Re: HTML images , spell check , window management bugs

[Brian Stafford <brian stafford uklinux net>]

On Wed, 23 January 12:26 Carlos Morgado wrote:

>> Briefly, Content-Location: specifies a full URL that can be fetched from 
>> the internet.  A UA may check URLs in the HTML document against the MIME
> 
> garrr. bad. death, dishonour or something far worse is farther down this
> path.

Its horrible, I agree.  I reckon the "something far worse" is having to 
implement a "standard" the describes two different ways to do the same thing.  
That just leads to bloated code programmer irritation and fried brain cells.  
I mean - hey I've just implemented tested and debugged my wonderful code which 
frobnicates wonderfully.  It was tough getting there but it works!  Oh dear, 
now I must write some completely different code to do the same thing.  Great!  
Oh, and when both mechanisms to do the same thing are used in the same context 
I must then test that they don't conflict.  So much for "simple" protocols!  
Bah!  Humbug!

>> parts with Content-Location: headers before attempting to fetch them from 
>> the internet.  This mechanism would have to be implemented as well as the 
>> cid: references as different UAs arbitrarly use one or both mechanisms when 
>> composing the aggregate document.
>> 
>> IMO there should be an option to prohibit automatic fetching content from 
>> the internet; the empty images could always be clicked or something similar 
>> to fetch them if the user wants them.
>> 
> IMO MUAs shouldn't be getting images from the internet. if the user wants 
> that
> he says 'open in browser'

True enough.  I guess this is more a philosophical point than a technical 
issue.  A HTTP client that's just good enough for fetching images is fairly 
trivial to implement, or the nanoHttp functions from libxml could be used.  In 
any case, if fetching related content is done, Content-Location: still needs 
to be checked in addition to the cid: stuff even if fetching from the net is 
ruled out of scope.

OTOH, balsa *does* fetch the primary content from multipart/related and there 
is no requirement I can find to fetch or process the related stuff.  So in 
that respect it is a correct implementation as it stands.

>> Just my tuppence worth!
>> 
> 
> my 0.02 EUR

2p == .03 euro :)  anyway don't get me started on Britain's refusal to join 
the Euro.  Does the government here think I *enjoy* paying all that commission 
to Bureau des Changes and having to change currency all the time when I travel 
in Europe?  Do they think I *like* paying international cheque clearance 
charges if I draw on a cheque made out in euros?  What is the big deal 
here?????  Rant ... fume ....

Brian