Sorry for the interruption. As someone who makes use of this tool regularly, being able to include the hash-checks in one shot is helpful. Having an alias is alright but in some situations, not enough. I've now decided to simply include it in makepkg. For the benefit of $anybody here it is attached. If you don't know how to apply a patch and you're reading this (gasp!), do the following as root: patch -Np0 /usr/bin/makepkg /path/to/this/makepkg-checkresume.patch

Summary:

+ variable GETINTEG

+ function insert_checksums

+ option -G,--getinteg

eg. To rule the world at once with style we run "makepkg -Gic"

It isn't that much warranted. However, it _can_ save some time. Something like this and Roman's download-resume patch would be good additions for the next update. If this has already been implemented in some other way, I'll be glad to see it soon.

Now for a Q&A session. It appears that makepkg will be in $srcdir when it's called. However, I noticed the following under devel_update function:

sed -i "s/^pkgver=[^ ]*/pkgver=$newpkgver/" ./$BUILDSCRIPT
source $BUILDSCRIPT

I didn't test this as I have no svn et al packages currently, but I see "source $BUILDSCRIPT" a lot. Just wondering whether we need to type in $startdir for every action on $BUILDSCRIPT. Anyway, I think some parts can directly make use of variables from makepkg.conf - messages for instance. I've reflected this to a little extent in this patch, like DB_CHECKSUMS. Do you guys think PKGBUILD-specific options would be useful? Say, something like --pkgver=$somever, so for trivial packages there'd be no need for further user intervention. You'd definitely have to rely on the beast called "sed" exclusively, though.

Thanks and regards