On 3/14/16, Didier Spaier <did...@slint.fr> wrote:
> Let me add: only make a change if there is a proven advantage, like
> _ old way failing (provide example), where the new one does not
I occasionally come across source tarballs where every file
unintentionally has the sticky bit set. (I don't have a clue how this
happens.) The find command doesn't correct this. Combine that with a
badly handwritten makefile or a binary repackaging and you can easily
end up with those sticky bits in the resulting package.
> _ new way *significantly* faster (a few less seconds when the end user
> builds the package is not significant as I see it)
I remember the find command taking about 4 minutes on the kernel
source, a minute or two on some larger packages like Octave. chmod -R
is instant.
> _ guarantee that the new formula neither fails nor introduce unexpected
> side
mkdir tmp
cd tmp
for i in {0..7}{0..7}{0..7}{0..7}; do touch $i; chmod $i $i; done
$YOUR_CHMOD_COMMAND_HERE
find . -exec stat --printf "%n %A\n" {} + | sort > manifest
Compare manifests over different runs to see how the permissions are affected.
-Kyle
_______________________________________________
SlackBuilds-users mailing list
SlackBuilds-users@slackbuilds.org
http://lists.slackbuilds.org/mailman/listinfo/slackbuilds-users
Archives - http://lists.slackbuilds.org/pipermail/slackbuilds-users/
FAQ - http://slackbuilds.org/faq/
