There is/was a bug in SRS (version 4.02?) in Unix file date
checking. On the version I work with, SRS was checking the
file "ctime" or change time, not the file "mtime" modify time.
I'm not sure if this is the problem you ran into, but I would
guess so. I do copy SRS indices among unix (sun) computers
and use them.
My cure was, in file srs/src/tm.c, function TimeFileCreate
change the returned time from "st_ctime" to "st_mtime"
like this:
fstat (fileNo, &buff);
return buff.st_mtime; /* dgg, was ctime (status change time, bad) */
In article <491kah$2qe at rc1.vub.ac.be>, Guy Bottu <gbottu at ben.vub.ac.be> wrote:
>We have two computers where SRS is running, a DEC OSF/1 and a slower
>DEC ULTRIX. With the increase of the databank sizes the DEC ULTRIX
>spends to much time making the indexes. Therefore I tried to ftp
>indexes from the OSF/1 machine to the ULTRIX machine. Now, SRS seems
>able to use the indexes to perform a search, but when srscheck is run,
>it does not see the indexes, and wants to reindex.
...
--
-- d.gilbert--biocomputing--indiana u--bloomington--gilbertd at bio.indiana.edu
