https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=290583
Jason E. Hale <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] Flags|maintainer-feedback? |maintainer-feedback+ Status|New |Open --- Comment #2 from Jason E. Hale <[email protected]> --- (In reply to Daniel Engberg from comment #1) Daniel, you have a valid point that the '-j' flag in ctest(1) functions differently than that of bmake(1), gmake(1), or ninja. The ports system, however, really expects MAKE_JOBS_NUMBER to be a positive nonzero integer and does absolutely nothing to check the sanity of this user-settable value. For example, bmake allows the -j flag to be a floating point or end in 'C' per the manpage, and ports that just use bmake will build with rather bizarre strings in /etc/make.conf like 'MAKE_JOBS_NUMBER=3.14159_gimme_some_pi'. *dog in burning house - this is fine - meme* Build systems like gmake or ninja would bark and bite at such a declaration. Since meta build systems like CMake and meson use ninja by default, they would also crap out. *qmake quietly chuckles to itself* That said, I'm more inclined to agree with sunpoet. This variable, however it is called, should be documented in Mk/Uses/cmake.mk, though. -- You are receiving this mail because: You are the assignee for the bug.
