Ping? I know that asking others to do work for them instead of doing it themselves is frowned upon in OpenBSD culture, but even though I might be able to hack something together if I persevered at it long enough, I unfortunately won't have the free time to do so for a while so I hope that one final ping after the holidays won't be too far out of line.
I would contact the maintainer directly and inquire privately, but the current maintainer is listed as the ports@ mailing list. In short, the version of inputmethods/uim currently in the tree is over six years old, and needs a (strongly advised against) environment variable set to work with GTK+3 applications, since it's so old that it doesn't support GTK+3 except via the old XIM module. I'm hoping that this will be a simple thing to update if the person doing so knows the ports-tree well, but if not and people are too busy with other projects, I'll probably try to see what I can do in a couple months when I have some more free time. -- Bryan On 2014-12-17 15:28:03, Bryan Linton <b...@shoshoni.info> wrote: > [Moving to ports@ from misc@] > > On 2014-12-11 19:12:50, Bryan Linton <b...@shoshoni.info> wrote: > > > > [...] > > > > Note that some applications refuse to accept Japanese input unless > > they're run with the correct locale settings *AND* an overridden > > input module, so I have > > > > bind C-g "env GTK_IM_MODULE=xim LC_CTYPE=ja_JP.UTF-8 gwaei" > > > > in my .cwmrc so that a Japanese dictionary program of all things > > will accept Japanese input. > > > > [...] > > > > I believe I have tracked down the source of this bug. Apparently > the version of UIM currently in the ports tree (1.5.3) simply does > not support GTK-3 applications as evidenced by a lack of a UIM > module in /usr/local/lib/gtk-3.0/3.0.0/immodules wheras one is > present in the GTK-2 directory /usr/local/lib/gtk-2.0/2.10.0/immodules > > Indeed, GTK-2 applications work fine when launched *without* > needing any of the overridden environment variables whereas the > GTK-3 applications need to have both their locale and input-module > overridden to work correctly. > > According to UIM's website, GTK-3 support was first added in > 1.7.0-alpha, and the current version of UIM is 1.8.6, so it would > seem that version 1.5.3 being released on 2008/09/07 is a bit out > of date. > > I tried to see if I could build a new version of UIM by just > bumping the version and adding a few tweaks here and there (such > as adding gtk3 to the build-depends, adding a "--with-gtk3" to the > configure flags, and updating AUTOCONF_VERSION) but it would seem > that the errors I was given were simply beyond my ability to > easily fix. > > I don't suppose some kind soul would be willing to look into the > possibility of churning out an update to UIM in the near future? > > It would certainly benefit people using GTK-3 applications, since > UIM is used for more than just Japanese input. It allows not only > Chinese/Japanese/Korean input, but also provides an easy way to > input IPA symbols (a must for an linguist) as well as a way to use > dead-keys to input the variety of diacritics used in the various > European languages (though I know there are other ways to > accomplish this). > > I would certainly be very appreciative of any efforts towards this, > and would of course be willing to test any updates since it > appears that that is all I can do with my current skill level. > > As GTK3 applications become more and more popular over GTK2 > applications, I imagine that this will become a more pressing > issue, though the fact that I appear to be the first one to have > noticed it probably shows how many people this is really affecting > at the moment... > > Regardless, I appreciate the effort the various OpenBSD developers > have put into bringing a coherent GNOME desktop to OpenBSD. Even > though I use CWM as my window manager, I have no doubt that this > cohesiveness is what has allowed me to simply pkg_add uim and > related components and have everything "Just Work (TM)" for all > these years, so I thank all the people involved. >
