On Thu, Nov 03, 2011 at 08:25:27AM +0100, Lukas Fleischer wrote: > On Wed, Nov 02, 2011 at 06:13:52PM -0700, Linus Arver wrote: > > On Wed, Nov 02, 2011 at 08:59:15PM -0400, Dave Reisner wrote: > > > On Wed, Nov 02, 2011 at 05:49:27PM -0700, Linus Arver wrote: > > > > For all vcs prototypes, we currently create a temporary build directory. > > > > However, we do not delete it after we are done creating the package. > > > > This wastes disk space. > > > > > > > > We already do a "rm -rf" on the build directories every time after we do > > > > a checkout anyway, so this patch doesn't really change anything. > > > > > > > > > > Big -1 from me. This prevents repackaging. > > > > > > dave > > > > Hmm, I am apparently not knowledgeable at all in what "repackaging" > > means. Since this is the first in the series of 6 patches, I'll have to > > do a rebase (taking into account your other comments). > > Repackaging is the main reason for having separate build() and package() > functions.
Ah, that makes sense... thanks! > `man makepkg | less -p repackage`. TIL about the -p flag... > > > > > I'm thinking of keeping this patch, but just as a comment for all the > > thousands of AUR packagers who blindly use the prototype without taking > > into account their particular needs. > > > > Any objections? > > Putting a comment there might encourage AUR packagers to uncomment that, > especially if they don't know what they're doing. On the other side, > users, who know how makepkg(1) works, also know how to clean/remove the > build directory and don't need such a comment... My two cents. So you're suggesting # code instead of # comment # code ? Well, if that's the case then I must disagree --- I *despise* it when there's code commented out with zero explanation. Looking back, I can see that the comment I wrote is rather verbose; I'll shorten it to a one-liner, for now, like this: # save disk space # rm -rf /temp/build/dir Everyone loves one-liner comments, right? -Linus P.S. I'll wait and let the discussion die down for a few days this time before releasing v3 of this patch series... the traffic for ABS patches is nearly non-existant anyway.
