On 02/05/14 at 04:25pm, Pierre Neidhardt wrote:
> On 14-02-05 09:51:17, Andrew Gregory wrote:
> > On 02/04/14 at 12:16am, Pierre Neidhardt wrote:
> > > # We grab the following fields: repo, name, ver, group, installed, and
> > > # desc. We grab leading space for 'group' and 'installed' so that we do
> > > not
> > > # need to test if non-empty when printing.
> > > - my @pkgfields = /^(.*?)\/(.*?) (.*?)( \(.*?\))?( \[.*\])?\n(.*)$/s;
> > > + my @pkgfields = /^(.*?)\/(.*?) (.*?)( \(.*?\))?( \[.*\])?$/s;
> > > + my $desc = <$syncout>;
> >
> > The description needs to be read after the check for a regex match,
> > otherwise that line will just be lost on failures.
>
> Actually this is done on purpose. If we do not read 'desc' right now, on
> failure
> the next line will be that description which will be parsed as if it were a
> 'repo/pkgname pkgver ...' line.
>
> I am getting you wrong on this?
Right now pacsearch prints any line it can't parse. If you read an
extra line for the description before you check for a regex match,
that line will simply be lost.
> > > if(not @pkgfields) {
> > > # skip any non-matching line and just print it for the user
> > > print $_, "\n";
> >
> > This newline needs to be removed.
>
> Right, forgot that. Thought I cannot find a case when that happens.
