On Jeu 28 avril 2005 15:29, Daniel Veillard a �crit :
> On Thu, Apr 28, 2005 at 02:41:31PM +0200, Nicolas Mailhot wrote:
>> >   now do man indent, and count the options.
>>
>> Well, I doubt any user wants
>>
>> <foo xmlns:first="long uri sqlrvnLZ NQLZTJQLTJBLERLZJEL
>> VJZELJZLRBLZREVLZERVLajzlrjalzrvjlazekjlaz" xmlns:second="long uri 2
>> q;sfdv kghb elrh lerhylzreyjnlerylerjlertlz" xmlns:third="long uri 3
>> ljsdg
>> hlberjynlsry ejnlrt kjnmdrgnmfd,hnmr"/>
>
>   I may take a patch, check include/libxml/xmlsave.h xmlSaveOption enums
> which would have to be augmented. The problem is that you need to keep
> track of the column number which is not fun, and probably very
> inefficient.
>
>> And you'll note your argument is somehow unfair - one of your example
>> shows no indenting at the element level though it didn't stop xmllint
>> (like most other tools) indenting them
>
>   I'm not unfair. I expose that writing that code is very special case,
> not fun, probably hard to make efficient, and so far just got something
> which looks like a rant.

I'm sorry if it came out like this. I know writing software is hard, and
"there is no time to do this right now" is a perfectly fine answer with
me.  Especially since doing this stuff requires column tracking, which I
was perfectly aware of (but I'd have loved to be wrong here)

"I don't see there is a problem, you must be mistaken" OTOH just asks for
more arguments, which I dutifully provided ;).

I fear the need will become more evident as namespaces are generalized -
they make the case for indenting attributes awfully clear ;(.

Regards,

-- 
Nicolas Mailhot

_______________________________________________
xml mailing list, project page  http://xmlsoft.org/
[email protected]
http://mail.gnome.org/mailman/listinfo/xml

Reply via email to