If we use NoSymbol for a key position, that key cannot be reassigned with xmodmap later on. If we use VoidSymbol instead, xmodmap works. Output from xmodmap -pk is correct, however. See a detailed example and a description at the URL mentioned in this bug report.
May be, it should be filed against xmodmap then? xkeyboard-config is just a configuration DB, we are not really responsible for the way Xorg server/utilities process the configuration:)
I will be filing a report under xorg as well. This affects the xkeyboard config in the sense that until this bug is fixed, it might be desirable to use VoidSymbol instead NoSymbol for positions where no character should be produced. I've specified the pros and cons on the website.
> I will be filing a report under xorg as well. This affects the xkeyboard config > in the sense that until this bug is fixed, it might be desirable to use > VoidSymbol instead NoSymbol for positions where no character should be produced. > I've specified the pros and cons on the website. IMHO the "cons" is too serious to do s/NoSymbol/VoidSymbol/g. Some people may be used to the idea that having Alt pressed on keys without 3rd/4th level gives them 1st/2nd level (for example, in sequences like "3rd-1st-3rd" level user does not have to release Alt in the middle). So I would prefer to have this problem fixed in xmodmap only.
The problem is either in XChangeKeyboardMapping() or XLookupString(). xmodmap just uses them. If one wants a character from columns 1 or 2 (from level 1), those can be placed in their correct columns explicitly. This makes it possible to avoid the bug, and still get behavior users want. In some cases using NoSymbol is OK. This is detailed at the URL.
(In reply to comment #4) > The problem is either in XChangeKeyboardMapping() or XLookupString(). xmodmap > just uses them. > > If one wants a character from columns 1 or 2 (from level 1), those can be > placed in their correct columns explicitly. Yes it can. The problem is that we do not know when people use implicit rule "1->3, 2-4" and when do not. Ok, as an experiment, let's change NoSymbol to VoidSymbol in just one symbols/fi. Could you submit the patch for all the cases when you think this substitution is fair?
I posted the new settings for the "kotoistus" layout at http://kotoistus.tksoft.com/linux/symbols_fi.txt.
committed, thanks
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.