ajordanr-google wrote: > Okay, I think I understand what's going on with fuchsia: they manually set > CURSES_INCLUDE_DIRS, CURSES_LIBRARIES, and PANEL_LIBRARIES. On line 17 we go > into the else-case because TINFO_LIBRARIES is not set.
That's correct! > So next we check if the provided CURSES_LIBRARIES has our symbol... I believe they don't even get this far. They apparently have an issue with find_package/find_library lookup mechanism, and that fails. So it breaks then without even trying to check for the symbol. > (As an aside, while I'm reading this: why is this checking TINFO_LIBRARIES > and not HAS_TERMINFO_SYMBOLS, that's what this package sets. I don't think we > should be mentioning TINFO_LIBRARIES at all, unless we know someone is > setting that manually?) Mostly this variable was derived from the spack commit which uses it. That's the only place I know it exists. Technically now fuchsia is using it too, after they implemented the fix (which also logically makes some sense, `TINFO_LIBRARIES` in their case IS `CURSES_LIBRARIES`. --- If we were to hide TINFO_LIBRARIES entirely in the conditional, we must check `CURSES_LIBRARIES` to see if what's passed has the necessary terminfo symbols (or just tell people it needs to be there). I'm fine with that. We shouldn't _need_ to check `HAS_TERMINFO_SYMBOLS` if they're providing this all directly. --- Cool cool, will do about the `CURSES_LIBRARIES` update. https://github.com/llvm/llvm-project/pull/126810 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
