On Fri, Dec 28, 2012 at 6:26 PM, Paul Gevers <elb...@debian.org> wrote:
> Hi,
>
> [Resent as my previous mail only went to the bts. Mail-Followup-To set
> to the bts. Sorry if you are subscribed and already read my original
> message.]
>
> I am touching on this bug as part of the request by the RT [1].
>
> Some remarks:
> - I suggest explaining in more detail why this version of the package
> deserves a freeze exception [2]. I don't see any RC or important bugs
> fixed, or release goals achieved, so I DON'T expect the RT to grant
> the unblock.
This unblock was filed on 18 Oct 2012, at that time freeze wasn't that deep yet.
> - If I understand correctly, at least, that is how I reproduced it,
> your code.diff is the difference of the 4.2.4.1-7 versus the
> 4.2.4.2-1 package with the patches applied, showing only the "new"
> code.
Right.
> - I wonder if you could explain in more detail what the changes
> are about. The changes are not minor (see below) and not easily
> understandable to an outsider (such as me).
>
See below:
> $ diffstat code.diff
> CMakeLists.txt | 2
Change the version from 4.2.4.1 to 4.2.4.2
> src/frontend/gtk2/fcitximcontext-gio.c | 139
> ++++++++++++++++++++-------------
>
> src/frontend/gtk2/fcitximcontext.c | 134
> ++++++++++++++++++++++---------
> src/lib/fcitx-gclient/fcitxclient.c | 81 ++++++++++---------
These add some checks about errors as well as some GtkIMContextClass
related codes to the input method module, which is loaded as plugin by
GTK+. As seen in the diff, many of the changes are add "if" to handle
unexpected situations, others are about adding a new function to
handle pre-edit and surrounding in a safer way.
These two cases, especially pre-edit, are very important to input
method users. Quoting some old GTK documentation[1]:
Input Methods provide a way for complex character sets to be
used in GTK+. Languages such as Chinese, Japanese, and Korean
(often abbreviated to CJK) use a large number of ideographs,
making it impossible to support all characters with a simple keyboard.
Instead, text is usually pre-edited using a phonetic alphabet
and then composed to form the ideographs.
In detail, alphabets are passed to the application before the
resulting character is committed to it, so that the application can
react to user input much earlier, an example is that search engines
can use the information to predict what the user is about to find so a
list of related candidate can be given before the user completing his
typing.
[1]http://www.gtk-server.org/gtk1/gdk/gdk-input-methods.html
> src/module/lua/luawrap.c | 2
> src/module/xkb/xkb.c | 2
These two are just formatting changes, no real content.
> 6 files changed, 231 insertions(+), 129 deletions(-)
>
> Paul
>
> [1] https://lists.debian.org/debian-devel-announce/2012/11/msg00003.html
> [ Unblocks and Freeze Policy ]
> [2] http://release.debian.org/wheezy/freeze_policy.html
>
>
>
--
Regards,
Aron Xu
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
