On 18/01/2013 10:23 a.m., Alex Rousskov wrote:
On 01/17/2013 10:56 AM, Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/16/2013 09:58 PM, Amos Jeffries wrote:
* timeout() happens first and runs doneConnecting() while connect() is
queued.
Yes (timeout wins and ends ConnOpener job).
It is
On 18/01/2013 10:53 a.m., Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/17/2013 10:56 AM, Rainer Weikusat wrote:
Amos, do you remember what "partially open FD" in the following comment
meant, exactly? Why cannot we just call comm_close() to cleanup?
// it never reached fully open,
On 18/01/2013 2:51 p.m., Alex Rousskov wrote:
On 01/17/2013 02:53 PM, Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/17/2013 10:56 AM, Rainer Weikusat wrote:
Then, the ::connect
method (or some code running on behalf of that) would close
temporaryFd_ before scheduling a 'delayed connect r
On 18/01/2013 5:37 p.m., Alex Rousskov wrote:
On 01/17/2013 06:35 PM, Amos Jeffries wrote:
Amos, do you remember what "partially open FD" in the following comment
meant, exactly? Why cannot we just call comm_close() to cleanup?
It means we have issued socket() and connect() OS calls
On 18/01/2013 6:47 p.m., Alex Rousskov wrote:
On 01/17/2013 10:28 PM, Amos Jeffries wrote:
On 18/01/2013 5:37 p.m., Alex Rousskov wrote:
On 01/17/2013 06:35 PM, Amos Jeffries wrote:
Amos, do you remember what "partially open FD" in the following comment
meant, exactly? Why cannot we
On 18/01/2013 2:34 a.m., Tsantilas Christos wrote:
Hi all,
this patch removes the "repinning" code from squid.
A related discussion in squid-dev mailing list is here:
http://www.mail-archive.com/squid-dev@squid-cache.org/msg18997.html
Detailed description of the patch follows:
Propagate
On 17/01/2013 6:47 a.m., Alex Rousskov wrote:
Hello,
clientProcessRequest() assumes that a CONNECT request is always
tunneled and sets flags.readMore to false. However, if url_rewriter
redirects the CONNECTing user, Squid responds with a redirect message
and does not tunnel. In that case,
On 19/01/2013 11:53 a.m., Alex Rousskov wrote:
On 01/18/2013 05:32 AM, Amos Jeffries wrote:
On 17/01/2013 6:47 a.m., Alex Rousskov wrote:
Hello,
clientProcessRequest() assumes that a CONNECT request is always
tunneled and sets flags.readMore to false. However, if url_rewriter
redirects
On 19/01/2013 12:19 p.m., Rainer Weikusat wrote:
Amos Jeffries writes:
On 01/17/2013 02:53 PM, Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/17/2013 10:56 AM, Rainer Weikusat wrote:
Then, the ::connect
method (or some code running on behalf of that) would close
temporaryFd_ before
On 19/01/2013 12:32 p.m., Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/17/2013 02:53 PM, Rainer Weikusat wrote:
[...]
If the write handler was scheduled
by the time timout runs, this happened because 'something definite' is
known about the state of the connection attempt: Either, it su
As I mentioned in the last release announcement for 3.3. This month I am
ticking off the requirements for 3.3 to go stable.
As it stands we have four bugs which appear to be major enough to block
that release:
* http://bugs.squid-cache.org/show_bug.cgi?id=3740
- "ssl_bump none ..." apparentl
On 21/01/2013 9:21 a.m., Rainer Weikusat wrote:
The attached patch removes the timeout checking code from
ConnOpener::connect and changes ConnOpener::timeout to actually force
the connection attempt to be terminated with 'a timeout error' instead
of invoking the connect method. The main motivatio
Hi Alex,
could you clarify the situation you mention in
http://bugs.squid-cache.org/show_bug.cgi?id=3111 please?
