Talk:Fcitx

From ArchWiki
Latest comment: 9 August 2023 by Lilydjwg in topic Questions about XIM section

Questions about XIM section

I don't understand this section. Firstly, it mentions:

You can use XIM ... without installing the above modules.

So in the case the user doesn't want to use the modules, doesn't that implies the user won't use fcitx? So how is XIM put in the "usage" of fcitx?

Also, isn't it wrong syntax to include spaces between variable names in bash script? am I missing something or the following syntax is wrong?

GTK_IM_MODULE DEFAULT=xim
QT_IM_MODULE  DEFAULT=xim

The section mentions "Using XIM can sometimes cause problems". Just curious about these problems. It would be helpful to include some links mentioning them.

--AvidSeeker (talk) 13:15, 6 August 2023 (UTC)Reply[reply]

According to this stackexchange answer, you might want to set your input method to "xim" if you are using compose keys. I used to set it to "xim" in the past when there was a bug in fcitx5 preventing me from using compose key properly, but it was fixed later on.
And yes, you cannot set environment variables with spaces. That not only violates bash, it actually violates any other things that handles environment variables (specifically it causes a panic in Rust iirc). I am not so sure why it is like that.
windowsboy111 (talk) 13:50, 6 August 2023 (UTC)Reply[reply]
It's leftover from when people prefer ~/.pam_environment to setup environment variables. I've updated it.
XIM programs will crash if fcitx5 is restarted. It may leak keys with e.g. browsers while typing fast. And of course it doesn't work with Wayland applications.
--Lilydjwg (talk) 04:14, 7 August 2023 (UTC)Reply[reply]
"It may leak keys"?! That's a serious issue, where can I read more about this? Also if there are links to other problems, that will be a good read. --AvidSeeker (talk) 06:02, 7 August 2023 (UTC)Reply[reply]
I have no links. They are talked about in chat group. It is key strokes, not crypto keys to be clear. --Lilydjwg (talk) 11:34, 9 August 2023 (UTC)Reply[reply]