HAR for SPDY traffic

141 views
Skip to first unread message

Andrea Cardaci

unread,
Mar 13, 2013, 6:47:46 AM3/13/13
to http-archive-...@googlegroups.com
Hi,
reading the HAR spec I noticed some incompatibilities with SPDY for what concerns the header size field:

headersSize [number] - Total number of bytes from the start of the HTTP request message until (and including) the double CRLF before the body. Set to -1 if the info is not available.

This make no sense in SPDY since the headers are sent in binary mode i.e. omitting CRLF and ": ".

I understand that the H of HAR stands for HTTP, and that SPDY is a different thing even if it carries HTTP data. But I was wondering what would be a reasonable compromise. It seems that the HAR saved from the Chrome Developer Tools considers that delimiters for SPDY pages.

Also, is an upgrade of the HAR spec planned to overcome this?

Andrea

Jan Odvarko

unread,
Mar 13, 2013, 3:49:35 PM3/13/13
to http-archive-...@googlegroups.com

Hi Andrea,

 

> But I was wondering what would be a reasonable compromise. It seems

> that the HAR saved from the Chrome Developer Tools considers that

> delimiters for SPDY pages.
Could we update the field spec definition and say that CRLF is not included

in case of SPDY?

 

> Also, is an upgrade of the HAR spec planned to overcome this?

No

 

Honza

 

--
--
You received this message because you are subscribed to the Google
Groups "HTTP Archive Specification" group.
To post to this group, send email to
http-archive-...@googlegroups.com
To unsubscribe from this group, send email to
http-archive-specif...@googlegroups.com
For more options, visit this group at
https://groups.google.com/forum/#!forum/http-archive-specification
 
---
You received this message because you are subscribed to the Google Groups "HTTP Archive Specification" group.
To unsubscribe from this group and stop receiving emails from it, send an email to http-archive-specif...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

Andrea Cardaci

unread,
Mar 14, 2013, 7:25:37 AM3/14/13
to http-archive-...@googlegroups.com
Hi Jan,


Il giorno mercoledì 13 marzo 2013 20:49:35 UTC+1, Jan Honza Odvarko ha scritto:

Hi Andrea,

 

> But I was wondering what would be a reasonable compromise. It seems

> that the HAR saved from the Chrome Developer Tools considers that

> delimiters for SPDY pages.
Could we update the field spec definition and say that CRLF is not included

in case of SPDY?

Well, it could be an option, but consider that SPDY uses header compression, so even without considering CRLF (and other separators) we wouldn't have a fair size of the headers. It would be more coherent to store in that field the actual size of the received headers (TCP payload), but unfortunately that's not so easy to get. At least, using Chrome (see http://crbug.com/154706) and even for HTTP, though I can calculate that number, while in the case of SPDY I'm not aware of a reliable way.

Note, I mentioned Chrome since it's my main concern.
 

> Also, is an upgrade of the HAR spec planned to overcome this?
No

 

Honza

 

From: http-archive-...@googlegroups.com [mailto:http-archive-specif...@googlegroups.com] On Behalf Of Andrea Cardaci
Sent: Wednesday, March 13, 2013 11:48 AM
To: http-archive-...@googlegroups.com
Subject: [har-spec] HAR for SPDY traffic

 

Hi,
reading the HAR spec I noticed some incompatibilities with SPDY for what concerns the header size field:

headersSize [number] - Total number of bytes from the start of the HTTP request message until (and including) the double CRLF before the body. Set to -1 if the info is not available.


This make no sense in SPDY since the headers are sent in binary mode i.e. omitting CRLF and ": ".

I understand that the H of HAR stands for HTTP, and that SPDY is a different thing even if it carries HTTP data. But I was wondering what would be a reasonable compromise. It seems that the HAR saved from the Chrome Developer Tools considers that delimiters for SPDY pages.

Also, is an upgrade of the HAR spec planned to overcome this?

Andrea

--
--
You received this message because you are subscribed to the Google
Groups "HTTP Archive Specification" group.
To post to this group, send email to
http-archive-...@googlegroups.com
To unsubscribe from this group, send email to


For more options, visit this group at
https://groups.google.com/forum/#!forum/http-archive-specification
 
---
You received this message because you are subscribed to the Google Groups "HTTP Archive Specification" group.

To unsubscribe from this group and stop receiving emails from it, send an email to http-archive-specification+unsub...@googlegroups.com.

Andy Davies

unread,
Mar 15, 2013, 10:44:47 AM3/15/13
to http-archive-...@googlegroups.com
I'd be tempted to raise it on the W3C Perf List (http://lists.w3.org/Archives/Public/public-web-perf/) as they're due to revise the HAR spec later in the year.

It's not a very active list and I've not had responses to the HAR issues I raised on it though.

Cheers

Andy
Reply all
Reply to author
Forward
0 new messages