You said there that ConnOpener leaves events scheduled on successful
exit. But, I was of the understanding that we should NOT explicitly
clear and cancel any Calls in ConnOpener(
On 21/01/2013 11:54 p.m., Steve Hill wrote:
On 18.01.13 18:37, Alex Rousskov wrote:
However, this brings us back to trunk r11920 and bug 3420.
Ok, reading that bug it seems I shouldn't be calling
enableAutoConsumption() directly - this needs to be handled through
expectNoConsumption()?
I
On 22/01/2013 3:04 a.m., Steve Hill wrote:
On 18.01.13 18:37, Alex Rousskov wrote:
Further to this...
* Use expectNoForwarding() instead of adding forwardToBitbucket().
* In ConnStateData::noteBodyConsumerAborted() do not stop reading if
flags.readMore is set. Just keep auto-consuming. This m
On 22/01/2013 9:15 a.m., Tsantilas Christos wrote:
On 01/18/2013 08:37 PM, Alex Rousskov wrote:
On 01/18/2013 08:14 AM, Steve Hill wrote:
The attached patch enables autoconsumption of the BodyPipe if access
isn't allowed in order to discard the request body rather than letting
it fill the buff
On 22/01/2013 5:53 a.m., Rainer Weikusat wrote:
Amos Jeffries writes:
On 21/01/2013 9:21 a.m., Rainer Weikusat wrote:
The attached patch removes the timeout checking code from
ConnOpener::connect and
[...]
the definitions of *F and *prt need to be inside their
relevant if() scopes.
+1
With HTTP/1.1 functionality Variants are much more common and the ICP
failure rate is growing quite high.
I propose altering the default behaviour of peering to using HTCP
queries by default.
This would entail adding "htcp=off" and "icp" options to cache_peer and
altering the build defaults
On 23/01/2013 11:52 p.m., Kinkie wrote:
Hi all,
the attached patch unpacks the unsigned int:1 declarations for
flags in PortCfg and turns them to unpacked bools.
--
/kinkie
This clashes with a patch I have nearly ready separating the flags into
groups and renaming for clarty. Can we
On 24/01/2013 4:30 a.m., Kinkie wrote:
This seems to be droppign the small mkstemp optimization makign use of its
FD return result as the already open file.
I think we should be calling umask() before before the whole #if segment and
reset it on each of "return" as well as on successful completi
On 24/01/2013 5:22 a.m., Kinkie wrote:
Hi,
this patch converts the unsinged integers in ACLUserData to unpacked bools.
--
/kinkie
Are we doing this to reduce the memory footprint? or only to simplify
the code?
This step goes some way towards memory savings by halving the memory
footp
On 24/01/2013 7:20 a.m., Kinkie wrote:
Hi all,
the attached patch turns the unsigned int:1 flags in CachePeer to bools.
--
/kinkie
Please retain the :1 bitmasking. My microbench is showing a consistent
~50ms speed gain on bitmasks over full bool, particularly when there are
multiple
On 24/01/2013 10:27 a.m., Tsantilas Christos wrote:
This is a new version of the patch which handles Amos comments.
If there is not any objection I will commit this patch to trunk...
+1.
Amos
On 24/01/2013 10:58 a.m., Alex Rousskov wrote:
On 01/18/2013 08:43 PM, Amos Jeffries wrote:
On 19/01/2013 11:53 a.m., Alex Rousskov wrote:
On 01/18/2013 05:32 AM, Amos Jeffries wrote:
On 17/01/2013 6:47 a.m., Alex Rousskov wrote:
Hello,
clientProcessRequest() assumes that a CONNECT
On 24/01/2013 11:13 a.m., Tsantilas Christos wrote:
There are cases where the generated certificates do not mimic enough
properties and secure connection with the client fails. For example,
Squid does not mimic Key Usage extensions. Clients using GnuTLS (or
similar libraries that validate server
On 24/01/2013 1:50 p.m., Alex Rousskov wrote:
On 01/23/2013 04:39 PM, Rainer Weikusat wrote:
Rainer Weikusat writes:
[...]
