Hi there,
I always use #!/usr/bin/env perl in my scripts, though in the case of using interpreter="..." it wouldn't look at the shebang since the command is specified at the command line.
I even tried without interpreter="..." just using <command>./my_script.pl ...</command> so that it would invoke shebang and I get this error:
/bin/sh: ./my_script.pl: No such file or directory
Hi Leandro,
A lot of tools that are perl or python scripts will have the shebang line hardcoded to the system version, e.g.:
#!/usr/bin/perl
If you change that to:
#!/usr/bin/env perl
it'll pick up the version in your path. For example, we had to do this with tophat (python).
However, that only applies to scripts that are run directly -- I'm not sure how Galaxy tool configs and the interpreter parameter work.
Best,
John
On Tue, 13 Jul 2010, Leandro Hermida wrote:
Dear all,
Sorry if I missed it, I looked in the list archives and couldn't find any mention of the topic.
I am developing tools for Galaxy and thus running a Galaxy server from my within my home directory under my username. I have my own Perl
installed in my home directory and this is in my $PATH, I don't use the system Perl, but when I run a Galaxy tool with <command
interpreter="perl"> it is using the system Perl, as if my $PATH is not used at all. Since the Galaxy server and tool runner is running
under my username and it looks like it just forks off a call to "perl ...." when it forks why doesn't it use my Perl?
best,
Leandro