Neither gnat-gpl nor gnat-gcc will build. They're both masked, and if you unmask them the compiles blow up. Please restore dev-ada/gnat until the recommended substitutes actually work. Reproducible: Always Steps to Reproduce:
(In reply to comment #0) > Neither gnat-gpl nor gnat-gcc will build. They're both masked, and if you > unmask them the compiles blow up. Then file bugs about them instead of requesting no longer supported stuff to get unmasked.
They actually do. There is only one outstanding bug (#139241) caused by the leak in the gcc build environment that I am yet to trace (any help even reproducing it is appreciated), however this seems to trigger only for a few people. In any case the legacy dev-lang/gnat is way more problematic than either of the new compiler ebuilds. In fact that masking was long overdue and was triggered by yet another bug report against legacy gnat. Please note - in order to successfully emerge gnat-gxx you *need* to completely unmerge legacy gnat. It *will* interfere (do "quickpkg gnat" before unmerging to prepare binary blob if you want to keep it handy). Please also check that, when unmerging, all traces under /etc/env.d have been removed. If not, remove /etc/env.d/55gnat by hand. Please also, as a precaution, unset all env vars having ADA in the name (set|grep ADA) in the session you will use to emerge new gnat compiler(s). If you will still have problems, please submit a bug report about the one that fails as usual (if it is the same as that bug I mentioned above, you may just add to it). The gnat-gcc/gnat-gpl ebuilds are not masked btw, they are in testing (try it again, there have been many changes, most happen in eclass so that you won't see an indication of them. As these are only compile-time fixes they do not require revision bump either). If you wish them to go stable just mention it. I think they can/should go stable, I just did not have any requests to make them so. George
George, Thanks for taking the time to offer some tips, but unfortunately I still can't get the alternatives to compile. I will file a separate bug report on the error. All that to the side, are you sure it's wise to deprecate GNAT while the alternatives are still designated as test versions? later,