The downside of this approach is that this means adding
future connect timeouts will become proportionally more expensive as
more and more connections are being establis
Unless there are any further objections I shall commit this patch
tomorrow with the following merge tweaks:
* whitespace cleanup via the maintenance script.
* narrow down the enable/disable debugs messages to only be about the
value found.
Cheers
Amos
On 24/01/2013 7:51 p.m., Alex Rousskov wrote:
On 01/23/2013 07:05 PM, Amos Jeffries wrote:
On 24/01/2013 7:20 a.m., Kinkie wrote:
the attached patch turns the unsigned int:1 flags in CachePeer to
bools.
Please retain the :1 bitmasking. My microbench is showing a consistent
~50ms speed
On 25/01/2013 12:05 p.m., Rainer Weikusat wrote:
Rainer Weikusat writes:
Alex Rousskov writes:
On 01/24/2013 02:46 PM, Rainer Weikusat wrote:
In my opinion, using
a more sensibly priority queue algorithm (still both extremely simple
to implement and 'readily available' in a canned form), in
On 24/01/2013 4:50 p.m., Amos Jeffries wrote:
Unless there are any further objections I shall commit this patch
tomorrow with the following merge tweaks:
* whitespace cleanup via the maintenance script.
* narrow down the enable/disable debugs messages to only be about the
value found
On 25/01/2013 6:12 a.m., Alex Rousskov wrote:
On 01/24/2013 09:27 AM, Kinkie wrote:
From what I understand, the current consensus is to use :1'ed bools, right?
I hope the consensus is to use full bools (because they are simpler and
often faster) except where memory footprint of the structure
On 25/01/2013 12:40 p.m., Alex Rousskov wrote:
On 01/18/2013 04:54 AM, Amos Jeffries wrote:
On 18/01/2013 6:47 p.m., Alex Rousskov wrote:
Will comm_close() break if we call it instead of reaching into Comm guts
and manipulating its tables? If yes, what is Comm missing? Perhaps we
can give that
On 24/01/2013 11:36 p.m., Kinkie wrote:
On Thu, Jan 24, 2013 at 11:28 AM, Amos Jeffries wrote:
On 24/01/2013 5:56 a.m., Kinkie wrote:
Hi Amos,
what is the best way to signal in a mempools client that the
constructor is complete and that thus zeroing is not necessary? It's
not immedi
On 25/01/2013 3:06 p.m., Alex Rousskov wrote:
Hello,
The attached patch fixes several ConnOpener problems by relying on
AsyncJob protections while maintaining a tighter grip on various I/O and
sleep states. It is in PREVIEW state because I would like to do more
testing, but it did pass basi
On 25/01/2013 3:24 p.m., Rainer Weikusat wrote:
Amos Jeffries writes:
[...]
Agreed. To cut this all short. I was under the impression that we had
proven the usefulness of add-and-remove model.
Under the current code that 'remove' would be an event delete ("good
enough
On 25/01/2013 5:34 p.m., Alex Rousskov wrote:
On 01/24/2013 06:37 PM, Amos Jeffries wrote:
/* When one connection closes, give accept() a chance, if need be */
Comm::AcceptLimiter::Instance().kick();
This one might kick us. (sorry bad pun).
Heh! AFAICT, this kick() might deprive
On 24/01/2013 8:52 a.m., Alex Rousskov wrote:
Hello,
I noticed that "tcp://host:port" logger does not get "rotate logs"
and other commands that the "daemon:foo" helper gets. The TCP logger
just gets the logging line, without any prefix to accommodate future
commands or extensions. This seem
On 26/01/2013 10:39 a.m., Tsantilas Christos wrote:
Hi all,
While we are working in a project we seen that the helpers do wrong
usage of the Notes interface.
This is maybe it is our mistake (and mostly my mistake) because we do
not have make good documentation of this interface.
