These involve:
- Formatting and spacing.
- Switching to using fetchFromGitHub where possible.
- Adding missing meta attributes.
- Correcting license values or adding missing license values.
- Adding vyp as maintainer for unmaintained packages.
None of these changes should affect the build result.
A different revision is used for luasqlite3, however, the source code
between these revisions is actually the same. (And the advantage is that
the new revision is a tagged release.)
The shared objects are not available in 'pcre.dev', so to build the rex_pcre.so
file we must specify 'pcre.out' as the directory to obtain shared objects
from, 'pcre.dev' is still needed for the header files though.
This has surfaced due to 505d7bea3a and
the fixup commit 77f5a50c40.
The individial platform attributes are a list rather than single
elements, so in this case we got:
with platforms; [darwin linux freebsd illumos]
Which results in:
[ ["x86_64-darwin"]
["i686-linux" "x86_64-linux" "armv5tel-linux" "armv6l-linux"
"armv7l-linux" "aarch64-linux" "mips64el-linux"]
["i686-freebsd" "x86_64-freebsd"]
["x86_64-solaris"]
]
So if you don't have allowBroken set in nixpkgs config, you end up with
an evaluation error because meta.platforms is expected to be a list
instead of a list of lists which we got here.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Cc: @mornfall, @vcunat
In line with the Nixpkgs manual.
A mechanical change, done with this command:
find pkgs -name "*.nix" | \
while read f; do \
sed -e 's/description\s*=\s*"\([a-z]\)/description = "\u\1/' -i "$f"; \
done
I manually skipped some:
* Descriptions starting with an abbreviation, a user name or package name
* Frequently generated expressions (haskell-packages.nix)
* removing luaMessagePack (only used by neovim)
* using libmpack (updated to latest master since they fixed a lot of packaging
issues)
* package libmpack lua bidnings
* neovim expression now excepts luaPackages as an input and not individual
lua packages (to avoid depending on different lua version)
The most complex problems were from dealing with switches reverted in
the meantime (gcc5, gmp6, ncurses6).
It's likely that darwin is (still) broken nontrivially.
