Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 07 Feb 2020 11:44:26 +0000
From:      bugzilla-noreply@freebsd.org
To:        python@FreeBSD.org
Subject:   [Bug 243937] lang/python38: Fails to package in certain conditions due to DISABLED_EXTENSIONS not working
Message-ID:  <bug-243937-21822-8TN82Uh1K2@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-243937-21822@https.bugs.freebsd.org/bugzilla/>
References:  <bug-243937-21822@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D243937

--- Comment #6 from John Hein <jcfyecrayz@liamekaens.com> ---
(In reply to Kubilay Kocak from comment #5)
python37 has patch-issue20210 to "solve" how to disable building/installing=
 the
extensions.  That got repocopied from python36 and python35/34 before that.=
=20
You probably already know this.

I don't know why it was not pulled into python38.  Maybe the patch no longer
applied?  I didn't check.  But from reading the upstream issue 20210, it's
clear they went another way than the original patch we've been carrying alo=
ng.=20
The  *disabled* marker method in Modules/Setup* was committed (looks like it
was committed in python37 in 2017).  So I think using that going forward is=
 the
right way to go (instead of trying to resurrect the old patch which upstream
decided not to use).

I just read bug 241416.  The extensions problem was known and deferred (unt=
il
someone could develop the fix).  So I guess it just hadn't gotten addressed
yet.

I think (but have not tried) that this patch would work for our latest
python37.  We could choose to apply it there as well or just live with the =
old
patch as long as it still works.  pro for the new patch in python37: future
python37 updates might break the old patch.  con argument: unnecessary
work/thrash (wait until current flavor stops working for py37, if ever).  In
any case, that decision can be separated from this fix for py38.

Re: your second question re: third party headers.  I don't know if having g=
dbm,
et. al., ports installed before building py38 affects whether those extensi=
ons
are built or not.  I suspect it does, but I have not done that test to conf=
irm.
 In my environment, they were installed.

--=20
You are receiving this mail because:
You are on the CC list for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-243937-21822-8TN82Uh1K2>