[Nix-dev] Re: Tracking upstream releases

[Ludovic Courtès]

Hello,

Michael Raskin <7c6f434c at mail.ru> writes:

> Ludovic Courtès wrote:
>>> Or do we want to add kind of scripting language to extract updates from
>>> *any* kind of html page presenting download links?
>>
>> IMO (and as discussed at
>> http://thread.gmane.org/gmane.linux.distributions.nixos/2421) there
>> should be a `meta.watchURL' attribute (or similar) specifying a regexp
>> of URLs to be watched.
>
> I have written (for my own needs) some Lisp code for comfortable
> extraction of urls (or what not) from a relatively regularly structured
> HTML page. It allows to express the condition in terms of page layout
> per se, not by regexps which have own problems for complicated parsing.

That's neat, but it's often more than needed.

For GNU projects, for instance, it's common for the web page to not
contain a link to the latest tarball nor even the last version number.
However, tarball URLs always follow the same scheme, so it's very easy
from a series of `wget' (or something like "echo ls | lftp
ftp.gnu.org/gnu/PROGRAM") to infer whether a new tarball is available.

Thanks,
Ludo'.