On Sat 21 Sep 2013 08:42:00 PM CEST, Jonas Sicking wrote:
Could you file a bug on this. I believe that for systemXHR we should
permit setting more headers.

maybe this bug[1] could be updated to include more headers?

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=832557


/ Jonas

On Thu, Aug 15, 2013 at 1:06 PM, Roger Dentz <[email protected]> wrote:
On Thursday, August 15, 2013 1:02:11 PM UTC-7, Roger Dentz wrote:
On Thursday, August 15, 2013 9:20:06 AM UTC-7, Andrew Sutherland wrote:

On 08/15/2013 11:29 AM, Roger Dentz wrote:



However it needs to be sent chunked - 
http://en.wikipedia.org/wiki/Chunked_transfer_encoding



The header needs to include "transfer-encoding: chunked" and each block needs 
to be preceded by a length field. This is a requirement of the server, whether we have 
all the info available on disk or not.







Chunked encoding doesn't seem to require any discussion between the



server and your client, so you should be able to just pre-can all of the



chunks.







Blobs know their length (.size) and can be sliced



(https://developer.mozilla.org/en-US/docs/Web/API/Blob#slice%28%29). If



the server is fine with a single chunk, it seems like you could add that



boilerplate.  If the server wants to hear about things in smaller



chunks, you can slice up the Blob to fake smaller chunks and provide the



boilerplate for those chunks too.  Those slices will still only be



streamed from disk as they are needed.







Unless XHR isn't letting you set the transfer-encoding header?







Andrew



Yeah, that's about the size of it.  It ignores setRequestHeader for 
transfer-encoding=chunked and sets the content-length to whatever the sum of 
all your blob parts is.



-R.

So I think I'll focus on TcpSocket - there were several code samples where they 
access port 80 so maybe that's a doc typo where it says apps have to be certified 
to access any port < 1024.  Seems benign enough to allow port 80 access since 
that's what XHR does anyway.  Right now it doesn't work for any port so still need 
to poke around a little.

Thanks fer the advice,
-R.
_______________________________________________
dev-webapps mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-webapps
_______________________________________________
dev-webapps mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-webapps



--
Mathieu Rochette

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

_______________________________________________
dev-webapps mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-webapps

Reply via email to