[l2h] A workaround for segmentations

[Jacques B. Siboni]

Dear colleagues  L2H

the problem I had is that there is a huge main program, and a very small
segment.

Namely there is a thesaurus of 5500 entries for the main and a search segment 
generated for each client demand. 

As I understand L2H even with a make, when there is a modification on the
segment
all of the main is recompiled (it last a few minutes)

And when recompiling only the segment it pollutes the main by adding many
references
which become unreachable by the main.

there might be some ways to do it but I failed. I finally found a workaround not
too dirty:

--- When the main has been built I use chattr +i for all the files (linux of
course)
the result is that all of the files depending of main become immutable.

--- Then when I run the segment, it runs ok but with a few set of warnings but
the result is
very satisfying:

--- All of the main references are accessible

--- All of the segment references refer perfectly to the main

If you're interested to test it (it's in French)

the main is at  http://www.lutecium.org/mathemes/index.html

The segment is at http://www.lutecium.org/search.html

In a future version of l2h maybe this kind of facility might me added
by modifying directly the l2h code. I am personally  to bad a programmer in Perl
even to try.

All the best

Jacques


-- 
Jacques B. Siboni mailto:jacsib at Lutecium.org8 pass. Charles Albert, F75018 Paris, FranceTel: +33 142 287 678 Port: +33 612 536 959Home Page: http://jacsib.lutecium.org/  Lutecium pages: http://www.lutecium.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/latex2html/attachments/20200308/701504ae/attachment.html>