In any case I b
On 26/01/2013 6:35 a.m., Alex Rousskov wrote:
On 01/24/2013 09:35 PM, Amos Jeffries wrote:
On 25/01/2013 3:06 p.m., Alex Rousskov wrote:
NP: This is way beyond the simple fixes Ranier was working on. The
changes here relys on code behaviour which will limit the patch to trunk
or 3.3. I was a
On 26/01/2013 6:56 a.m., Alex Rousskov wrote:
On 01/24/2013 07:27 PM, Amos Jeffries wrote:
On 24/01/2013 11:36 p.m., Kinkie wrote:
On Thu, Jan 24, 2013 at 11:28 AM, Amos Jeffries wrote:
On 24/01/2013 5:56 a.m., Kinkie wrote:
Hi Amos,
what is the best way to signal in a mempools client
This patch has passed the 10-day standown period and has no audit objects.
I only note that it uses class member dynamic array definitions (foo[];)
definitions. The C++ feature will break build on some compilers we need
to support (clang for FreeBSD9).
Once that is fixed this patch can go in.
On 27/01/2013 8:27 a.m., Tsantilas Christos wrote:
On 01/26/2013 06:05 AM, Amos Jeffries wrote:
On 26/01/2013 10:39 a.m., Tsantilas Christos wrote:
Hi all,
While we are working in a project we seen that the helpers do wrong
usage of the Notes interface.
This is maybe it is our mistake (and
On 28/01/2013 12:04 a.m., Tsantilas Christos wrote:
On 01/26/2013 12:11 PM, Amos Jeffries wrote:
This patch has passed the 10-day standown period and has no audit objects.
I had forgot it. Thank you.
I only note that it uses class member dynamic array definitions (foo[];)
definitions. The C
On 28/01/2013 6:32 a.m., Tsantilas Christos wrote:
On 01/27/2013 07:12 AM, Amos Jeffries wrote:
On 27/01/2013 8:27 a.m., Tsantilas Christos wrote:
On 01/26/2013 06:05 AM, Amos Jeffries wrote:
On 26/01/2013 10:39 a.m., Tsantilas Christos wrote:
I did not found such cases in the existing code
Coverity has detected a long list of small defects in the snmp_core.cc code.
Lots of alignment, sizeof(), and memory uninitialized.
They all stem from a few uses of "sizeof(name)" where name is a pointer
to an OID object rather than the object itself. This code is present as
far back as squid-2
Alex and co., I'd like your feedback on this please...
The issue outlined very clearly in this squid-users mail:
http://www.squid-cache.org/mail-archive/squid-users/201301/0358.html
is that ICAP server was broken and producing a stupid response.
On 1/28/2013 11:24 AM, Michael Graham wrote:
ICA
On 29/01/2013 11:50 p.m., Reiner Karlsberg wrote:
I am new here, so please be kind :-)
In case, the reader considers "Egoless Programming" an "out of date"
principle, then simply ignore this mail.
But in case, the reader of this mail can detect some interesting
issues, a discussion about it i
On 30/01/2013 2:47 p.m., Rainer Weikusat wrote:
What's the point of putting these onto the event queue? Wouldn't this
be better done with something like this?
The Call queue and the Events queue are drained individually in
alternating fashion. The first 0-delay event will not run until after
On 30/01/2013 2:07 p.m., Eliezer Croitoru wrote:
This topic was discused couple times and I want to get back to it with some
highlights.
Since we want to maximize cache efficiency we must apply something that will
allow partial content caching.
The situation now is kind of all or nothing.
One o
On 30/01/2013 9:21 a.m., Alex Rousskov wrote:
On 01/29/2013 01:49 AM, Tianyin Xu wrote:
3. For configuration like A B=C, A and B should be sensitive while C
should be insensitive.
The case sensitivity of squid.conf directive parameters (i.e., all the
stuff that follows the directive name) depe
On 30/01/2013 10:31 a.m., Rainer Weikusat wrote:
Rainer Weikusat writes:
[...]
