Path: utzoo!attcan!uunet!ginosko!gem.mps.ohio-state.edu!tut.cis.ohio-state.edu!bloom-beacon!EXPO.LCS.MIT.EDU!rws
From: rws@EXPO.LCS.MIT.EDU (Bob Scheifler)
Newsgroups: comp.windows.x
Subject: Re: XRebindKeysym and XLookupString considered inefficient (LONG)
Message-ID: <8908281908.AA00456@expire.lcs.mit.edu>
Date: 28 Aug 89 19:08:27 GMT
References: <14453@haddock.ima.isc.com>
Sender: daemon@bloom-beacon.MIT.EDU
Organization: The Internet
Lines: 30


    I haven't yet seen the R4 tape

You don't have an "R4" tape, you have an alpha release.

    Typically, in order to be
    fully portable, you have to query the server to find out what
    keysyms cause the state mask that you want to be generated.

That's because you're thinking in terms of modifier bits, while the Xlib
interface is in terms of modifiers.  For example, the client might well
want to deal with "Hyper-A", without regard to which modifier bit Hyper
is assigned to.  I'm not saying a bit-interface doesn't make sense, but
a keysym-interface also makes sense.

    It keeps the list of modifier keysyms that
    you passed in, even though their only use is to compute the state bits.

No, they are also used if XRefreshKeyboardMapping is called.

    The data structure for maintaining keysym mappings should be changed
    for speed and flexibility.

You are encouraged to send implementation suggestions/code to xbugs, and we'll
take a look at them.

    The parameters of XRebindKeysym should be changed for ease of use.

An incompatible change is not a possibility.  An alternate interface could
be considered.