Re: [evolution-patches] #63731 - You made Japanese input worse than ever

[Not Zed <notzed ximian com>]

Must be some serious performance issues with the im stuff.  But anyway, a comment below.

BTW nobody knows how EText works, its left over from the previous maintainer who dropped it in our lap, so basically if the patch works, its probably ok with us.  Suresh does this look ok to you?

This patch will also need ChangeLog entries but assuming it does what it needs to do and the F10 thing is sorted, it looks reasonable to me.

On Tue, 2004-08-31 at 15:54 +0900, Nakai wrote:
> http://bugzilla.ximian.com/show_bug.cgi?id=63731
> http://bugzilla.ximian.com/showattachment.cgi?attach_id=10003
>
> Hi,
>
> Some fixes in #63731 made Japanese input so slow even on my Xeon 3.2Gx2 with HT machine.
>
> The newly attached patch does:
> - Remove preedit_len and add
> im_preedit,im_preedit_attrs,im_preedit_cursor to e-text.h
> - Fix memory leak in e_text_preedit_changed_cb()
> - Reduce the redundant gtk_im_context_get_preedit_string() call
> - Remove slow codes from e_text_draw()
> - Use gchar* instead of GString
> - Remove pango_layout_*() funcs from preedit handling to speed up
> - Move gtk_im_context_filter_keypress() funcs to
>   GDK_KEY_PRESS section. Because the keypress func
>   doesn't make sense in GDK_KEY_RELEASE section.
This is done before the shift-F10 keypress check, should this be the case?  Will it interfere with accessibility?

> And now it works for me.
> Japanese input feature is indispensable on mail client, so this fix is
> important for Japanese.
>
> Thank you.
> --
> Nakai
> _______________________________________________
> evolution-patches mailing list
> evolution-patches lists ximian com
> http://lists.ximian.com/mailman/listinfo/evolution-patches

--

Michael Zucchi <notzed ximian com> "born to die, live to work, it's all downhill from here" Novell's Evolution and Free Software Developer