Re: [Bug-wget] Wget bug - srcset

[Tim Ruehsen]

Am Donnerstag, den 01.02.2018, 14:00 +0100 schrieb Bence Bojti:
> Dear Developers!
> 
> I have a problem with the Wget 1.18 (installed from standard repo of
> Debian
> 9).
> 
> I need wget to mirroring a website with --page-requisites.
> 
> Unfortunately on my website, which I want to mirror ( https://mno.hu
> ),
> there are lot of <picture> tag with a few <source srcset="...">. This
> is
> valid html5 code.
> 
> For example:
> 
> <picture>
>     <source media="(min-width: 1238px)" srcset="/test/1.jpg">
>     <source media="(min-width: 926px)" srcset="/test/2.jpg">
>     <source media="(min-width: 614px)" srcset="/test/3.jpg">
>     <img src="/test/4.jpeg" alt="Alt text">
> </picture>
> 
> In this situation the wget doesn't download the images given in the
> srcset.
> Just the img src.
> 
> On this page https://savannah.gnu.org/forum/forum.php?forum_id=8568 I
> can
> read, that the version 1.18 can handle the srcset: "Parse <img
> srcset>
> attributes on a recursive download. ".
> 
> But I dont know that this refers just to the srcset within the <img>
> tag,
> and not the <source>, or this is a bug in this release. But all the
> same.
> It would be great, if wget would follow the new html stanards, 'cause
> I
> cannot use so now for my site. :(

You are right, the srcset attr is only supported within img tags by
wget 1.x.

Wget2 supports all those HTML tags / attrs out of the box.
Debian unstable (maybe testing ?) has an older version packaged (search
the net to find how to install packages from Debian unstable/testing).
Even better might be to git clone from latest master and build wget2 by
yourself (https://gitlab.com/gnuwget/wget2.git).

Regards, Tim

signature.asc
Description: This is a digitally signed message part