> The first thing I noticed after using Devel::Cover was how much output
> it generates. The HTML files depicting the line-by-line status of the
> coverage are enormous.
I assume you're talking about the large amount of data reported and not
the actual file sizes? The "HTML files are massive" problem was fixed
quite a while ago.
> As an alternative, I generated a plain-text file like so:
>
> cover cover_db -report=text > LC_coverage_20040814.txt
>
> Even that file was large.
You're generating basically the same report in a different format. You
should expect the same volume of data.
> Since my coverage was fairly good to begin with, what I *really*
> wanted was a smaller file which reported on the uncovered elements.
> [...]
> Can Devel::Cover's 'cover' program be used to generate reports of
> uncovered statements/branches/conditions/only?
No. There's (currently) no option for doing so. The HTML is well formed,
though, which should make building a filter easy if you know how the
formatting works. ;) e.g.:
#!/usr/bin/perl
use strict;
use warnings;
while (<>) {
# remove source lines with no coverage data
next if (/class="s"/ && ! /class="c."/);
# remove source lines without at least some missing coverage
next if (/class="c."/ && ! /class="c[012]"/);
print;
}
__END__
Add a little nice packaging and that should do the trick.
-mjc
