memory is going)?
I had a very similar situation to this, when a bug in a third-party
library caused it to buffer everything in my filter. mod_diagnostics
rapidly tracked that down for a x300 performance improvement.
--
Nick Kew
architecture, so if you tie something to them,
you're more-or-less fighting against the architecture. That means
your module is likely to be non-portable and at risk of breaking on
future Apache updates.
--
Nick Kew
In urgent need of paying work - see http://www.webthing.com/~nick/cv.html
.
Your best bet may be to try and use a persistent connection,
and handle breaking the connection as an error.
--
Nick Kew
and modularity.
I intend to implement this using httpd 2.0, if that makes any difference.
Yes, that's a far more powerful development platform than 1.x.
--
Nick Kew
horribly
when an installed module has been updated. We could do with a mechanism
for that, and if Gerardo is going to implement it then great (though
I can't help thinking it ought to be simpler: an unload-all-modules
thing).
Perhaps this wants a bug report.
--
Nick Kew
Apache?
--
Nick Kew
Apache header files.
--
Nick Kew
with
the linkage from C. I've done that experimentally, building the W3C
CSS validator with gcj, but considered this just too huge and unwieldy
to contemplate for operational use.
--
Nick Kew
to that. And it might also be very useful
for one of my wishlist-projects if that ever happens.
Thanks for clarifying. I'm just off to your sourceforge page to
learn more:-)
--
Nick Kew
into the developer docs?
--
Nick Kew
third-party patches. It's most likely to get
adopted if you raise it as a bug report, then attach the patch to that.
--
Nick Kew
another module is
setting it, but I should implement a configuration directive as an
alternative. Make existing behaviour the default, but let
httpd.conf override it and force or suppress output compression.
--
Nick Kew
As discussed previously, here are my updates as a patch against 2.0.49.
They serve to enable working with compressed data coming from a proxy
(or other backend) and processing content in the output filter chain.
--
Nick Kew
Nick's manifesto: http://www.htmlhelp.com/~nick/
--- mod_deflate.c.old
in mod_deflate
destroy their brigades. A look at apr_brigade.c shows that it's not
in fact necessary, but maybe a note to that effect would be in order?
--
Nick Kew
Nick's manifesto: http://www.htmlhelp.com/~nick/--- mod_deflate.c 2004-04-18 13:06:13.0 +0100
+++ mod_deflate.c.old 2004-04-18
graceful recovery when proxying content containing bogus characters
through a markup filter.
Rather than use hacks like AP_FTYPE_RESOURCE-1, would it not be
better to introduce a new output filter type AP_FTYPE_PREPROCESS
below RESOURCE for this kind of application?
--
Nick Kew
Nick's manifesto: http
. This seems to offer the compiler
more scope for generating thread-safe code than setjmp/longjmp,
but I really don't know if that's wishful thinking ...
Where do I stand using either setjmp/longjmp or try/throw/catch
with different MPMs?
--
Nick Kew
Nick's manifesto: http://www.htmlhelp.com/~nick/
that (originally
mine, but now more actively developed by others as mod_transform).
Perhaps mod_perl users might benefit from switching?
--
Nick Kew
for it - although there would seem to be no reason
to do so other than particular system optimisations.
--
Nick Kew
://www.theatreroyal.com/showpage.php?dd=1theid=2578
which now has three nights left to run).
More interesting is the entire subject of filtering in a dynamic
context such as a proxy. The directives available to control filtering
are simply not up to it. Watch this space:-)
--
Nick Kew
is made.
Indeedie.
--
Nick Kew
...
grumbleWhy isn't this documented in the manpages or in zlib.h/grumble?
--
Nick Kew
. The latter have just come from the backend
(the server proxied). But how/why should there be (any) cookies in
r-err_headers_out at this point? Presumably they'd be from the proxy
rather than the backend? And why merge them into a normal 2xx response?
--
Nick Kew
On 25 Jun 2004, Joe Schaefer wrote:
Jim Jagielski [EMAIL PROTECTED] writes:
Nick Kew wrote:
[...]
It appears just to merge Set-Cookie headers in r-err_headers_out with
those in r-headers_out. The latter have just come from the backend
(the server proxied). But how/why should
with yours, but at least it's now clear what's going on.
--
Nick Kew
Dieterich [EMAIL PROTECTED]
--
Nick Kew
I've rolled a fairly extensive mod_proxy patch, which seems rather
big to commit without review. Comments please:
(1) Bug #10722 - cookie paths and domains in reverse proxy
Following my patch to 2.0.49, I've adapted it for 2.1, taking into account
Jim's patch. In doing so, I made some
On Sat, 26 Jun 2004, Graham Leggett wrote:
Nick Kew wrote:
Can anyone see why proxy_fixup should not be removed altogether?
