* remove intall script in favor of hooks
* drop msysize patch by pretending to be cygwin
* drop autotools patch, and depend on autoconf-archive instead
In some cases this will introduce cycles which could likely be cleaned
up by only depending on some unix tools instead of the whole autotools.
Rebuilding doesn't make much sense at this point since it just adds a subset
of base-devel which is still installed by default.
gawk – 4.2.0-1 - Update to latest 4.2.0 version, remove unnecessary patch, rebuild with MPRF 4.0.0.1-3
gcc – 6.4.0-2 – rebuild 6.4.0 with isl 0.18, mpc 1.1.0-1 and mpfr 4.0.0.1-3
isl - 0.18-1 – Update to 0.18 version 4.0.0.1-3
Mingw-w64-crt-git – Update to latest mingw-w64 from git
mingw-w64-gcc – 6.4.0-2 – Update to gcc 6.4.0 – rebuild with isl 0.18-1 and mfpr 4.0.0.1-3. Skipped 1 rev to match gcc 6.4.0-2
mingw-w64-headers-git – Update to latest mingw-w64 from git
mingw-w64-tools-git – Update to latest mingw-w64 from git
mingw-w64-winpthreads-git – Update to latest mingw-w64 from git
mingw-w64-winstorecompat-git – Update to latest mingw-w64 from git
mpc – 1.1.0-1 – Update to mpc 1.1.0
mpr - 4.0.0.1-3 – Update to mpfr 4.0.0.1 – added workaround for issue I described above, rekey patches
msys2-w32api-headers-git – Update to latest mingw-w64 from git
msys2-w32api-runtime-git – Update to latest mingw-w64 from git
A number of install scripts were testing for install-info to no effect,
then calling it regardless. For minimal installs, this makes pacman fail
on them visibly.
This fix replicates what other MSYS2 packages, e.g. autoconf, do.
Also none of these packages actually depend on GNU info.
mpfr package had invalid version 3.1.4.-3 when it should have been
3.1.4-3. This was caused because an empty _patchlevel was set appended
to the package version separated with a dot.
This commit changes the package versioning so that the _patchlevel is
appended to the package version only if it isn't empty.
Similar commit 0c86b53 was done to mpfr package in MINGW-packages.
* upgrade some broken packages
* correct some broken urls/checksums
* use secure urls where possible
* use stable url for ncurses
* some whitespace fixes
Remaining md5sums either didn't download or
didn't pass checksum tests.
Christoph Reiter