Re: Fwd: Bug#740862: duck: Should cease to pass "mailto:" URLs to curl
Hi Simon,
On Fri, Mar 07, 2014 at 09:09:02AM +0100, Simon Kainz wrote:
> >
> >> My *personal* opinion is that I would be fine with either of
> >> these two options:
> >
> >> 1. Bug-Submit might allow Hyperlinks and mailto: links 2.
> >> Bug-Submit should be a Hyperlink and we should enable another
> >> field Bug-Mailto for e-mail addresses.
> >
> 2 is currently implemented and working, 1 is close to being finished,
> with the following check:
>
> If Bug-Submit is mailto: - Style: Grab the email address, and try to
> find out if an A(AAA) record for the given domain exists. Checking for
> MX record is not enough in my opinion, as MTA use A(AAA) records to
> deliver mail if no MX was found. This should at least report email
> addresses which cannot receive emails because there is no domain
> registered.
>
> If it's a Hyperlink: business as usual.
>
> Please share your thoughts about this.
I'm pretty mazed whyt nifty stuff you can do with lintian! :-)
> Another issue:
>
> The Repository: field is specified as "URL to a repository containing
> the upstream sources." This is just the plain URL, without indication
> what kind of VCS upstream might use. So I'm having a hard time to
> distinguish an https:// based SVN url from an https:// based GIT url.
> Same for hg,darcs and some others.
>
> Currently I try to be smart about detecting this (grepping for shh://,
> git://, etc, but for those not-so-clear cases i can currently only
> check whether the url is not 404, without actually being able to check
> if the given VCS would work.
>
> Should this field be changed somehow, like Repository-Git? I'm not
> sure if trying to find out which kind of repo upstream uses by trying
> to appen well known postfixes (eg. _darcs/hashed_inventory for DARCS)
> is the right way. I think this is also an issue for humans, trying to
> use the Repository URL: One would havve to first open the url using a
> browser and then try to find out which VCS was used.
>
> Please share your thoughts about this.
Since in d/control files we are distinguishing between
Vcs-Browser
Vcs-Git
Vcs-Svn
...
it might make sense to apply the same philosophy to upstream metadata
and I somehow guess that the according tests would exist.
> Thank you,
Thank you for your serious work on this
Andreas.
--
http://fam-tille.de
Reply to: