/3.11.0-2

Patch Description Author Forwarded Bugs Origin Last update
icamax-izamax-extended-precision.patch Fix i{c,z}amax behavior on any-i386 and m68k On arches with extended precision (any-i386 and m68k), the icamax and izamax
routines must be compiled with -ffloat-store, in order to avoid the extra
precision bits which make the test fails.
.
The FFLAGS_FLOAT_STORE variable is set by debian/rules.
Sébastien Villemot <sebastien@debian.org> not-needed 2015-12-24
doxygen-downsize.patch Downsize documentation (HTML and manpages) - Set EXTRACT_ALL=no, to avoid generating tons of files containing only doc
stubs.
- Do not include verbatim source code in Doxygen output (redundant with the
source package).
- Drop the call graphs.
Sébastien Villemot <sebastien@debian.org> not-needed debian 2017-07-04
makefile-blas-remove-dep.patch remove the target dependency in order to avoid unnecessary rebuild that would incur FTBFS. Mo Zhou no 2019-08-28
verbose-tests.patch sed -i -e 's#\(>\)\s*\($@\)\s*\(2>&1\)#\3 | tee \2#g' TESTING/Makefile Make the tests verbose (printing something to the screen) so that they
won't timeout on buildd.
Mo Zhou not-needed 2019-09-13
cblas64-test.patch Fix CBLAS tests with 64-bit indexing This patch replaces all instances of “int” by “CBLAS_INT” in CBLAS/testing/*.c
(except for RowMajorStrg which has must remain an “int”, since it is declared
as such in CBLAS main code).
.
Interestingly, the bug would not manifest on low endian architectures,
because interpreting an int64_t pointer as an int32_t pointer works there (as
long as the integers fits in 31 bits, which is the case here). But on big
endian architectures, this of course fails badly.
Sébastien Villemot <sebastien@debian.org> yes 2022-11-27

All known versions for source package 'lapack'

Links