GNOME Bugzilla – Bug 313533
Allow EContactPhoto to refer to images with URIs
Last modified: 2013-09-14 16:49:42 UTC
The vCard specification allows PHOTO fields to either contain inline image data, or to point at an image with a URI. Currently EContactPhoto only supports inline image data, attaching a patch to extend this.
Created attachment 50728 [details] [review] Allow photo to be a URI This patch adds a type member and a union to EContactPhoto. It also sneaks in a patch to e-vcard.c to add: GList *e_vcard_attribute_get_param(EVCardAttribute *attr, const char *name); Which is pretty useful on the whole.
For now changing EContactPhoto would break ABI freeze.
Yes it would. I filed now as I'm working on it now and didn't want to forget about it. :)
Marking as an enhancement.
*** Bug 346544 has been marked as a duplicate of this bug. ***
Created attachment 68352 [details] [review] Revised patch This patch is C89 compliant (no anonymous unions) and is tested: it's in Maemo 2.0.
Any chance of this landing in 1.8?
oh my god, i want this. a friend of mine asked exactly *that* question yesterday and if its possible in evolution. /me wants an ABI break.
ok this can be committed.
This will break Evolution builds and introduce an ABI break. I don't have an Evolution checkout at the moment, will someone else fix that? Also will the versions in configure.ac be increased on release, or should I do that when I commit?
Ok Ross wait till monday i will update you about versions stuff, whether you have to update that or it will be done at release time.
Devashish: a gentle reminder ping re versioning and breaking EDS. :)
you can commit the patch ross, versioning will be handled at the release time. if you want me to commit lemme know.
Whilst preparing for committing I noticed that this patch depends on an API addition to libebook. If #347752 is rejected I can re-code this patch, but that function call is very useful.
Committed to CVS, thanks!
Created attachment 69041 [details] [review] Patch to fix the build. The above change is breaking the build. I have not tested this with ldap backend, but it fixes the build.
Arse, I appear to not have LDAP headers so didn't catch this. You'll want to initialise the data.inlined.mime_type field to NULL too, but that looks reasonable to me.
there is also a patch by hiroyuki at bug 347870 to fix this. well, reopening this bug as long as head is not fixed.
Created attachment 69092 [details] [review] Revised patch I noticed that the previous LDAP patch was committed without incorporating my feedback. Attached is a patch that finished the migration.
Created attachment 69093 [details] [review] Revised patch GCC kindly pointed out that a == b == c in C doesn't do what one expects.
Committed, thanks. I think that is everything in EDS sorted now.
thanks for the effort Ross.