[pkg-bioc] pbuilder --create [...] should be pbuilder create [...]
David Vernazobres
dv at uni-muenster.de
Sun Oct 21 09:35:29 UTC 2007
On Mon, Oct 15, 2007 at 03:54:44PM -0700, Don Armstrong wrote :
[...]
> > > As a side note, there are multiple locations in the code where you
> > > should not be using system SCALAR, but system LIST instead.
> >
> > Could you give me an example ? I am not 100% sure to follow you.
>
>
> --- r_pkg_ordering.pl (revision 351)
> +++ r_pkg_ordering.pl (working copy)
> @@ -741,8 +741,8 @@
> # expand archives for package building
> if (
> 0 != system(
> - "ln -sf ../$packagehashref->{TarGz} "
> - . $builddirectory
> + 'ln','-sf',"../$packagehashref->{TarGz}",
> + $builddirectory
> . $origsourcefile
> )
> )
> @@ -766,7 +766,7 @@
> # renaming the extracted directory to the Debian model
> if ( -d $packagehashref->{BuildDir} ) {
> print "\tnuking BuildDir for $package\n";
> - system( "rm -rf " . $packagehashref->{BuildDir} );
> + system( 'rm','-rf',$packagehashref->{BuildDir} );
> }
>
>
> For example. In the original code, it's bad news if BuildDir contains
> anything that should be escaped. In the latter cases, you're exec'ing
> rm directly without splitting on word boundaries or in the case of
> shell metacharacters, a shell in between. [As a matter of style too,
> you generally put the result your testing for after the command that
> generates the result.]
I am not also sure to follow on this point, could you give me an
example, or the code part where to look at ?
OK, thanks for your advice, I will take more care on those point. Do you have a full patch
available with those change ? So I can commit the full modification ?
Thanks for your interrest,
David
--
Be realistic. Demand the Impossible.
More information about the pkg-bioc-devel
mailing list