[Bug 281756] lang/tcl90: Update to 9.0.0 release
- In reply to: bugzilla-noreply_a_freebsd.org: "[Bug 281756] lang/tcl90: Update to 9.0.0 release"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sun, 29 Sep 2024 10:20:46 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=281756 --- Comment #5 from Daniel Engberg <diizzy@FreeBSD.org> --- (In reply to Pietro Cerutti from comment #4) We can refer DISTVERSION_R and _RS to as MAJOR/MINOR/PATCH (version) in some way which some ports uses although that leads to slightly quirky abbreviations at times but more commonly used. I would also argue that the first section of the port Makefile should be easily readable, ie not (unintentionally) obfuscate version and package metadata. Most changes are actually fixes, BUILD_DEPENDS (missing) WRKSRC + WRKSRC_SUBDIR is the correct way of defining a WRKSRC subdir (See Porters Handbok) First instance INSTALL_TARGET+= should be INSTALL_TARGET= as it's not defined earlier (see Mk/bsd.port.mk) GNU_CONFIGURE_MANPREFIX is unnecessary/obsolete as of commit 8f80d6dde15c23f550d7f2201299c04e49d91ecc MAKE_ENV is unused USES= pathfix doesn't match anything and it's patched in post-patch section OPTIONS section doesn't follow recommended layout, see "Example 43. Simple Use of OPTIONS" "Example 45. Practical Use of OPTIONS" for example in Porters Handbook and a bit akwardly placed Might be something else I've forgot... Given that the Makefile isn't long and the new version is actually shorter it's easier to compare both side by side. Unless you have a fix I think we should hold off until atleast Tier 1 platforms passes unit tests with or without DTRACE enabled and report failures upstream/try to figure out what's makes it fail. Best regards, Daniel -- You are receiving this mail because: You are the assignee for the bug.