Proxy fixup seems to do the job of making sure the URL /%41%42%43
matches ProxyPass /ABC http://xxx/ABC;, so I don't think it should be
removed altogether.
I don't
On Sat, 26 Jun 2004, Graham Leggett wrote:
Nick Kew wrote:
Can anyone see why proxy_fixup should not be removed altogether?
Proxy fixup seems to do the job of making sure the URL /%41%42%43
matches ProxyPass /ABC http://xxx/ABC;, so I don't think it should be
removed altogether.
OK
minor hacks to 2.0. Need time to
generalise/abstract it into a proposal.
--
Nick Kew
.
Without DEBUG, it will default to current (fixups) behaviour.
Of course, DEBUG won't work with conditional (Request)Header directives.
In addition to documenting this, attempts to do so will log a warning.
--
Nick Kew
On Sun, 4 Jul 2004, Nick Kew wrote:
(If this gets the thumbs up, I'll be happy to do the work:-)
Since reaction seemed broadly positive, I've checked it in to HEAD.
Following suggestions in the replies, it's invoked by the keyword
early which takes the place of the env clause.
--
Nick Kew
this kind of thing work?
As I said, the trivial cases should (transparently) be treated separately
and more simply. Otherwise ... well, as discussed on IRC.
--
Nick Kew
serving from a complete document in-memory or on-disc, it's
straightforward.
- Teach certain problem modules (mod_gzip if appropriate) to react
That'll be mod_byteranges.
--
Nick Kew
On Tue, 13 Jul 2004, Joe Orton wrote:
On Mon, Jul 12, 2004 at 03:35:12AM +0100, Nick Kew wrote:
This doesn't completely address the issue that this might cause
excessive memory usage; particularly if we have to serve ranges in a
perverse order. I would propose two admin-configurable limits
options for the proxy to pass range
requests through or not.
--
Nick Kew
On Tue, 13 Jul 2004, Graham Leggett wrote:
Nick Kew wrote:
Indeed. In a straight-through proxy that's right.
But in the case of a cacheing proxy, it may be better for it to retrieve
the entire document and manage byteranges locally. And in the case of
a content-transforming proxy
content transformation
Each filter is the stack could promote the complexity but should never set
it to a lower state. This would allow http/proxy modules to negotiate less
complex transformations in more efficient ways.
Nicely put. Thank you!
+1
--
Nick Kew
been sent down the wire). With my patch it might still
risk minor breakage (repeated gzip header) if the headers filter changes
sometime in future.
Any more issues with this? If not I'll make nd's fix and leave it.
--
Nick Kew
a simpler solution but found it didn't work. Just enough
to persuade me to preserve the loop-over-buckets test.
--
Nick Kew
/longjmp with Worker or Windoze MPM asking for trouble?
And if so, is there an alternative approach I could try?
--
Nick Kew
that
mod_proxy and friends don't provide?
--
Nick Kew
of it, and implement load-balancing and fault-tolerance
in additional modules. As a matter of fact, one of my wishlist-projects
is a connection-pooling module for backend HTTP connections in a proxy.
That might actually be the same as your project.
--
Nick Kew
?
Well let see my suggestion :
Makes sense.
With the caveat that proxying plain HTTP can do much more than some
posts in this thread seem to think. So the motivation has to be
people want AJP, not HTTP can't do things.
--
Nick Kew
with using apr.
Agreed. But I don't have APR karma to introduce the idea there.
--
Nick Kew
:-(
--
Nick Kew
On Thu, 22 Jul 2004, Nick Kew wrote:
On Mon, 19 Jul 2004, Joe Orton wrote:
Nothing like that was posted to the list, at least. Patch below is
still sufficient to fix the proxy+304 case; does it work for you too?
Yes, mostly (it fixes the important bug that was previously a
showstopper
on IRC. Perhaps it can lead to a
general-purpose module for 2.0 and an architecture update for 2.2?
http://www.apachetutor.org/dev/smart-filter
--
Nick Kew
cannot see any circumstance
in which it's necessary in an output filter. But it's not a
requirement of the proposed archictecture, except insofar as
it saves the overhead of a filter initialising when it's not
going to be run.
--
Nick Kew
On Fri, 30 Jul 2004, Joe Schaefer wrote:
Nick Kew [EMAIL PROTECTED] writes:
[...]
