After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 660698 - scanner: remove backcompat copies more safely
scanner: remove backcompat copies more safely
Status: RESOLVED FIXED
Product: gobject-introspection
Classification: Platform
Component: general
unspecified
Other All
: Normal normal
: ---
Assigned To: gobject-introspection Maintainer(s)
gobject-introspection Maintainer(s)
Depends on:
Blocks: 646742
 
 
Reported: 2011-10-02 15:22 UTC by Torsten Schoenfeld
Modified: 2015-02-07 17:01 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
scanner: remove backcompat copies more safely (2.35 KB, patch)
2011-10-02 15:22 UTC, Torsten Schoenfeld
reviewed Details | Review
scanner: remove backcompat copies more safely (2.39 KB, patch)
2012-08-21 19:43 UTC, Torsten Schoenfeld
none Details | Review
scanner: remove backcompat copies more safely (2.37 KB, patch)
2013-02-07 20:55 UTC, Torsten Schoenfeld
committed Details | Review

Description Torsten Schoenfeld 2011-10-02 15:22:30 UTC
Instead of modifying the namespace while we are iterating over it, which leads
to nodes not being tested, set a 'prune' property on the function and read it
in the GIR writer.
Comment 1 Torsten Schoenfeld 2011-10-02 15:22:32 UTC
Created attachment 198018 [details] [review]
scanner: remove backcompat copies more safely

Instead of modifying the namespace while we are iterating over it, which
leads to nodes not being tested, set a 'prune' property on the function
and read it in the GIR writer.
Comment 2 Colin Walters 2011-10-03 14:32:49 UTC
Review of attachment 198018 [details] [review]:

How about "internal_skipped" or "was_moved"?   I guess we could also just do the test directly in girwriter.py.
Comment 3 Torsten Schoenfeld 2011-10-03 14:39:59 UTC
I don't care much about the actual name of the attribute.  But I do prefer the conceptual separation between 'moved_to' and 'prune' (or whatever we end up calling it).  With this patch, 'prune' is equivalent to 'moved_to && !introspectable'.  But with the current patch set attached to bug 646742, 'prune' is also used for another case.  So I think this should stay in IntrospectablePass as opposed to GIRWriter.
Comment 4 Colin Walters 2011-10-03 21:17:20 UTC
(In reply to comment #3)
> I don't care much about the actual name of the attribute. 

Ok.  Would you mind using 'internal_skipped' then?

> But I do prefer the
> conceptual separation between 'moved_to' and 'prune' (or whatever we end up
> calling it).  With this patch, 'prune' is equivalent to 'moved_to &&
> !introspectable'.  But with the current patch set attached to bug 646742,
> 'prune' is also used for another case.  So I think this should stay in
> IntrospectablePass as opposed to GIRWriter.

I see - yes, makes sense.
Comment 5 Torsten Schoenfeld 2012-08-21 19:43:32 UTC
Created attachment 222068 [details] [review]
scanner: remove backcompat copies more safely

Instead of modifying the namespace while we are iterating over it, which
leads to nodes not being tested, set an 'internal_skipped' property on the
function and read it in the GIR writer.
Comment 6 Torsten Schoenfeld 2012-08-21 19:44:13 UTC
Above is an updated patch implementing the name change you suggested.
Comment 7 Torsten Schoenfeld 2013-02-07 20:55:27 UTC
Created attachment 235449 [details] [review]
scanner: remove backcompat copies more safely

Instead of modifying the namespace while we are iterating over it, which
leads to nodes not being tested, set an 'internal_skipped' property on the
function and read it in the GIR writer.
Comment 8 Torsten Schoenfeld 2013-02-07 20:56:56 UTC
Above is a rebased patch.  I think it should be applied irregardless of the outcome of bug 646742.
Comment 9 Colin Walters 2013-02-10 18:23:41 UTC
Review of attachment 235449 [details] [review]:

This looks good to me, thanks!
Comment 10 Torsten Schoenfeld 2013-02-10 19:02:03 UTC
Committed with a small tweak to make the new code more closely match the old code.  Thanks for the review.
Comment 11 André Klapper 2015-02-07 17:01:25 UTC
[Mass-moving gobject-introspection tickets to its own Bugzilla product - see bug 708029. Mass-filter your bugmail for this message: introspection20150207 ]