There are probably all kinds of 'weird, stylistic issues' and
possible, yet unknown bugs,
Attached is a slightly updated version of this. Mostly, it fixes an
accidentally inverted comparison in down_heap, uses the
On 31/01/2013 6:28 a.m., Alex Rousskov wrote:
On 01/30/2013 09:30 AM, Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/28/2013 03:29 PM, Rainer Weikusat wrote:
so easy that even using the STL implementation wouldn't be worth the
effort
Can you clarify why we should not just use one of the
On 31/01/2013 2:11 a.m., Eliezer Croitoru wrote:
On 1/30/2013 8:48 AM, Amos Jeffries wrote:
Does not matter. For starters we want to be *able* to store any
cacheable response.
Since I dont know the exact progress of collapse forwarding.
The request contains the desire to fetch partial data so
On 31/01/2013 3:12 p.m., Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/30/2013 09:30 AM, Rainer Weikusat wrote:
Alex Rousskov writes:
On 01/28/2013 03:29 PM, Rainer Weikusat wrote:
so easy that even using the STL implementation wouldn't be worth the
effort
Can you clarify why we shoul
On 29/01/2013 6:32 a.m., Alex Rousskov wrote:
Hello,
The attached patch fixes several ConnOpener problems by relying on
AsyncJob protections while maintaining a tighter grip on various I/O and
sleep states. This patch is identical to the last ConnOpener patch
posted in PREVIEW state. I am n
On 31/01/2013 6:44 p.m., Alex Rousskov wrote:
On 01/30/2013 06:30 PM, Amos Jeffries wrote:
I've kind of gone off the idea of leaving old code using a wrapper API
with the old behaviour. All that we have got out of that approach is
added technical debt for others whoc have to ome
On 1/02/2013 5:17 p.m., Alex Rousskov wrote:
Hello,
Many SslBump deployments try to minimize potential damage by _not_
bumping sites unless the local policy demands it. Unfortunately, this
decision must currently be made based on very limited information: A
typical HTTP CONNECT request does
ETA for 3.3.1 as of this writing is 9th Feb.
That said, we still have over 60 unresolved major bugs from older
versions affecting 3.3.
Amos
On 2/02/2013 8:27 a.m., Alex Rousskov wrote:
On 02/01/2013 12:06 PM, Kinkie wrote:
BTW: c++11 defines intXX_t as part of the core language; I believe
that to be an indication that the language will evolve in that
direction.
I sure hope C++ does not evolve in the direction of uint64_t! :-)
Jus
On 2/02/2013 1:42 p.m., Eliezer Croitoru wrote:
On 2/2/2013 2:35 AM, Eliezer Croitoru wrote:
I was trying to access my /squid-internal-mgr/* and it seems to have a
problem.
The visible host name is www1.home ip 192.168.10.1
I have the proper acls to allow manager access and I get:
1359765266.4
On 28/01/2013 2:07 p.m., Amos Jeffries wrote:
Coverity has detected a long list of small defects in the snmp_core.cc
code.
Lots of alignment, sizeof(), and memory uninitialized.
They all stem from a few uses of "sizeof(name)" where name is a
pointer to an OID object rather than
On 4/02/2013 7:54 a.m., Eliezer Croitoru wrote:
On 2/3/2013 8:29 PM, Alex Rousskov wrote:
The cause of this specific problem is_not_ rock behavior (correct or
not) but a mismatch between a startup script and squid.conf -- the
script does not test for all the right directories when running squid
On 5/02/2013 8:48 a.m., Alex Rousskov wrote:
On 02/03/2013 10:06 PM, Alex Rousskov wrote:
On 02/03/2013 08:31 PM, Amos Jeffries wrote:
On 2/3/2013 8:29 PM, Alex Rousskov wrote:
To be consistent with ufs, we should probably change rock behavior from
initializing the database to doing nothing
http://bugs.squid-cache.org/show_bug.cgi?id=3686
The maximum cached object size gets capped to the highest specified
max-size of any cache_dir, even if there is cache_dir available with no
max-size specified.
