[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Linux collation problem
From: |
Mikko Huhtala |
Subject: |
Re: Linux collation problem |
Date: |
Mon, 14 Feb 2000 00:28:05 +0200 (EET) |
> . TAB is collated before printable characters
> . blank space and punctuation are *not* ignored in collation
>
> Unresolved @Roman&&1 xref looks hints that the first assumtion doesn't
> seem to hold. This is easy to verify, look at the standard.li and
> check that 2&&1 line goes before 2&&10 line (where 2 is the number
> assigned to @Roman, i.e. there's a line 00symbol 2 @Roman at the
> beginning of the file).
Indeed, lout.li file does not look right. I guess blank space is ignored
(and '&' is collated after number characters). These bits are cut and
pasted in order:
00symbol 18 @BasicSetup @NumberMarker
00symbol 1 @BasicSetup @DocumentSetup @TheoremCounterMarker
00symbol 20 @BasicSetup @DocumentSetup @ReportSetup @SubSectionList
00symbol 21 @BasicSetup @DocumentSetup @ReportSetup @Section @SubSection
00symbol 2 @BasicSetup @DocumentSetup @DefinitionCounterMarker
[...]
12&3722.test.lout.1 0 00012 1544 116 test.lout
1&2537.test.lout.1 0 00001 0 1 test.lout
13&1036.test.lout.1 0 00013 1625 123 test.lout
[...]
21&tag 0 00038 6826 491 test.lout
2&2596.test.lout.1 0 00002 218 17 test.lout
3&2651.test.lout.1 0 00003 366 27 test.lout
Locale seems to have no effect on the collation in this case. I'm using
glibc 2.1.2 on Linux. Is there anything that can be done about this?
Could lout pad the numbers with preceding zeros (02 instead of 2) or does
that lead to problems elsewhere?
Mikko