[pacman-dev] [arch-dev-public] info pages

Allan McRae allan at archlinux.org
Sun Aug 10 09:33:29 EDT 2008

Dan McGee wrote:
> On Sat, Aug 9, 2008 at 11:00 AM, Miklos Vajna <vmiklos at frugalware.org> wrote:
>> On Sat, Aug 09, 2008 at 12:12:29PM +1000, Allan McRae <allan at archlinux.org> wrote:
>>> CC: pacman-dev - We should also automatically compress the info files like
>>> we do the man pages.
>> That should not be hard, we have this trivial commit since a while:
>> http://git.frugalware.org/gitweb/gitweb.cgi?p=pacman-g2.git;a=commitdiff;h=e804504286ecdb6eee56a26bfaf7431728ff2507
> Do we do something like the following, or do we hardcode the info dirs
> as well? It seems odd how we laid this out looking back, where the
> DOC_DIRS location is specified but not the mandirs.
> diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in
> index c5dbed9..8feeed3 100644
> --- a/scripts/makepkg.sh.in
> +++ b/scripts/makepkg.sh.in
> @@ -685,9 +685,10 @@ tidy_install() {
>         fi
>         if [ "$(check_option zipman)" = "y" ]; then
> -               msg2 "$(gettext "Compressing man pages...")"
> +               msg2 "$(gettext "Compressing documentation...")"
>                 local manpage mandirs ext file link hardlinks hl
>                 mandirs="usr/man usr/share/man usr/local/man
> usr/local/share/man opt/*/man"
> +               mandirs="$mandirs ${DOC_DIRS}"
>                 find ${mandirs} -type f 2>/dev/null |
>                 while read manpage ; do
>                         # check file still exists (potentially
> compressed with hard link)

Won't that compress e.g. html documentation in ${DOC_DIRS} as well?  In 
general, that would be bad.

I know the reasons are historical, but from a makepkg point of view, why 
are man pages treated as different from info pages?  I see a good reason 
for treating the rest of documentation different.  As far as makepkg is 
concerned man and info pages both need treated the same.  So should info 
be removed from the DOC_DIRS variable?  Kind of makes the including docs 
by default redundant if we do that though...


More information about the pacman-dev mailing list