Matthias Kilian <[email protected]> writes: > Hi, and sorry for the delay, > > On Sun, Dec 21, 2025 at 08:46:14PM -0800, Greg Steuck wrote: >> I previously didn't want to update happy due to the infra missing a >> piece, so I implemented that. What do you think about the following? >> The main point of interest is MODCABAL_DATA_DIR. > [...] >> One side effect of this patch is that MODCABAL_DATA_DIR no longer needs >> to have any particular value; it just needs to be enabled for this >> automated machinery to find the relevant pieces. I could change the >> variable's meaning to be a simple flag enabling the feature. Since only >> two packages use the feature so far, it would be a simple documentation >> update with no ports churn. > > I'd prefer a better name for it, but don't have an idea for a good one ;-) > > In general, I like your diff. Except for the invocation of > populate-datadir.pl. IMHO this should be explicitely run using perl, > like this: > > perl ${PORTSDIR}/devel/cabal/populate-datadir.pl
Good call, done! > The reason is that I'm not sure wether executable bits are always > correctly set when checking this out via cvs. And they aren't set > for sure when just applying your diff. > > Otherwise, ok for me. Thanks, committed now.
