Re: GCC_REQD/USE_LANGUAGES

[nia <nia%mailfence.com@localhost>]

I really respect that USE_LANGUAGES is confusingly named, and misunderstood by many pkgsrc developers as a result (including clearly some senior developers) -- it should really be WRAP_LANGUAGES and it does not pull in any kind of  dependency.

However, I currently don't see anyone with motivation for making this change, just as I don't currently see anyone pouring a ton of effort into maintaining support for older compilers. I had a little bit of capacity for doing so this week, so chose to put my time into committing fixes for bulk build failures (especially on older OSes), but I don't see myself doing this as a commitment, it's just a way to be a little bit productive in a way that poses very minimal risk to pkgsrc use.

The decision to name the pkgsrc variables this way and have them provide this functionality was made many years before I joined the project, so I don't see myself as at fault for either the design of USE_LANGUAGES or for using the variables as they were designed -- all over the tree, not just in this one commit that Greg noticed. I've been commiting recipes that use USE_LANGUAGES and GCC_REQD in this way for a long time now, and it was never a problem until now.

I feel extremely uneasy reverting a fix for bulk build fallout. I would be re-introducing breakage to pkgsrc by reverting this change.

I revently contacted pmc about another fix I committed which was reverted by another developer without discussion, and I've been really put off by the lack of communication. I don't think PMC exists to ensure that bulk build failures persist. :/