The solution:
Re-document the global as a default limit with cache_dir specific val
This audit round all appear to be documentation an debugs polish. Any
objection to me applying those changes and comitting?
Amos
On 8/02/2013 5:23 a.m., Alex Rousskov wrote:
On 02/06/2013 08:45 PM, Amos Jeffries wrote:
This audit round all appear to be documentation an debugs polish. Any
objection to me applying those changes and comitting?
None from me, especially if you can remove code duplication that Eliezer
cannot
On 8/02/2013 10:07 p.m., Tianyin Xu wrote:
It's not wrong but when I see the code I want to refine it since it
looks uncomfortable.
680 if (Config.Announce.period > 0) {
681 Config.onoff.announce = 1;
682 } else if (Config.Announce.period < 1) {
683 Config.Announce.p
On 6/02/2013 8:14 a.m., Alex Rousskov wrote:
On 02/05/2013 03:36 AM, Amos Jeffries wrote:
http://bugs.squid-cache.org/show_bug.cgi?id=3686
+if (static_cast(newMax) > maxSize()) {
+debugs(47, DBG_PARSE_NOTE(2), "WARNING: object 'max-size' option for " <
On 9/02/2013 5:58 p.m., Alex Rousskov wrote:
On 02/08/2013 05:54 PM, Amos Jeffries wrote:
-virtual int64_t maxObjectSize() const { return max_objsize; }
+int64_t maxObjectSize() const;
Please do not remove "virtual". It is a useful reminder that this is a
part of the Store
On 2/02/2013 12:01 a.m., Kinkie wrote:
Hi all,
I'm done with the bool-ification effort (only pending is the
TrafficMode merge from Amos).
This is now done. But the PortCfg flags are still not bools. It seems
you will need to re-do that part of it.
Amos
On 11/02/2013 9:18 a.m., Eliezer Croitoru wrote:
On 2/8/2013 11:46 AM, Amos Jeffries wrote:
On 8/02/2013 5:23 a.m., Alex Rousskov wrote:
On 02/06/2013 08:45 PM, Amos Jeffries wrote:
This audit round all appear to be documentation an debugs polish. Any
objection to me applying those changes
On 11/02/2013 6:02 a.m., Kinkie wrote:
Hi all,
the attached patch aims to c++-ify and standardize a bit more mime.cc
and the auxiliary classes it contains.
No functional changes.
It'd be nice to change the mime table from being a table of regexes to
a suffix search - after all, all mime extens
On 12/02/2013 1:37 a.m., Kinkie wrote:
* the string created by tempnam() apparently needs to be xfree()'d when
done.
doh! Ok.
* please add a comment in the #else condition noting the tempnam() is
officially obsolete from POSIX-2008.1.
(in a few more years we should drop the #else code enti
On 11/02/2013 9:13 a.m., Eliezer Croitoru wrote:
On 2/2/2013 6:23 AM, Amos Jeffries wrote:
On 2/02/2013 1:42 p.m., Eliezer Croitoru wrote:
On 2/2/2013 2:35 AM, Eliezer Croitoru wrote:
Sorry missing part.
When I am doing it using as forward proxy and use the url to the
intercept port 3127
On 13/02/2013 12:01 a.m., Eliezer Croitoru wrote:
On 2/12/2013 1:25 AM, Amos Jeffries wrote:
What are via and visible_hostname configured to in your Squid? (or in
your /etc/resolv.conf for hostname)
Amos
visible_hostname www1.home
via off
Ah. Squid detects loops by looking for its visible
The hierarchy_stoplist directive has been a rarely needed and redundant
directive for some time now.
It's useful operation is also performed by this following config:
acl stoplist url_regex cgi-bin \?
always_direct allow stoplist
Any objections?
