On Mon, 13 Aug 2018 13:09:53 +0200 Agustin Henze <[email protected]> wrote:
> On Wed, 25 Jul 2018 12:04:01 +0200 Agustin Henze <[email protected]> wrote:
> [...]
> > I'm struggling to get a debian/watch rule to fetch new upstream releases of
> > gcc-arm-none-eabi... It seems that uscan is getting some trouble with the 
> > href
> > I need to match, they appear to be empty which is not true. Just please see 
> > on
> > the paste[0] what I'm talking about.
> > I have to admit that the href are totally weird with spaces and a lot of
> > strange characters... But I think it should handle this case.
> > 
> > Thanks in advance,
> > 
> > [0] https://paste.debian.net/1034027/
> 
> Ok, I've tried locally removing the part `data-eula-href=""` from the href and
> it works. I have no clue why uscan is dropping the entire href because of
> this... But it seems to be valuable information to get the bug fixed.

Please review if this patch is ok for apply, it should work ok as far as I
understand. But it'd be good if someone with experience on perl, regex and
uscan review it carefully.

--- /usr/bin/uscan      2018-08-13 15:12:31.374367891 +0200
+++ uscan       2018-08-13 15:12:27.758379028 +0200
@@ -3294,7 +3294,7 @@
            uscan_verbose "Matching pattern:\n   @patterns\n";
        }
        my @hrefs;
-       while ($content =~ m/<\s*a\s+[^>]*href\s*=\s*([\"\'])(.*?)\1/sgi) {
+       while ($content =~ m/<\s*a[^>]*\s+href\s*=\s*([\"\'])(.*?)\1/sgi) {
            my $href = $2;
            my $mangled_version;
            $href = fix_href($href);

The modified regex forces to recognize that before href there is a space in
contrast with the previous regex which matches for example with
something-weird-href="" which is not ok.

-- 
TiN

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to