David Nebauer
2006-08-18 10:06:23 UTC
Hi Markus,
The following manpages cause groff errors when they are being processed
for display: refdbib, bib2ris, db2ris, en2ris, marc2ris, med2ris,
refdbxml, refdba, refdbc and refdbd.
The error for refdbc is representative:
----------------------------------------------------------------------------
/usr/share/man/man1/refdbc.1:143: warning: can't find numbered character 160
/usr/share/man/man1/refdbc.1:143: warning: can't find numbered character 160
----------------------------------------------------------------------------
There are two Latin-1 characters 160 (non-breaking space) in line 143 of
refdbc.1.
Line 143 is: 'Table 1. refdbcrc'.
The non-breaking spaces can be made visible by changing them to hashes:
-------------------------------------------------------------------------
<< first line proves there are no hash ('#') characters in manpage source >>
$ cat refdbc.1 | grep '#'
<< now transform non-breaking space to hash
***@hezmana:temp$ cat refdbc.1 | tr '\240' '#' | grep '#'
Table#1.#refdbcrc
$
-------------------------------------------------------------------------
The hashes show the location of the two non-breaking spaces.
In every case where this error occurs it is due to a table caption line
similar to the one for refdbc.1.
These characters are not present in the xml source and are obviously
introduced during the transformation of xml to groff.
On my xterm the relevant region of transformed groff input is displayed as:
-------------------------------------------------------------------------
...
refdbc evaluates the refdbcrc configuration file at startup to
initialize it-
self. Table1.refdbcrc
| Variable | Default | Comment |
...
-------------------------------------------------------------------------
You can see groff strips out the offending characters. Incidentally, the
layout would look better if the table caption had its own line.
I know I'm being picky here but an error is an error...
Regards,
David.
The following manpages cause groff errors when they are being processed
for display: refdbib, bib2ris, db2ris, en2ris, marc2ris, med2ris,
refdbxml, refdba, refdbc and refdbd.
The error for refdbc is representative:
----------------------------------------------------------------------------
/usr/share/man/man1/refdbc.1:143: warning: can't find numbered character 160
/usr/share/man/man1/refdbc.1:143: warning: can't find numbered character 160
----------------------------------------------------------------------------
There are two Latin-1 characters 160 (non-breaking space) in line 143 of
refdbc.1.
Line 143 is: 'Table 1. refdbcrc'.
The non-breaking spaces can be made visible by changing them to hashes:
-------------------------------------------------------------------------
<< first line proves there are no hash ('#') characters in manpage source >>
$ cat refdbc.1 | grep '#'
<< now transform non-breaking space to hash
***@hezmana:temp$ cat refdbc.1 | tr '\240' '#' | grep '#'
Table#1.#refdbcrc
$
-------------------------------------------------------------------------
The hashes show the location of the two non-breaking spaces.
In every case where this error occurs it is due to a table caption line
similar to the one for refdbc.1.
These characters are not present in the xml source and are obviously
introduced during the transformation of xml to groff.
On my xterm the relevant region of transformed groff input is displayed as:
-------------------------------------------------------------------------
...
refdbc evaluates the refdbcrc configuration file at startup to
initialize it-
self. Table1.refdbcrc
| Variable | Default | Comment |
...
-------------------------------------------------------------------------
You can see groff strips out the offending characters. Incidentally, the
layout would look better if the table caption had its own line.
I know I'm being picky here but an error is an error...
Regards,
David.