Amos
On 13/02/2013 11:33 a.m., Henrik Nordström wrote:
tis 2013-02-12 klockan 14:41 -0700 skrev Alex Rousskov:
Hello,
Could somebody with better authentication and helper knowledge clarify
whether the token field is indeed required for Nagotiate ERR and NA
responses? If not, can we just remove the ab
On 11/02/2013 7:57 p.m., raja wrote:
Hi,
while my traffic is going through squid proxy some times browsing becomes
damn slow..after some time getting icap protocol error ..after few minutes
able to browse again.Let me know if anybody have idea about this issue
Detected protocol errors cause t
On 13/02/2013 2:37 p.m., Alex Rousskov wrote:
On 02/12/2013 03:33 PM, Henrik Nordström wrote:
tis 2013-02-12 klockan 14:41 -0700 skrev Alex Rousskov:
Could somebody with better authentication and helper knowledge clarify
whether the token field is indeed required for Nagotiate ERR and NA
respon
On 13/02/2013 3:00 p.m., Alex Rousskov wrote:
On 02/12/2013 06:41 PM, Amos Jeffries wrote:
On 13/02/2013 11:33 a.m., Henrik Nordström wrote:
tis 2013-02-12 klockan 14:41 -0700 skrev Alex Rousskov:
Hello,
Could somebody with better authentication and helper knowledge clarify
whether the token
On 13/02/2013 4:41 p.m., Amm wrote:
I had reported this bug earlier in Dec 2012 but probably went unnoticed in
squid-dev group
http://www.squid-cache.org/mail-archive/squid-dev/201212/0099.html
So just re-posting as it still exists in stable branch 3.3.1
For the record this is being tracked
On 12/02/2013 12:01 a.m., Kinkie wrote:
Since this is a refactor please fix the coding guidelines differences as
well. I have added some here:
* private variables suffixed with '_' (ie MimeIcon::url_, MimeIcon::icon_)
Done.
* please make MimeEntry a MEMPROXY class. Possibly also MimeIcon.
Do
On 13/02/2013 3:31 p.m., Amos Jeffries wrote:
On 13/02/2013 2:37 p.m., Alex Rousskov wrote:
On 02/12/2013 03:33 PM, Henrik Nordström wrote:
tis 2013-02-12 klockan 14:41 -0700 skrev Alex Rousskov:
Could somebody with better authentication and helper knowledge clarify
whether the token field is
Since we now support HTTP/1.1 storage and revalidation of
Cache-Control:no-cache it is important that we at least detect the cases
where no-cache= and private= contain parameters.
AFAIK these are still rare occurances due to the historic lack of
support. So for now Squid just detects and exemp
On 17/02/2013 8:10 p.m., Tianyin Xu wrote:
Hi, all,
I proposed to make it consistent for the case sensitivity of
configuration parsing couple weeks ago. I saw the replies from Alex,
Reiner, and Amos. All of you agreed on this (thanks a lot!). The
attached patch is for this purpose.
The patch fo
On 10/08/2012 1:57 p.m., Amos Jeffries wrote:
On 9/08/2012 12:12 a.m., Alexander Komyagin wrote:
Following patch fixes nested debug() calls problem.
Since debugs() is a macros, it should not change static Debugs::level
before putting debug message to the internal stream. Otherwise we
encounter
On 19/11/2011 9:02 a.m., Henrik Nordström wrote:
lör 2011-11-19 klockan 00:56 +1300 skrev Amos Jeffries:
This removes the domain from the server pconn key.
I would not recommend doing this.
We started out with only IP:PORT as the key, but then had to add doamin
to origin server connections
On 19/02/2013 9:22 a.m., Alex Rousskov wrote:
On 02/18/2013 04:39 AM, Amos Jeffries wrote:
On 10/08/2012 1:57 p.m., Amos Jeffries wrote:
On 9/08/2012 12:12 a.m., Alexander Komyagin wrote:
Following patch fixes nested debug() calls problem.
