[sword-devel] Diatheke - locale - sword - verse parsing segfault

Slavo slavo.bb at freemail.sk
Mon Jun 19 02:14:08 MST 2006

Dear friends in Sword Community

Diatheke on my Mandrake 9.2 (Multimedial 2.4 kernel) ends up with 
segfault, when wrong range on phrase search is set
for example gdb output
without setting of locale ususally this kind:
Starting program: /usr/local/bin/diatheke -b KJV -s phrase -r RM -k love

Program received signal SIGSEGV, Segmentation fault.
0x40060d3d in sword::ListKey::decrement(int) (this=0x8401178, step=1)
     at ../src/keys/listkey.cpp:173
with setting of locele, sometimes previous, and sometimes this kind:
Starting program: /usr/local/bin/diatheke -b KJV -l sk -s phrase -r RM 
-k love

Program received signal SIGSEGV, Segmentation fault.
0x404b78a7 in __after_morecore_hook () from /lib/libc.so.6
When first book is correct then it works, even when second book in range 
is inco                               rrect
However, even more spaces, for example "I  Peter" cause segfault

2.And another thing:
It was inpossible to change cs or sk locales.d names of books.
My changes in those files was just ignored
Even when I deleted cs.conf, cs-utf8.conf, sk.conf, sk-utf.conf in 
/usr/local/sh                               are/sword/locales.d and also 
in build directory (and I did not found such files 
           elsewher) diatheke just ignored it and show results with 
czech (Iso8859-2) local                               e, when I run 
query with -l cs option.
Where does sword get these information? How can I changed it I have even 
restarted computer and deleted /tmp
I am really confused.

3. Previous problem is conncted to fact that diahteke, with output set 
for -e UTF8 show verse names in ISO8859-2 format (maybe taken from shell 
or whatever) and not in UTF8. (if this is so It would be nice, if 
diatheke would rather follow command line options and not it's own 
gueses from environment.)



More information about the sword-devel mailing list