[Xapian-tickets] [Xapian] #707: Problems on OS X El Capitan

Xapian nobody at xapian.org
Thu Mar 8 04:30:46 GMT 2018


#707: Problems on OS X El Capitan
-----------------------------+-----------------------------
 Reporter:  james            |             Owner:  james
     Type:  defect           |            Status:  new
 Priority:  normal           |         Milestone:  1.4.x
Component:  Xapian-bindings  |           Version:
 Severity:  normal           |        Resolution:
 Keywords:                   |        Blocked By:
 Blocking:                   |  Operating System:  Mac OS X
-----------------------------+-----------------------------

Comment (by olly):

 I think the patch on your github branch was since addressed (in a similar
 but slightly different way) by [4c60bf7fc0d8c90d2d65f5bc96acefcf5998cea9].

 And I've just pushed [b0f8bfb073fe71dda699304312bb71b0fa0b3852] which runs
 `prove` via `App::Prove` so specifying `PROVE` is no longer necessary.

 There's still stuff unaddressed here I think though - going through the
 list in the original report:

 Point 1 - we don't seem to have documented that anywhere yet.  However, it
 seems #732 may offer a way to make this work for system interpreters.

 Point 2 (ensuring `DYLD_PATH` is specified so that at least non-system
 interpreters work) should be addressed by
 [4c60bf7fc0d8c90d2d65f5bc96acefcf5998cea9].

 Point 3 (python generation of `.pyc` and `.pyo`) is I think also address
 by [4c60bf7fc0d8c90d2d65f5bc96acefcf5998cea9].

 Point 4 (perl and prove) should be addressed by
 [b0f8bfb073fe71dda699304312bb71b0fa0b3852].

 I think we don't know about Java yet.

 Point 5 (faketime in omegatest) should be addressed by
 [2f535671ef3d74bf74abb945f1170ef973769110], though it might be better to
 test it on omega rather than on the `date` command.  But I think it'll
 fail on the `date` command because that's protected by SIP, while it'll
 fail on the in-tree omega because that's actually a libtool shell-script
 wrapper, and gets SIP protection thanks to the shell being invoked to run
 it.  I'm not sure I see a nice way to get this to actually work (best way
 to make it work I can see is to link a second version of omega with `-no-
 install` so there's no shell wrapper), but disabling tests that need
 `faketime` isn't a big deal really - there aren't many and the get
 exercised on other platforms.  It's not nice that they get disabled for
 kind of the wrong reason though.  If you installed coreutils from source
 in a non-SIP directory and put that on your path, then `faketime date`
 could work while `faketime ./omega` would still fail.

 So if you get a chance it would be good to:

 * Test that the patches applied address the points I think they do.
 * See what the situation is with Java.

 Beyond that:

 * We ought to document whatever cases remain that don't work, but perhaps
 that's better done once we've looked at #732 (at least if we can do that
 fairly soon).
 * Perhaps adjust faketime handling.  The downside of testing if faketime
 works by running omega under it is that if omega is buggy these tests
 could be skipped because we think faketime doesn't work.  I think I'm OK
 with this as it is overall, unless someone has a better plan.

--
Ticket URL: <https://trac.xapian.org/ticket/707#comment:7>
Xapian <https://xapian.org/>
Xapian



More information about the Xapian-tickets mailing list