On Tue, 2 Mar 1999, Uli Wortmann wrote:
>
> me> 1) l2h generates a directory filename/l2h####,
> me> where #### seems to be a process ID number.
> me> This seems to be a temporary directory for
> me> holding images. Is this supposed to be removed
> me> at the end of the run?
>
For those running l2h on an OS (such as redhat linux) where
this problem occurs, you might be interested in a (not too
delicate) solution. I don't guarantee it to work,
especially if you set the temporary directory. But, it
works for me. Here it is:
1) To make the WARNING statement go away, comment out lines
(circa) 7522 like such:
# &write_warnings("\n\n Couldn't remove .$dd$TMP_$dd:$!")
# unless (rmdir ".$dd$TMP_$dd");
and add a line right underneath like:
rmdir ".$dd$TMP_";
This added line is a feeble attempt to delete the directory.
It will not work, but I think it worthy to try.
2) Substitute your path to latex2html for <YOURPATH> in the
following cshell script, which I called l2h_linux:
#!/bin/csh
#-------------------------------------------------
# Script to run latex2html on linux machine
#-------------------------------------------------
/<YOURPATH>/latex2html $argv[*]
echo -------------------------------------------
echo Post latex2html processing
echo -------------------------------------------
while($#argv > 1)
shift
end
rm -r -f $1/l2h*
Then just call l2h_linux <name> rather than latex2html
<name>.
>
>As far as I understand the problem, it is caused by opening $TMP in
>one module of l2h and trying to clean it up in another. I.e., it is a
>matter of scope. Different os, treat the problem differently. Most
>unix dialects allow to unlink a file handle while the filehandle is
>still in use. They quietly wait until the filehandle is closed, and
>unlink the file afterwards. However, quite a few os are not so
>clever, and complain right away.
>
>AFAIK, the problem could only be solved if $TMP becomes a local
>variable to a subroutine which opens and closes $TMP, and calls all
>intermediate steps inbetween. After leaving this subroutine another
>subroutine could clean up. But I fear that this problem will persist,
>unless l2h is redesigned from scratch......(using all these wonderfull
>features of perl like prototyping, object orientation......
-------------------------------------------------------------------
Andy Felt | [EMAIL PROTECTED] |
Dept. of Pure and Applied Mathematics| |
Washington State University | http://www.wsu.edu/~felta/ |
-------------------------------------------------------------------
We could save billions of dollars if we just all agree to call next
year "00".
