Oops, apparently you've sent this answer only to me and it's been
sitting on my inbox for a few days. CC'ing the list back.

At Tue, 19 Oct 2010 11:30:02 +0900,
Carsten Haitzler (The Rasterman) wrote:
> 
> On Mon, 18 Oct 2010 22:55:49 -0200 Raphael Kubo da Costa <[email protected]>
> said:
> 
> > At Tue, 19 Oct 2010 08:26:37 +0900,
> > Carsten Haitzler (The Rasterman) wrote:
> > > 
> > > On Mon, 18 Oct 2010 14:29:24 -0200 Raphael Kubo da Costa
> > > <[email protected]> said:
> > > 
> > > > Sorry, I've kind of missed the point while reading your message ;)
> > > > 
> > > > Are you saying we should use an unsigned long even though it will be
> > > > read as a long by curl, or that we shoulduse POSTFIELDSIZE_LARGE, or
> > > > that it's OK as it is?
> > > 
> > > point: if curl has a limit implicitly placed by using signed long (on 
> > > 32bit
> > > systems this is 2gb - on 64bit systems it's too big to bother
> > > discussing :)). but if OUR api should directly reflect such limits is
> > > another question.
> > > 
> > > is it really valid/sensible to be sending 2gb post data in 1 go? remember
> > > that this will need to be memcpy'd on a 32bit system into an internal
> > > buffer so it can be spooled out - so you will already fail as simply there
> > > is not enough memory space to even do this. for 64bit systems.. is it
> > > sensible to even be sending 2gb posts? i have my doubts. as the webserver
> > > will need to accept such massive posts. i don't see that being sensible at
> > > all. if your connection drops you have to re-send.. the 2gb blob. as such
> > > thats a MASSIVE waste. wouldn't reality be that any such massive file
> > > uploads would always be split over many posts of much smaller chunks? if
> > > that is the case... we could just use an int and be done with it. you 
> > > still
> > > can't even manage 2gb on a 32bit system - and on a 64bit box that will be
> > > the theoretical limit, but in practice any such protocol "design" that
> > > requires you post entire complete 2gb (or bigger) blobs of data is a 
> > > broken
> > > protocol anyway?
> > 
> > Well, I might be wrong here because it's too late to be digging
> > through libcurl's internals right now, but at first glance, it seems
> > to do any required splitting of the content into various requests when
> > it cannot be sent all at once.
> 
> for a post? i thought an http post is a 1 time thing? (a get can get specific
> regions of a file with size+offset, but a post can only send a post as a while
> as a mime encoded file? am i wrong?).

Well, I haven't looked at libcurl's internals ever since, so I can't
tell for sure.

--
Raphael Kubo da Costa
ProFUSION embedded systems
http://profusion.mobi

------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store 
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to