On Thu, 14 Jun 2012 17:11:14 +0200 Stephan Beal <[email protected]> wrote:
> But i think i see the problem now: the fossil binary is not meant (AFAIK) > to be used _directly_ as a CGI binary, and that is in effect what you are > doing. It is meant to be fed "the two-line config file" stub which is > itself the CGI script. i have never tried to use it the way you're trying > to now. > > Another option to force the repo path: pass -R /path/to/repo.fossil > > but note that "fossil -R ..." without any command won't don't anything > useful. > > -- > ----- stephan beal > http://wanderinghorse.net/home/stephan/ > http://gplus.to/sgbeal The option -R doesn't work at all, because fossil needs open checkout, not simply a repository. On the other hand I am almost sure it runs exactly inside the open checkout working directory. All bash documentation points that the executed command will inherit the working directory from the bash script. The last script I tested is: ( http://asm32.hopto.org/update_short.cgi ) ............begin code........................................... #!/bin/bash echo "Content-type: text/html" echo "" echo "<html><head><title>Update the site from the repository.</title></head><body><pre>" pwd cd /home/www/asm32 ls fossil ls --nocgi 2>&1 fossil help --all --nocgi 2>&1 echo "</pre></body></html>" ..........end code............................................... "fossil help" works as expected, but "fossil ls" returns "not in checkout" error. You can see on the result page, the directory listing contains _FOSSIL_ file. So, the question arise: How fossil determines whether it is in the working checkout or not? -- John Found <[email protected]> _______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
