Re: using uscan with Fossil SCM

[Sergei Golovan <sgolovan@gmail.com>]

Hi Romain,

On Sat, May 1, 2021 at 1:26 AM Romain Porte <debian@microjoe.org> wrote:
>
> Hi fellow Debianites,
>
> I am currently in the process of packaging grammalecte [1] that relies
> on the Fossil SCM [2]. However on the taglist page [3] the links are
> directing to Fossil HTML pages, but not .tar.xz files. When browsing out
> the 2.1.2 tag HTML page [4], one has to follow the "check-in" link [5]
> before being able to download an archive from the resulting page [6].

With Fossil you don't have to know the exact artifact hash in order to download
tagged revisions. Knowing the tag name is sufficient. You can construct the
download URL yourself given the tag name. The following debian/watch seems
to work, it uses filenamemangle and downloadurlmangle to manualy specify
the filename and URL to download:

=========================
version=4

opts="searchmode=plain, \
      filenamemangle=s/<a\sclass="taglink"\shref="\/timeline\?t=v([\d.]*)">/grammalecte-$1.tar.gz/,
\
      downloadurlmangle=s/<a\sclass="taglink"\shref="\/timeline\?t=v([\d.]*)">/http:\/\/code.grammalecte.net:8080\/tarball\/Grammalecte.tar.gz?r=v$1/"
\
    http://code.grammalecte.net:8080/taglist \
    <a\sclass="taglink"\shref="\/timeline\?t=v([\d.]*)">
=========================
Cheers!
-- 
Sergei Golovan