GNOME Bugzilla – Bug 99759
Remove unnecessary markup from glade messages in gnome-control-center
Last modified: 2011-07-11 13:03:05 UTC
#: capplets/keyboard/gnome-keyboard-properties.glade.h:2 msgid "<small><i>Fast</i></small>" #: capplets/keyboard/gnome-keyboard-properties.glade.h:3 msgid "<small><i>Long</i></small>" #: capplets/keyboard/gnome-keyboard-properties.glade.h:4 msgid "<small><i>Short</i></small>" #: capplets/keyboard/gnome-keyboard-properties.glade.h:5 msgid "<small><i>Slow</i></small>" #: capplets/mouse/gnome-mouse-properties.glade.h:2 msgid "<i>Fast</i>" #: capplets/mouse/gnome-mouse-properties.glade.h:3 msgid "<i>High</i>" #: capplets/mouse/gnome-mouse-properties.glade.h:4 msgid "<i>Large</i>" #: capplets/mouse/gnome-mouse-properties.glade.h:5 msgid "<i>Low</i>" #: capplets/mouse/gnome-mouse-properties.glade.h:6 msgid "<i>Slow</i>" #: capplets/mouse/gnome-mouse-properties.glade.h:7 msgid "<i>Small</i>" #: capplets/network/gnome-network-preferences.glade.h:3 msgid "<b>_Automatic proxy configuration</b>" #: capplets/network/gnome-network-preferences.glade.h:4 msgid "<b>_Direct internet connection</b>" #: capplets/network/gnome-network-preferences.glade.h:5 msgid "<b>_Manual proxy configuration</b>" #: capplets/network/gnome-network-preferences.glade.h:6 msgid "<b>_Use authentication</b>" This use of markup inside messages marked for translation is totally unnecessary, and a problem for the translation process. For details, see the problem description in bug 96836. Since these are glade messages, it's probably not as easy to solve as in plain C messages. I'll mark it dependant on the glade bug 97061 in the mean time.
I disagree with the proposed solution and the fact that it is a bug. Translators are the only ones in the position to be able to correctly place some markup. Admittedly this markup is the entire string and hence handleable elsewhere. However, as people become more accustomed to being able to use markup the issue will re-assert itself and we'll be forced to look at it again. If I have a string 'this is <b>bold</b> but this is not' that needs translating. Only the translators can place that correctly. Sure a developer can dick around and do printf (_("this is %sbold%s but this is not"), "<b>", "</b>") or some other silliness but it boils down to the developer hard coding where things go or the translators putting them there. Only the latter seems feasible.
Jody, did you read the rationale in bug 96836? If not, that clearly explains that this only applies to the second case of use of markup in messages, and goes into detail of the many disadvantages of including markup in those cases. If you haven't read it, I suggest you do so and reconsider the resolution of this bug.
I did read it when it appeared, disagreed, and forgot to make comment then. To my mind the burden placed on translators to place the markup is not different than that required to place the printf expansions.
Are you sure you read the right bug report? I'm not convinced. Your comment is about the other type of messages, the type that rationale specifically says it's not about. Also, this is not about a "one-time burden" for translators. Translators face the problems of this whenever they try to update the translation or re-use the translated messages, and the bulk overhead of the markup effectively destroys the usefulness of both fuzzy matching and translation re-use, the biggest advantages of the po format, and the ones that really help consistency.
Please fix, bug 97061 has been fixed, and http://l10n.gnome.org/POT/gnome-control-center.HEAD/gnome-control-center.HEAD.pot still contains markup.
I'm going to address this issue
Fixed in http://git.gnome.org/cgit/gnome-control-center/commit/?id=339738a567fa77ac1f4d3991f5565ffc3c88ff7b There are still some strings containing markup, but either the markup is included inside the string, or the string is in C code (maybe another bug should be opened for the latter).
(In reply to comment #7) > There are still some strings containing markup, but either the markup is > included inside the string, or the string is in C code (maybe another bug > should be opened for the latter). If so, please do. Thanks a lot!
(In reply to comment #7) > There are still some strings containing markup, but either the markup is > included inside the string, or the string is in C code (maybe another bug > should be opened for the latter). > Some of these are from libslab, and I have a patch for these bitrotting in bug #403191 Maybe that can be used :)
See http://bugzilla.gnome.org/show_bug.cgi?id=590933 for the xrandr-capplet.c file.
I had to reintroduce some strings with markup because GTKBuilder format does not support <attributes> tag on Gtk*Button. See #591821.
There's only one bit of markup left, and it cannot be removed (it's a word in the middle of a sentence). commit 5b4d433f772b5538abd19928dc6d62fddc6a2a7a Author: Bastien Nocera <hadess@hadess.net> Date: Mon Jul 11 13:59:29 2011 +0100 universal-access: Remove unused column in model And the markup in the contents as well. commit 0f5e7554fff9f6e9ff2d179f6216c5e835c7da2e Author: Bastien Nocera <hadess@hadess.net> Date: Mon Jul 11 13:40:31 2011 +0100 region: Don't use markup in bold labels commit a24c6a7f6609e5bbb15abf43d9ad62862c84ba2b Author: Bastien Nocera <hadess@hadess.net> Date: Mon Jul 11 13:37:47 2011 +0100 user-accounts: Remove unneeded markup in label