GNOME Bugzilla – Bug 694593
doctool: Various patches
Last modified: 2015-02-07 16:58:02 UTC
See attached.
Created attachment 237278 [details] [review] scanner: Synthesize argument names if we don't see one While this is a bit lame, we need to do something. We preserve the warning message.
Created attachment 237279 [details] [review] doctool: Switch default language to C
Created attachment 237280 [details] [review] scanner: Also traverse GObject properties when walking namespace This could be done manually by the caller, but it's better if we're consistent here, since we do traverse fields.
Created attachment 237281 [details] [review] doctool: Use builtin namespace walker Rather than hand-rolling our own. This ensures we traverse all nodes.
Created attachment 237282 [details] [review] scanner: internals cleanup: Key more things off Namespace The .gir format has a weird legacy where stuff like the includes are outside of the <namespace>. But conceptually they're tied together, so let's start reflecting this in the code. This way we can just pass around and look at a Namespace object instead of a 4-tuple of (namespace, includes, c_includes, pkg_config).
Review of attachment 237278 [details] [review]: OK.
Review of attachment 237279 [details] [review]: OK.
Review of attachment 237280 [details] [review]: OK.
Review of attachment 237282 [details] [review]: This looks OK, but you clarify the second paragraph of the commit message to say that this will be something done in the future -- right now we just strip off the includes. ::: tests/scanner/Bar-1.0-expected.gir @@ -7,3 @@ xmlns:c="http://www.gtk.org/introspection/c/1.0" xmlns:glib="http://www.gtk.org/introspection/glib/1.0"> - <include name="GLib" version="2.0"/> Is it intentional that this is removed here?
Review of attachment 237281 [details] [review]: Looks to be a semi-nice cleanup. The _chain stuff isn't too pretty but it works for now.
(In reply to comment #9) > Review of attachment 237282 [details] [review]: > > This looks OK, but you clarify the second paragraph of the commit message to > say that this will be something done in the future -- right now we just strip > off the includes. Hmm? Oh I see, no - we aren't stripping the includes. This patch just ends up fixing the bug where we recursively traversed dependencies of includes, and added them to the generated .gir file. See below. > ::: tests/scanner/Bar-1.0-expected.gir > @@ -7,3 @@ > xmlns:c="http://www.gtk.org/introspection/c/1.0" > xmlns:glib="http://www.gtk.org/introspection/glib/1.0"> > - <include name="GLib" version="2.0"/> > > Is it intentional that this is removed here? Yes because the scanner already processes dependencies, it was busted that we were putting the full dep tree in every .gir file.
(In reply to comment #11) > (In reply to comment #9) > > Review of attachment 237282 [details] [review] [details]: > > > > This looks OK, but you clarify the second paragraph of the commit message to > > say that this will be something done in the future -- right now we just strip > > off the includes. > > Hmm? Oh I see, no - we aren't stripping the includes. This patch just ends up > fixing the bug where we recursively traversed dependencies of includes, and > added them to the generated .gir file. See below. erm, I mean that we're not passing just namespace yet, we're passing a 3-tuple
commit dfeaf33c8a4ae7e25b4a83d0b31c4d435b4ef7de broke the pango build...it looks like we're trying to find --pkg-export at build time.
Created attachment 237529 [details] [review] 0001-scanner-Don-t-re-parse-includes-we-already-found.patch
[Mass-moving gobject-introspection tickets to its own Bugzilla product - see bug 708029. Mass-filter your bugmail for this message: introspection20150207 ]