Re: GIT version - no display of html messages

[Jack <ostroffjh users sourceforge net>]

On 2018-07-07 01:14 PM, Albrecht Dreß wrote:
> Am 07.07.18 18:55 schrieb(en) Jack via balsa-list:
> > Unfortunately, that's just not always realistic.  I agree, almost
> > always, those images are completely unnecessary (even from legit
> > senders) but I seem to frequently get emails where the entire message
> > is embedded in the image.
>
> Really?  From my experiences, it's either dumb spam, or the image is
> embedded into a <a href=…> container to redirect you to an infected
> web site or even immediately download malware if you click on it. 
> Putting the message text into an image prevents the security scanners'
> text analysis heuristics from detecting it (that's why Spamassassin's
> HTML_IMAGE_ONLY_XX rules exist).
>
> Anyway, balsa (and other mua's) let *you* choose: if you are /really
> sure/ you want to load the image, just click the button…
I agree that MOST of the time this happens, you want to stay far away. 
However, I routinely get emails from stores where I HAVE signed up for
the list, or my government representatives, or non-profits, including
major international ones, where the plain text just says "Click here to
see online" or "Your email reader can't read HTML - see the online
version..."  and the HTML version is just boiler-plate, with the actual
message embedded in images.  And I'm talking about legitimate
organizations, where there is unlikely to be any malware, but may well
be trackers.
> > I tried again just prior to your commit, and I get instead:
> >
> > lbh_navigation_policy_decision uri about:blank, type 5, used
>
> Hmmm, this indicates that the second approach (the “if
> (g_ascii_strcasecmp(uri, "about:blank") != 0)…” variant) may fix your
> issue, /without/ removing the EFail mitigation.
Yes, this does now work for me.  Again, for info, what SHOULD that URI
be?  Should it be anything for the basic message itself, on only
something for a link within the message?
> > which would have been printed by .._TYPE_OTHER or ..._TYPE_RELOAD,
> > which still doesn't make much sense to me.
>
> Yes, agree.
>
> > Out of curiosity - clicking WHERE would trigger the
> > WEBKIT_NAVIGATION_TYPE_LINK_CLICKED ?  Would that be clicking some
> > link withing the displayed content?
>
> Yes.  At least for the current git version, it is displayed when you
> click on a link in the message (which in turn opens the browser).
>
> > Also, I'm using webkit-gtk 2.18.6, and have been for at least several
> > months.  It's currently the only version available on Gentoo, so I'll
> > have to check their forums/bugsystem to see if there is any move to
> > upgrade.
>
> I see.  Debian stretch also comes with 2.18.6, I'll install a vm with
> it and try if I can reproduce the issue, stay tuned.
It certainly seems odd.
> Best
> Albrecht.
Jack