Skip to main content.
home | support | download

Back to List Archive

Re: WordCharacters not working in 2.1

From: Bill Moseley <moseley(at)>
Date: Tue Nov 28 2000 - 15:30:03 GMT
Hi Jose,

At 09:40 AM 11/28/00 +0100, wrote:
>The characters are stored alphabetically sorted. I made this for

But the headers shown by swish are different than the headers in my config
file used for indexing.  The headers printed by swish contained characters
that are not in my wordcharacter settings.  And indeed my wordcharacter
settings are being used as I needed to add #&; to get Entities to index.

Look at just IgnoreFirstChar:

~/swishperl/swish-e-2.1.8/t > ../src/swish-e -c swish.cfg > /dev/null   

~/swishperl/swish-e-2.1.8/t > ll index.swish-e
-rw-r--r--   1 lii      users       41810 Nov 28 07:12 index.swish-e

~/swishperl/swish-e-2.1.8/t > date
Tue Nov 28 07:12:44 PST 2000

~/swishperl/swish-e-2.1.8/t > fgrep IgnoreFirst swish.cfg
IgnoreFirstChar  ,

~/swishperl/swish-e-2.1.8/t > ../src/swish-e -w sksks  |grep IgnoreFirst
# IgnoreFirstChar: "'(

>BTW, someone had found a problem with isspace in solaris (need to 
>check other OS) with some accuted chars like í or ú. It returns TRUE
>when parsing lines from config file:

You know, I saw this on Solaris yesterday, but then I forgot about it when
testing on linux.  I wonder why that is.  I also noted I couldn't cut
accented text from my PC ssh window to the solaris ssh window -- and didn't
really think about it.  Humm.  Not sure if that's related or not.

>Wordcharacters abcdefgíopuy
>will give abcdefg as the wordchar string because isspace returns 
>TRUE with í in function getword. It is easy to fix changing from
>issspace(int)(unsigned char)var)

Is that in all versions of swish, or just 2.1?


Bill Moseley
Received on Tue Nov 28 15:32:36 2000