(2) I propose getting rid of it because I cannot see any circumstance
in which it's necessary in an output filter.
But you still need a simple way for an output filter to run some code
before
On Fri, 30 Jul 2004, Joe Schaefer wrote:
Nick Kew [EMAIL PROTECTED] writes:
[...]
If there's a better approach, I'd be glad to update those docs.
Wouldn't an ap_hook_insert_filter() handler be the ideal spot for
that?
I thought such hooks were run on *every* request, not
just
from Bill Rowe above,
or merely with my proposed solution to that problem?
--
Nick Kew
On Sun, 1 Aug 2004, Justin Erenkrantz wrote:
--On Sunday, August 1, 2004 8:24 AM +0100 Nick Kew [EMAIL PROTECTED] wrote:
I'm not sure what 'match' is in this context.
In the above case, it could be text/html or latin1.
ap_register_smart_filter(transcode, latin1, charset_filter, ctx
need that level of
flexibility.
I haven't figured out how I'd implement that ...
--
Nick Kew
fall out of
a general review of this.
--
Nick Kew
On Mon, 2 Aug 2004, [ISO-8859-15] Andr Malo wrote:
Now we have your additional charsets twice...
Erk! So we do. I guess the best fix is just another update to chop the
duplicates?
--
Nick Kew
/a into their HTML.
Does this make sense? Or could we simply drop the AddDefaultCharset
from the installation default as suggested by Duerst and others?
--
Nick Kew
, this is supposed to be working on Bug 23421 - see my other post.
BTW2, there seems to have already been a duplicate entry for Big5 :-)
--
Nick Kew
On Mon, 2 Aug 2004, [ISO-8859-15] Andr Malo wrote:
* Nick Kew [EMAIL PROTECTED] wrote:
Our shipping with AddDefaultCharset preconfigured is causing lots of
pages to be served with a bogus charset, typically where authors
rely on meta http-equiv ... and either don't know how to fix
the admin has configured it?
--
Nick Kew
that, it begs the question of how to treat
unknown HTTP/extension methods (cf DAV), and suggests perhaps
RequireRequestBody should be made a configuration directive.
--
Nick Kew
the
request came with C-L but an input filter invalidated it.
--
Nick Kew
body with the
connection-close afterwards as 'persistent connection' MUST NOT be present.
Nope. GET requests routinely have keep-alive, but don't have bodies.
--
Nick Kew
On Sun, 8 Aug 2004, [ISO-8859-15] André Malo wrote:
A CGI script therefore should never trust Content-Length, but just read
stdin until it meets an EOF.
That is well-known to fail in CGI. A CGI must use Content-Length.
--
Nick Kew
.
--
Nick Kew
. Was this necessary?
(2) ISTR some discussion of generic connection pooling, but I don't
see it in the code. Am I missing something, or is this still TBD?
--
Nick Kew
implemented using apr_reslist.
/me kicks himself for not looking inside proxy_util.c :-(
Thanks.
--
Nick Kew
On Wed, 18 Aug 2004, Brian Akins wrote:
Shouldn't we still set Vary: Accept-Encoding if no-gzip is set?
Hmmm, makes sense. +1
Should we be looking at a wholesale backport from 2.1-head? There are
a number of minor bugfixes, as well as the inflate output filter.
--
Nick Kew
Server website.
Please don't do that. A URL for a tar.gz file is much friendlier on
peoples inboxes.
--
Nick Kew
it clear what's happening there.
--
Nick Kew
On Tue, 24 Aug 2004, Nick Kew wrote:
I actually have an implementation based on the discussion document and
addressing the concerns people raised in the thread. I hope to find
time to finish the accompanying documentation and post it here round
about this coming weekend.
OK, since you seem
variable with
mod_rewrite and dispatching on an env= variable (example: insert
DEFLATE depending on both Accept-Encoding request header and
Content-Type response header. mod_rewrite can't do that because
it runs too early to be sure to have the response headers).
--
Nick Kew?xml version=1.0
On Sat, 28 Aug 2004, Graham Leggett wrote:
Nick Kew wrote:
I posted my proposed smart filter module a few days ago, in response
a post here identifying a situation where it is relevant.
Ideally there should be one way of loading modules, not two - if it is
practical to fix
to suppress it.
The patch is a one-liner. Unless anyone can come up with a reason
why it might open a security hole, I'll apply it.
--
Nick Kew
for proprietary extensions (which HTTP explicitly permits) becomes
to allow them in a 304.
--
Nick Kew
for
someone else to fix it), and crucially also taking responsibility
for closing bugs with INVALID, WONTFIX or WORKSFORME where no
satisfactory resolution is/seems feasible.
