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 98617 - Latest h2defs.py produces is-constructor-of entries
Latest h2defs.py produces is-constructor-of entries
Status: RESOLVED FIXED
Product: gtkmm
Classification: Bindings
Component: build
2.0
Other other
: Normal trivial
: ---
Assigned To: gtkmm-forge
gtkmm-forge
Depends on:
Blocks:
 
 
Reported: 2002-11-15 16:31 UTC by Tassos Bassoukos
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: ---


Description Tassos Bassoukos 2002-11-15 16:31:15 UTC 
The latest h2defs.py from CVS produces defs files that
contain the is-constructor-of parameter for functions
that look like they are constructors. This produces 
warnings with the current GtkDefs.pm:

GtkDefs.pm: Unhandled function parameter ( (is-constructor-of
"GconfSchema") ) in gconf_schema_new
Comment 1 Murray Cumming 2002-11-15 18:45:15 UTC 
h2defs.py is not part of gtkmm. Reassigning.
Comment 2 James Henstridge 2002-11-16 11:04:33 UTC 
Murray: when reassigning accross product boundaries, use "reassign bug
to owner of selected component".  Otherwise I get no notification (you
still own the bug).

Tassos: h2def.py is not intended to create a fully correct defs file.
 Quite often the resulting defs file will need some additional
tweaking to add some information that could not be derived from the
header.

Working out what functions are constructors and what aren't is one
such thing that can't always be deduced.  It seems that the current
heuristics result in more correct defs entries than incorrect ones.

If your tools can't handle (is-constructor-of "...") lines, then that
is a bug in your tools (it has been in all the drafts of the current
defs spec).  If the information is not of any use for you, then your
tool should be ignoring it.

I will leave it up to murray to close this if appropriate.
Comment 3 Murray Cumming 2002-11-20 01:02:35 UTC 
Fixed in cvs. We will just ignore those entries for now - maybe we
could complain about unwrapped constructors in future.