Gaetano Giunta a écrit :
> ...
> ps: one more thing about the "email" context: how does it take into 
> account mime types and uuencoding/quoted_printable/base64?
>
>   
I just cannot wrap my head around the "context is only managed 
automatically" part.

Here are common use cases I see:

- most common format have slightly different requirements depending on 
"subcontext":
  - for xhtml it is eg. quite easy to use htmlspecialchars, but if the 
output is to be in a comment tag, it needs to have any "--" sequence 
either ripped off or replaced (while & and < and > are permitted). The 
same applies to CDATA sections that cannot contain ]]>, and, more 
interestingly, "should not" contain escaped &lt; and &amp;
  - for javascript, a literal \n char needs to be escaped if it is part 
of a string, but it can appear as part of processing instructions
  - outputting javascript within html: output needs first to be escaped 
according to js, then re-escaped according to html (usually in comment-mode)

Will this subcontext-encoding be left to user-defined custom-blocks? 
Will it be possible to actually have it working (eg. disabling automatic 
escaping of < and& while in html / avoiding strange interactions)?

Thanks for any clarification
Gaetano
-- 
Components mailing list
[email protected]
http://lists.ez.no/mailman/listinfo/components

Reply via email to