Re: stop the press!

On Thu, Oct 8, 2015 at 12:06 PM, Stefan Eissing wrote: > I need a tiny patch for mod_http2 that solves an empty response problem with > the latest nghttp2 libraries: > > Jim, can we get that still in? Commited in trunk as r1707468. I quickly ask (since I did not

Re: stop the press!

Yann, care to apply your eyeballs to this one? Old one is reverted. Thanks! Index: modules/http2/h2_util.c === --- modules/http2/h2_util.c (revision 1707479) +++ modules/http2/h2_util.c (working copy) @@ -539,14 +539,36 @@

Re: about those async filters

OK, found the bug. Seems an update to the lastest nghttp2 lib co-incided with the checkin of your async filter changes. Everything is fine now. At least I learned some more about core filters, cannot hurt. Thanks for the help. //Stefan > Am 07.10.2015 um 18:40 schrieb Graham Leggett

Re: about those async filters

On 07 Oct 2015, at 4:30 PM, Stefan Eissing wrote: > In http2 land, the request happen on "pseudo" connections, connections > properly created by ap_run_create_connection(), but with own filters of type > AP_FTYPE_PROTOCOL and AP_FTYPE_NETWORK, registered by

Re: about those async filters

On 08 Oct 2015, at 1:45 PM, Jim Jagielski wrote: > Yeah... it was 'always' foreseen that mod_h2/mod_http2 would provide > useful clues on how to make 2.6/3.0 better, especially w/ the idea of > slave connections; basically, as you say, let the MPM make mod_http2's > job easier

Re: mod_http2 version number

On Thu, Oct 8, 2015 at 8:08 AM, Stefan Eissing wrote: > Well, it is used only in one INFO log statement on startup... > > There are now several people who participated in early trials if mod_h2 > and gave really useful feedback and found bugs. Without a version

Re: stop the press!

On Thu, Oct 8, 2015 at 1:56 PM, Yann Ylavic wrote: > > So we can change that later if necessary Something like this maybe: Index: modules/http2/h2_util.c === --- modules/http2/h2_util.c(revision 1707505)

Re: stop the press!

Valid question. The failure can only happen when a bucket of indeterminate length fails the read. I think the prototype of h2_util_has_eos() needs to change to return arp_status_t and have an int* for the flag. But that change I do not want to do right now for the 2.4.17... //Stefan > Am

Re: about those async filters

Graham, I am very happy to get such feedback and start thinking about how to make the http/2 integration tighter. As it is designed now, it was intended to fit into a 2.4.x server which limited several design decisions and resulted in the separate worker pool. Since HTTP/2 changes the

Re: stop the press!

On Thu, Oct 8, 2015 at 1:38 PM, Stefan Eissing wrote: >> >> Can't tell about the other uses of ast_not_included() though... > > Often, mod_http2 needs to transfer a max amount of data. Sometimes > it "moves" buckets from one brigade to another. Sometimes, it just >

Re: about those async filters

> Am 08.10.2015 um 13:53 schrieb Graham Leggett : > >> HTTP/2 on https:// connections, as reported by Firefox, is currently at ~30% >> of requests. > > We need to make sure we haven’t made a step backwards in the core of the > server. The event MPM solved the C10k problem

Re: svn commit: r1707468 - in /httpd/httpd/trunk/modules/http2: h2_util.c h2_version.h

All green from my side now. //Stefan > Am 08.10.2015 um 12:50 schrieb Jim Jagielski : > > No rush... :) > >> On Oct 8, 2015, at 6:45 AM, Stefan Eissing >> wrote: >> >> Jim, give me a moment. Yann's code reviewing skill are just too good... >>

Re: stop the press!

OK, fair enough. On Thu, Oct 8, 2015 at 12:16 PM, Stefan Eissing wrote: > Valid question. The failure can only happen when a bucket of indeterminate > length fails the read. > > I think the prototype of h2_util_has_eos() needs to change to return > arp_status_t

Re: svn commit: r1707468 - in /httpd/httpd/trunk/modules/http2: h2_util.c h2_version.h

No rush... :) > On Oct 8, 2015, at 6:45 AM, Stefan Eissing > wrote: > > Jim, give me a moment. Yann's code reviewing skill are just too good... > > //Stefan > >> Am 08.10.2015 um 12:44 schrieb Jim Jagielski : >> >> I am guessing this should be

Re: stop the press!

On Thu, Oct 8, 2015 at 1:26 PM, Yann Ylavic wrote: > > Maybe last_not_included() could "eat" leading metadata buckets when > called with maxlen == 0, that is remove the leading "if (maxlen > 0)" > in there. Possibly the original code with this only change would work too...

Re: mod_http2 version number

Well, it is used only in one INFO log statement on startup... There are now several people who participated in early trials if mod_h2 and gave really useful feedback and found bugs. Without a version number of the module being visible in the log somewhere, it would make this kind of tests much

Re: stop the press!

Hi Stefan, On Thu, Oct 8, 2015 at 12:06 PM, Stefan Eissing wrote: > > Index: modules/http2/h2_util.c > === > --- modules/http2/h2_util.c (revision 1707467) > +++ modules/http2/h2_util.c

Re: stop the press!

