Re: One last question

Hi,

Rich Shepard writes:
 >    I really don't try to break things that work for others; sometimes it just
 > happens. Attached is the output of:
 > 
 > refdbc: getref -t ris -o editme.ris :KW:=shortnose
 > 
 > that shows the entries that produce two records in the keyword list. That's
 > also attached.
 > 

 > shortnose
 > shortnose 

These entries differ by a trailing space at the end of the second one
(users with graphical mail clients may have to view the raw data of
the email message to see the difference). This explains why the
shortnose appears twice in the keyword list.

It does not explain why the two references are not separated by a
space when you retrieve them. The relevant lines in backend-ris.c look
like this:

if (print_field_ris(item, ptr_rendinfo, "\nTY  - ") == NULL) {
[...] 
if ((new_ref = mstrcat(*(ptr_rendinfo->ptr_ref), "\nER  - \n", ptr_rendinfo->ptr_ref_len, 0)) == NULL) {

That is, a newline is inserted in front of each type tag. As each end
of reference tag is closed with a newline as well, two
consecutive references are always separated by an empty line. You
don't, by any chance, use a filter or an editor that skips empty
lines?


As I'm more or less stuck here, I think it is best to have a look at
your original data. Are you sufficiently familiar with PostgreSQL to
create a SQL dump of your reference database? Something like:

pg_dump -C -O -U username dbname > dumpfile.txt

I'll then try to restore the database here and see if your data are
broke or something else.

regards,
Markus


-- 
Markus Hoenicka
[email protected]
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de