[links] Links and javascript

Mikulas Patocka mikulas at artax.karlin.mff.cuni.cz
Thu Apr 12 13:58:12 MDT 2007


>> Edward Rosten wrote:
>>>
>>>> k at libsee.  The problem is that links would have to generate a DOM
>>>> for javascript to work.
>>>>
>>>
>>> That's for javascript to work fully. However even without a DOM, links
>>> still has some ability to work on those really stupid websites which
>>> replace all normak links with dodgy javascript code.
>>>
>>> I really don't want to loose the ability to do that.
>>>
>>>
>> I see.  I have a different need; I am working on an embedded controller
>> that needs a browser that can respond to user input.
>
> Maybe if someone had the time to reverse engineer and remove the javascript
> bugs
> - the original author (perm) doesn't have time because of school.
>
> Mikulas actually decided to remove the javascript. Most segfaults always came
> from the javascript. I don't know if Mikulas thinks it's a good idea to pass
> the javascript debugging to someone else.
>
> CL<

Javascript code is simply undebuggable and unfixable. I once tried to fix 
one bug and found that Martin Pergel copied the same piece of code 38 
times instead of making it a function. It is like that when graph theorist 
tried to do programming :)

For embedded applications you can simply use old version of links or you 
can add javascript to the new version (I will keep all #ifdefs there 
although they'll never be activated from ./configure). Releasing that 
javascript interpreter to general public is dangerous --- someone may find 
some exploit in it and...

Mikulas


More information about the links-list mailing list