https://bugzilla.novell.com/show_bug.cgi?id=705238
https://bugzilla.novell.com/show_bug.cgi?id=705238#c9
--- Comment #9 from Thomas Dickey 2011-07-15 23:19:25 UTC ---
hmm - taking more time to look. It's the encoding that is the issue.
I'm used to considering "uxterm" to be the normal way that utf-8
encoding is used.
Using ltrace, I see that if xterm uses utf-8 encoding, it makes
10 calls to XQueryLoadFont. ltrace shows enough context to see
which fonts are being loaded, e.g.,
1) nil2 (which "should" be a small font),
2) fixed
3) a request for a bold version of "fixed", which is 8859-1
4) double-width medium 10646
5) double-width bold 10646
6) the "font" value from UXTerm app-defaults, which is 10646-1
7) a 10646 version of #3
8) same as #4
9) same as #5
10) unknown (context doesn't show a string)
The basic reason why it's preloading those is so that it can refer
to GC's for each. I can probably improve on the caching, e.g.,
so that 3-5 and 7-9 weren't accessed immediately (the caching work
that I did a couple of years ago simply didn't go far enough).
Still, that's only a factor of 2-3 over where it is now - so you'd
be downloading a megabyte. Using Xft doesn't affect that much,
unless xterm is unable to find the bitmap fonts - in that case it
relies only on "fixed". Thinking about _that_, I've already modified
it to work with just Xft. Another configuration tweak would allow it
to ignore bitmap fonts. Then you'd only have Xft to be concerned with.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.