Dan McGee wrote:
On Thu, Oct 8, 2009 at 9:10 AM, Cedric Staniewski <[email protected]> wrote:
Currently, a changelog is added to a package if a specific file with a
hardcoded name exists in the PKGBUILD's directory. This approach is not
pretty and also inconsistent with the handling of install files, but it
works.
With the introduction of split PKGBUILDs, however, a drawback in this
old behavior has arisen: you only have the possibility to include one
specific changelog file in either every package defined in the PKGBUILD
or in none.
The use of an additional variable, `changelog`, works around this issue
and makes it possible to include a changelog in only some of the
packages, and besides, each package of the PKGBUILD can have its own
changelog file.
Signed-off-by: Cedric Staniewski <[email protected]>
---
Before we pull this, can you (or Allan) make some changes? It isn't
quite ready for primetime yet.
<snip>
+*changelog*::
+ Specifies a changelog file which is supposed to be included in the
package.
"supposed"? Since we fail hard if it is missing, just keep the
language more direct.
"Specifies a changelog file which is to be included in the package."
In reality, we should keep the language for this and the install file
as similar as possible, as I believe they are under the same set of
rules.
Fixed.
@@ -1059,9 +1059,13 @@ create_srcpackage() {
fi
fi
- if [ -f ChangeLog ]; then
- msg2 "$(gettext "Adding %s...")" "ChangeLog"
- ln -s "${startdir}/ChangeLog" "${srclinks}/${pkgbase}"
+ if [ -n "$changelog" ]; then
+ if [ -f "$changelog" ]; then
+ msg2 "$(gettext "Adding package changelog...")"
+ ln -s "${startdir}/$changelog" "${srclinks}/${pkgbase}/"
+ else
+ error "$(gettext "Changelog file %s not found.")"
"$changelog"
+ fi
fi
local netfile
@@ -1193,6 +1197,11 @@ check_sanity() {
return 1
fi
+ if [ -n "$changelog" -a ! -f "$changelog" ]; then
+ error "$(gettext "Changelog file (%s) does not exist.")"
"$changelog"
Not cool for translators to have two messages that say the same thing,
but are not the same. Please choose one (that hopefully resembles the
install script missing message).
However, why do we do this check in two places? Is it due to split
packages? And we fail in one, but not in the other.
These exactly the same as the two messages for install files. I can
make the two messages the same and just pull in a similar change for the
install file.
And the second check is because the check_sanity script misses variables
in package() functions (http://bugs.archlinux.org/task/16004). The lack
of fail in the second one is a bit weird, but that chack is done at the
end of the packaging rather than at the start where we do fail.
Allan