--
Nick Kew
, including test-driving with existing
filter modules, and will make it easier to coordinate the API
updates in util_filter.
Is this something that wants a vote? Anyone else have strong
feelings for or against putting mod_filter under CVS?
--
Nick Kew
it, we should have a reason that
addresses Doug's and Ryan's arguments on the record.
--
Nick Kew
On Wed, 1 Sep 2004, Jeff Trawick wrote:
I can't see how this ever worked before :( Any comments from the crowd?
FWIW, I fised that one in the proxy context about two months ago.
But I haven't looked at it in the general case.
--
Nick Kew
On Wed, 1 Sep 2004, Jeff Trawick wrote:
On Wed, 1 Sep 2004 20:36:07 +0100 (BST), Nick Kew [EMAIL PROTECTED] wrote:
FWIW, I fised that one in the proxy context about two months ago.
But I haven't looked at it in the general case.
was that this change entry?
*) mod_proxy: multiple
to look at the
source.
--
Nick Kew
the new code in a stable harness.
--
Nick Kew
timescale: what should their
status be? I wouldn't suggest removing them, but perhaps we could
flash up a prominent WARNING when you configure/build them?
--
Nick Kew
.
--
Nick Kew
or info.
--
Nick Kew
with util_filter (protocol should be
configured in ap_ calls rather than httpd.conf), and to investigate
whether it can be harnessed to fix architectural bugs like PR#17629.
--
Nick Kew
with #if APR_HAS_MMAP or is MMAP 'manatory'?
Thanks for the feedback. Yes of course, I've just patched that
(and grepped apr_buckets.h for any other APR_HAS_* that might
bite on some other platform).
The whole function it's in exists purely to support reporting bucket
types for the FilterDebug option.
--
Nick
first floated the concept, less when I posted a first-
pass implementation, and only yours on introducing it to CVS. Maybe
what it needs now is an updated roadmap to stimulate discussion?
As for a hooked-in build file, I have yet to RTFM for what that involves.
--
Nick Kew
binary-compatibility (ABI) is sufficient for that, innit?
--
Nick Kew
be included in the httpd.conf). The more translations we add, the less
applicable is it to include the whole doc tree.
Hmmm, does that risk generating a higher volume of dumb-newbie questions
in all the public fora? And perhaps also apache-is-hard articles in
the press?
--
Nick Kew
in no hurry to do that. What I can also do is re-implement all the
outputfilter directives within mod_filter and its updated framework.
--
Nick Kew
, including
but not limited to fixing the problems identified above?
--
Nick Kew
On Wed, 22 Sep 2004, Justin Erenkrantz wrote:
--On Wednesday, September 22, 2004 5:01 PM +0100 Nick Kew [EMAIL PROTECTED]
wrote:
I've said it before and I'll say it again: AddOutputFilterByType is
fundamentally unsatisfactory. This confusion is an effect, not cause.
Suffice to say, I
On Wed, 22 Sep 2004, Justin Erenkrantz wrote:
--On Wednesday, September 22, 2004 6:17 PM +0100 Nick Kew
[EMAIL PROTECTED] wrote:
It seems to me heavily counterintuitive that mixing ByType directives
with anything else means that the ByType filters *always* come last.
And that Remove
would fall straight out of his Capturing a
Subrequest. But it seems we're all stuck on partial insights.
I'm provisionally +1 on Paul's proposed fix, but I wonder if it should be
conditional on ap_is_initial_req, to leave untouched the 'normal' CGI
case.
--
Nick Kew
anything.
Ask yourself: is your code sufficiently different to anything we already
have to merit releasing separately as a third-party module? If yes, then
do that. If no, then it's probably appropriate to offer a patch.
My guess would be no.
--
Nick Kew
that. People who want it can
help themselves, without compromising stability.
--
Nick Kew
.
--
Nick Kew
about housekeeping and what ends up
in the logs. Do we want to log a successful POST or PUT when it wasn't?
--
Nick Kew
On Wednesday 13 September 2006 22:31, Jeff Trawick wrote:
On 9/13/06, Nick Kew [EMAIL PROTECTED] wrote:
On Wednesday 13 September 2006 20:33, Ruediger Pluem wrote:
Wouldn't it make sense to return OK even if rv != APR_SUCCESS in the
case that c-aborted is set, just like in the default
that's not OK/DECLINED and is out of range
for HTTP, then return 500 to the client, and log a buggy content
generator message to error_log.
Thoughts?
--
Nick Kew
1 - 100 of 1431 matches
Mail list logo