Re: [Pan-users] Reading an HTML posting when using 'old' Pan

[Duncan]

thufir posted on Fri, 23 Mar 2012 03:10:52 +0000 as excerpted:

> On Thu, 22 Mar 2012 17:49:40 +0000, Duncan wrote:
> 
> 
>> I think pan would be really nice with something similar.  Combine that
>> with the idea of inserting some sort of visible marker between text-
>> parts, thus eliminating /that/ long-standing issue, and it should work
>> quite well.
> 
> 
> Huh.  If I could help, I would.  I'll test or something, if it helps.
> 
> Can you require claws, and then send it through claws (or a portion
> thereof)?  Or, might that be intolerably slow, I wonder.

That's not really feasible.  But one (individually) /could/ perhaps try 
some sort of trick like setting up claws as pan's external editor.  I'm 
not real sure how that would work; it might well require a shell script 
or the like as an "impedance matching layer", but it should certainly be 
possible for someone sufficiently determined.  (See for example my pan-
attach-kd script from some years ago, using a script with a kdialog UI to 
select a file, encode and attach it, as text.  The script was designed to 
be invoked using pan's external editor feature.  Of course these days, 
thanks to HMueller, at least pan's git version handles attachments 
itself, but I had my script running to do that, well before the 0.90 C++ 
rewrite, even.)

Not being a coder I can't say for sure, but particularly given that claws-
mail is already gtk-based, and I THINK (but am not sure) it's C++ like 
pan, I'd /guess/ that the code to handle the HTML code stripping should 
be reasonably easy to either lift and convert, or simply see what they're 
doing, and write something similar for pan.

Either way, once you've tried claws-mail and have seen just how effective 
the HTML stripper solution works for there, you'll want it for pan!  I 
was a bit skeptical before I tried it, but it certainly impressed me.  
Every once in awhile, there's a case where say a photo caption gets 
blended in with the main page text, but such issues are WAY less frequent 
than I'd have expected, to the point that in general, normal human error, 
spelling or grammar mistakes, seem to occur more often, and in practice, 
one deals with the blended in photo caption in about the same way, 
reading along until something doesn't make sense, stopping, backing up a 
bit and reparsing, to figure out what happened and what was actually 
intended.  Since a person is usually already used to that, and jumbled 
plain-from-html is generally less frequent, in practice, there's little 
or no problem and things "just work"... certainly WAAYYY better than 
trying to visually/manually parse the text from the raw HTML, anyway, 
which is what we're effectively doing with pan ATM.

As I said, it works well enough to be able to read XML-based atom and rss 
feeds as plain text, anyway, to the point that claws-mail with the feed-
plugin is my feed-reader of choice, these days.  If it can handle that, 
and it already handles mail (and can handle news, but claws sucks 
compared to pan for news in other respects) it should handle news in pan 
quite well indeed!

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman