> The web-scraper algorithm only finds the earliest version of each
> Ubuntu package.
>
> I hope there's a better way to get this information, instead of
> page-scraping, but don't know it to suggest it.
It's still a web-scraper, but the attached patch makes it find
all Ubuntu versions of a package.
Steve
commit cac56bacf9abfbaca6834c90e47591335443a581
Author: Steve Cotton <st...@s.cotton.clara.co.uk>
Date: Tue Dec 30 15:41:50 2008 +0000
Show all Ubuntu versions
diff --git a/whohas b/whohas
index 0de8f47..7815980 100755
--- a/whohas
+++ b/whohas
@@ -279,19 +279,25 @@ sub ubuntu {
my @groups;
my $now = 0;
for (my $i = 0; $i<@lines; $i++) {
- if ($lines[$i] =~ /<h3>Package /) {
- my $name = (split /h3>Package |<\/h3>/, $lines[$i])[1];
- push @names, $name;
- my @parts = split /href\=\"|\"\>|<\/a\>/, $lines[$i+3];
- $parts[4] =~ s/ \(|\)://g;
- push @groups, $parts[4];
- push @repos, "";
- push @urls, $base.$parts[2];
- push @sizes, "";
- push @dates, "";
- @parts = split />|: /, $lines[$i+6];
- push @versions, $parts[1];
- $i += 11;
+ (my $name) = ($lines[$i] =~ /<h3>Package (.+)<\/h3>/);
+ if ($name) {
+ $i += 3;
+ # There are now one or more 8-line blocks that are
approximately
+ # $lines[$i] <li class="intrepid"><a
class="resultlink" href="/intrepid/dpkg">intrepid</a> (base):
+ # $lines[$i+3] <br>1.14.20ubuntu6: amd64 i386
+ while ($lines[$i] =~ /class="resultlink"/) {
+ push @names, $name;
+ my @parts = split /href\=\"|\"\>|<\/a\>/,
$lines[$i];
+ $parts[4] =~ s/ \(|\)://g;
+ push @groups, $parts[4];
+ push @repos, "";
+ push @urls, $base.$parts[2];
+ push @sizes, "";
+ push @dates, "";
+ @parts = split />|: /, $lines[$i+3];
+ push @versions, $parts[1];
+ $i += 8;
+ }
}
}
for (my $i = 0; $i < @repos; $i++) {
