[Bug 268018] textproc/py-sphinx: Updating to 5.3.0, ensuring themes are installed, using flit and pyproject-build

2024-01-23 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=268018

Alastair Hogge  changed:

   What|Removed |Added

 Attachment #247782|0   |1
is obsolete||

--- Comment #15 from Alastair Hogge  ---
Created attachment 247886
  --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=247886&action=edit
Update to 7.2.6

Drop DOCS, it potentially causes cyclic dependency problems.

-- 
You are receiving this mail because:
You are the assignee for the bug.


[Bug 276478] Mk/*: Build with a clean environment

2024-01-23 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=276478

Tijl Coosemans  changed:

   What|Removed |Added

 Attachment #247801|0   |1
is obsolete||

--- Comment #4 from Tijl Coosemans  ---
Created attachment 247897
  --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=247897&action=edit
patch2

To fix the build failures this patch introduces a WRK_ENV variable and adds the
following environment variables to it: HOME (elixir ports), PWD (go ports),
UNAME_* (nss, python, reduce-psl), PATH (rubygem ports).

Also add TERM to restore colored output to terminals.  This does not affect
output to files.  And TMPDIR for users who define that.

math/gap: This build failure is unrelated.  The port Makefile defines
SHEBANG_FILES=. which means shebangfix touches every file changing their
timestamp.  This causes the build to think some Makefile.in files are
out-of-date so it tries to regenerate them with automake without regenerating
prerequisite autotools files.  Perhaps shebangfix could restore timestamps like
run-autotools-fixup does, but that's for somebody else to look into.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


[Bug 276478] Mk/*: Build with a clean environment

2024-01-23 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=276478

--- Comment #5 from Tijl Coosemans  ---
(In reply to Mathieu Arnold from comment #2)
I don't mind making poudriere the default but that shouldn't mean patches for
the non-poudriere case are rejected.  I think that would only lead to unneeded
drama in the community.  In this case I think the patch also helps poudriere
because it also eliminates environment variables with common names like ARCH
that some ports need patches for currently.

-- 
You are receiving this mail because:
You are on the CC list for the bug.