Re: Built-in HTML renderer in Emacs?

[Ted Zlatanov <tzz@lifelogs.com>]

On Wed, 01 Sep 2010 22:26:35 +0200 Lars Magne Ingebrigtsen <larsi@gnus.org> wrote: 

LMI> Ted Zlatanov <tzz@lifelogs.com> writes:
>> OK, s/spammers/evildoers and other miscreants/.  I believe 99% of the
>> users will set this to nil or give up.  We should at least try to give
>> this some nuance; a regex is not the best solution.

LMI> Well, we could have a more fine-grained solution.  Like "I want to have
LMI> privacy", "do a best effort for privacy, but let me loot at purdy
LMI> pictures", and "whatever".

That would be swell (with "custom" as the fourth option).  And it should
have the granularity to specify some combination of symbolic
parameterized rules in an alist, e.g. 
((no-get-params) (min-dimensions 2) (min-area 8)) in addition
to multiple URL regexes.  WDYT?

Finally, there should be a way to get an image in the article buffer
even if it was blocked by going to it and hitting `i' or whatever.

>> Also image URLs that have GET parameters could be filtered out
>> explicitly as a separate rule.  That's very likely to be a web bug.

LMI> There are plenty of images out there that have get parameters.

Yes.  As a rule it would make sense but I agree it's not for everyone.

>>>> 1x1 images are suspicious but a 800x600 image is probably safe.
>>>> Basically the pixel count should be over 8 and each dimension should
>>>> be over 2.
>> 
LMI> You don't know the size of the image before you download it.
>> 
>> It's pretty common to set the width and height explicitly to 1 for web
>> bug images in the HTML.  That's what I was thinking of; sorry for being
>> unclear.

LMI> Ah, right.  Yes, that makes sense.  Please make it so.  :-)

Let's figure out the above rule structure and then I'll write it.

Ted