Since debugs() is a macros, it should not change
On 19/02/2013 6:48 a.m., Alex Rousskov wrote:
On 02/16/2013 06:51 PM, Amos Jeffries wrote:
Since we now support HTTP/1.1 storage and revalidation of
Cache-Control:no-cache it is important that we at least detect the cases
where no-cache= and private= contain parameters.
The patch appears to
A few things are on my radar that need improving in the Squid DNS
components. I am proposing these as a batch, any which we can agree on
will be scheduled for fixing in 3.4.
1) renaming all DNS options to begin with dns_* prefix.
This will allow the cfgman documentation to collate these optio
On 20/02/2013 8:45 p.m., Alex Rousskov wrote:
On 02/19/2013 05:08 PM, Amos Jeffries wrote:
A few things are on my radar that need improving in the Squid DNS
components. I am proposing these as a batch, any which we can agree on
will be scheduled for fixing in 3.4.
1) renaming all DNS options
On 18/02/2013 8:38 p.m., Ahmad wrote:
Dear all ,
ive installed centos 5.9 32 bit .
ive compiled the kernel with balabit "tproxy patch" i downloaded the kernel
& patch from balabit site .
i also patched the iptables from balabit site with verison 1.4.10.
i followed the articale of how to patch t
On 21/02/2013 10:20 a.m., Alex Rousskov wrote:
On 02/20/2013 01:46 AM, Amos Jeffries wrote:
On 20/02/2013 8:45 p.m., Alex Rousskov wrote:
On 02/19/2013 05:08 PM, Amos Jeffries wrote:
2) adapting append_domains from a string type to a custom type
This will allow us to do additional
On 21/02/2013 2:47 p.m., Alex Rousskov wrote:
Hello,
It looks like NTLM and possibly Negotiate authentication is broken
in trunk because Squid eats the bare backslash that AF responses use to
separate authentication domain and user names. With the backslash
removed, the merged domainuser na
On 21/02/2013 6:15 p.m., Alex Rousskov wrote:
On 02/20/2013 06:56 PM, Amos Jeffries wrote:
I'm only proposing for now that dns_defnames directive be enabled *if*
resolv.conf is loaded containing search directive and nothing is in
squid.conf setting it explicitly.
Yes, that would make sen
On 21/02/2013 6:41 p.m., Alex Rousskov wrote:
On 02/20/2013 07:26 PM, Amos Jeffries wrote:
On 21/02/2013 2:47 p.m., Alex Rousskov wrote:
Hello,
It looks like NTLM and possibly Negotiate authentication is broken
in trunk because Squid eats the bare backslash that AF responses use to
Adjusted patch to drop the odd NP, rework CC:private operation on broken
parameters, and fix the segfault.
Since we now support HTTP/1.1 storage and revalidation of
Cache-Control:no-cache it is important that we at least detect the
cases where no-cache= and private= contain parameters and hand
On 23/02/2013 10:20 a.m., Alex Rousskov wrote:
On 02/21/2013 06:26 PM, Amos Jeffries wrote:
Adjusted patch to drop the odd NP, rework CC:private operation on broken
parameters, and fix the segfault.
Hi Amos,
Thank you for addressing most of my concerns.
+} else if (/* p
As you may be aware OpenSSL had some API changes which we dutifully
wrote #if-#else conditional code for using the mechanisms provided by
OpenSSL for the purpose.
Then somebody in Fedora or RHEL decided to back-port the functionality
into their older OpenSSL version. This corrupted the Fedora
I'm trying to get the MacOS builds of Squid going again but having some
problems with shm_open() in the Rock storage unit-tests.
1) MacOS defines the max name length we can pass to shm_open() at 30
bytes. "/squid-testRock__testRockSearch" being 35 or so bytes.
Cutting the definition in testRo
1 - 100 of 3984 matches
Mail list logo