The mod_http2 stuff for 2.4.x is CTR, meaning that one does not need to propose a backport in STATUS for it to be included; rather, one simply patches 2.4 as required, while ensuring that one doesn't break anything :) > On Oct 8, 2015, at 6:06 AM, Stefan Eissing >

Re: svn commit: r1707468 - in /httpd/httpd/trunk/modules/http2: h2_util.c h2_version.h

Jim, give me a moment. Yann's code reviewing skill are just too good... //Stefan > Am 08.10.2015 um 12:44 schrieb Jim Jagielski : > > I am guessing this should be backported to 2.4.17, right? > >> On Oct 8, 2015, at 6:03 AM, ic...@apache.org wrote: >> >> Author: icing >>

Re: stop the press!

On Thu, Oct 8, 2015 at 1:05 PM, Stefan Eissing wrote: > Yann, care to apply your eyeballs to this one? Old one is reverted. Thanks! > > Index: modules/http2/h2_util.c > === > --- modules/http2/h2_util.c

Re: stop the press!

Thanks for the review. Comments below. > Am 08.10.2015 um 13:26 schrieb Yann Ylavic : > > On Thu, Oct 8, 2015 at 1:05 PM, Stefan Eissing > wrote: >> Yann, care to apply your eyeballs to this one? Old one is reverted. Thanks! >> >> Index:

Re: about those async filters

On 08 Oct 2015, at 1:23 PM, Stefan Eissing wrote: > I am very happy to get such feedback and start thinking about how to make the > http/2 integration tighter. As it is designed now, it was intended to fit > into a 2.4.x server which limited several design

mod_http2 version number

Can we get rid of this? The only meaningful version is the httpd version. (I guess mod_ssl still pretends to have a meaningful version number.) I could see a potential use of a separate version number in the very short term to indicate API breakages in this "experimental" component that we

stop the press!

I need a tiny patch for mod_http2 that solves an empty response problem with the latest nghttp2 libraries: Jim, can we get that still in? Commited in trunk as r1707468. //Stefan Index: modules/http2/h2_util.c === ---

Re: stop the press!

Ah, good and useful hint! Currently working on a better patch and backport that myself then. Thanks! > Am 08.10.2015 um 12:46 schrieb Jim Jagielski : > > The mod_http2 stuff for 2.4.x is CTR, meaning that one does not > need to propose a backport in STATUS for it to be

Re: T of 2.4.17 this week

Jus curious, why is e.g. LUA there ? -Original Message- From: Jim Jagielski Sent: Thursday, October 8, 2015 12:44 PM To: dev@httpd.apache.org Subject: ** Re: T of 2.4.17 this week Since we are not bundling them, we don't. On Oct 7, 2015, at 6:40 PM, Steffen

Re: svn commit: r1707468 - in /httpd/httpd/trunk/modules/http2: h2_util.c h2_version.h

I am guessing this should be backported to 2.4.17, right? > On Oct 8, 2015, at 6:03 AM, ic...@apache.org wrote: > > Author: icing > Date: Thu Oct 8 10:03:30 2015 > New Revision: 1707468 > > URL: http://svn.apache.org/viewvc?rev=1707468=rev > Log: > fix for empty responses with nghttp2 v1.3.x,

Re: T of 2.4.17 this week

Since we are not bundling them, we don't. > On Oct 7, 2015, at 6:40 PM, Steffen wrote: > > Do we need in LICENSE.txt the license of nghttp2 (MIT) ? > > Also missing there libxml2. > > On Wednesday 07/10/2015 at 15:46, Jim Jagielski wrote: >> So the plan is to T

Re: about those async filters

Yeah... it was 'always' foreseen that mod_h2/mod_http2 would provide useful clues on how to make 2.6/3.0 better, especially w/ the idea of slave connections; basically, as you say, let the MPM make mod_http2's job easier by abstracting out a lot of the tricks that mod_http2 needs to do to

Re: stop the press!

Yes, that's what I realized after your first review. I think the currently applied patch works fine and any improvement is to much for 2.4.17. Time to revisit when Graham rips it all apart... ;-) //Stefan > Am 08.10.2015 um 14:36 schrieb Yann Ylavic : > > On Thu, Oct 8,

Re: stop the press!

On Thu, Oct 8, 2015 at 2:12 PM, Yann Ylavic wrote: > On Thu, Oct 8, 2015 at 1:56 PM, Yann Ylavic wrote: >> >> So we can change that later if necessary > > Something like this maybe: Hm, no, this changes "the whole brigade" with *plen == 0 semantic in

RE: Better SO_REUSEPORT

Great!! Thank you very much for sharing this with us! Thanks, Yingqi -Original Message- From: Yann Ylavic [mailto:ylavic@gmail.com] Sent: Thursday, October 08, 2015 10:46 AM To: dev@httpd.apache.org Cc: Lu, Yingqi Subject: Better SO_REUSEPORT Looks like we can do even better/faster

Better SO_REUSEPORT

Looks like we can do even better/faster with it (and latest Linux kernels), soon :) https://www.mail-archive.com/netdev@vger.kernel.org/msg81804.html Promising!

Re: T of 2.4.17 this week

With the changes, I decided to wait a day and let things settle: Will T tomorrow when I return from keynoting at Drupal Camp Baltimore. > On Oct 7, 2015, at 9:45 AM, Jim Jagielski wrote: > > So the plan is to T sometime tomorrow, likely > later on in the afternoon, eastern

Re: T of 2.4.17 this week

On 10/8/2015 3:57 AM, Steffen wrote: Jus curious, why is e.g. LUA there ? Probably because when mod_lua was added it got added and not the case when mod_proxy_html and mod_xml2enc got added nor did I think about it when adding mod_h2. My fault. This is beyond the scope of httpd's source