MAYBE this could be mitigated by providing RPMs for RHEL6/CentOS 6
that are built on a custom server with a recent gcc but older
libraries?
What do you guys think?
On Wed, Nov 5, 2014 at 1:27 PM, Marcus Kool marcus.k...@urlfilterdb.com wrote:
On 11/05/2014 02:01 AM, Amos Jeffries wrote:
You want to ask this question to the polygraph authors; squid users
and developers are not really the right place to ask.
I don't know about the details, but that error message probably means
that your bots are overwhelming the TCP stack, which can't free
ephemeral ports quickly enough due to TCP
There seems to be some extraneous code (documentation updates, formatting, etc).
Is it intentional?
On Mon, Nov 10, 2014 at 2:53 PM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Most of the uses of CbDataList appear to be abusing it for regular
list
root@rs-ubuntu-precise:~# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:Ubuntu 12.04.5 LTS
Release:12.04
Codename: precise
root@rs-ubuntu-precise:~# gcc --version
gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3
On Mon, Nov 10, 2014 at 11:50 AM, Amos
Hi all,
I've updated the build farm a bit, adding a few nodes and
simplifying build dependencies for trunk.
3.HEAD-amd64-centos-7 is the cornerstone build; if it succeeds all
others are immediately triggered (except for east). Hopefully this
results in more optimized build times.
--
Also note that ccache is an opimization, it's not strictly needed.
On Tue, Nov 18, 2014 at 1:16 AM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 18/11/2014 12:31 p.m., Eliezer Croitoru wrote:
I wanted to build an ubuntu build node but there is
relevant?
Thanks
--
Kinkie
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Great!
Thanks :)
On Thu, Dec 18, 2014 at 2:14 PM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Thanks to the issue behind rev.13760 (Support http_access denials of
SslBump 'peeked' connections.) I intend to release a new beta approx.
20hrs from
ok, I'll recheck all jobs.
On Fri, Dec 19, 2014 at 9:57 PM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 20/12/2014 9:19 a.m., Kinkie wrote:
Hi all, currently farm build failures are notified to the NOC list.
Does it make sense? I mean
Hi,
the msntauth helper is triggering some race condition in
Makefile.am, needlessly failing many builds. Furthermore, it still
follows the obsolete authentication + authorization are one and the
same pattern.
This patch simplifies the basic/MSNT auth helper, removing its
reliance on a
:///home/kinkie/squid/workspace/trunk/
# testament_sha1: f965cabac98083ebf3b7830322685d08ae20c2be
# timestamp: 2015-01-02 11:21:21 +0100
# base_revision_id: squid...@squid-cache.org-20150102061206-\
# 7e9auwijezfkedc5
#
# Begin patch
=== modified file 'include/splay.h'
--- include/splay.h 2014-12-20
I believe this is missing several delete Foo; in the ACL
destructors. You used new() to allocate the Splay objects, there
should be matching delete() at the same owner/abstraction level.
For example:
ACLIP::~ACLIP()
{
if (data) {
data-destroy(IPSplay::DefaultFree);
);
}
I believe that the line marked with the ASCII arrow is instead meant to be
assert(fd_table[params.fd].ssl != NULL);
Before going ahead and touching it, I'd like if anyone could confirm
that I'm not misinterpreting the code intentions here.
Thanks
--
Kinkie
I also have (of course) no objection with the proposed course of
action, and share the hope that the generic crypto library will be
useable for this.
On Wed, Feb 4, 2015 at 3:02 AM, Amos Jeffries squ...@treenet.co.nz wrote:
I've been expecting this to come up for a while now, but had hoped the
iterator whose dereference
can cast to an int, and support standard begin() and end()
operations).
I'd like to get your feedback on it, before bringing it in-tree and
starting to use it.
You can find the code at https://github.com/kinkie/tst
Thanks!
--
Kinkie
Hi,
sometimes it's useful to be able to rely on debugs output in unit tests.
This patch enables that. In order to turn on debugs, the unit test .cc file must
#define ENABLE_DEBUG_SECTION {section_number}
BEFORE the #include block.
--
Francesco
enable-debugs-in-unit-tests.patch
Hi all,
trunk now requires c++11; does this mean can we now start using lambdas?
--
Francesco
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Hello,
the attached patch changes some users of splay to std::set. The aim
is to get more predictable (if not necessarily better) performance and
leverage a cleaner API resulting in more readable code.
This is a merge snapshot of branch lp:~squid/squid/replace-splay revno
13835; I will continue
keep in mind I am conducting work to replace Splay with
std::map or std::unordered_map where it makes sense (which is almost
everywhere, IMO).
On Tue, Jan 6, 2015 at 9:21 PM, Kinkie gkin...@gmail.com wrote:
Hi,
I'm working on this right now. I missed constructing several Splays
know about where this may be originating from?
--
Kinkie
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
code.
So far I've converted src/acl/Arp* and src/acl/Eui64*; before moving
on to other users I'd like to get some feedback that I am on the
correct path.
Feature-branch is at lp;~squid/squid/remove-splay ; I'm attaching the
current delta snapshot.
Thanks
--
Kinkie
replace-splay-rfc1
On Tue, Jan 13, 2015 at 3:09 AM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 13/01/2015 8:51 a.m., Kinkie wrote:
Hello, the attached patch changes some users of splay to std::set.
The aim is to get more predictable (if not necessarily better
Attached is the current patch in case it gets approved for an interim merge.
On Tue, Jan 13, 2015 at 11:19 AM, Kinkie gkin...@gmail.com wrote:
On Tue, Jan 13, 2015 at 3:09 AM, Amos Jeffries squ...@treenet.co.nz wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 13/01/2015 8:51 a.m
Hi all,
the attached bundle is a patch proposal for CompactTrie, a generic
library implementing efficient prefix matching.
It is meant to be a more feature-complete replacement for
lib/libTrie, and has the machinery to effectively replace splay when
doing dstdomain-type lookups. In fact, when
:)
Kinkie
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Hi,
I've managed to fix the FreeBSD 9.3 build node:
gcc now uses gcc-4.9; it is a bit messy in resolving needed libraries,
but it should be stable
clang has clang-3.6 available, however in order to have it work it
also requires to use libc++ in place of gcc's libstdc++, and it also
requires the
I've been testing it for a few days; +1.
On Mon, Feb 23, 2015 at 5:37 AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 24/01/2015 8:37 a.m., Amos Jeffries wrote:
Updated patch including all changes made after previous review.
Most significantly:
* rearranging the cascade of Tokenizer calls
Looks good to me.
On Fri, May 8, 2015 at 10:23 AM, Amos Jeffries squ...@treenet.co.nz wrote:
Pull the IcmpConfig object out of the global SquidConfig structure and
updates it to processing its own parse logics. Bringing it inline with
the per-component configuration design in SourceLayout and
Hi,
the following patch enables using libc++ in place of libstdc++.
This is necessary with recent clang++ on some platforms, such as
FreeBSD 9.3 and reportedly newer MacOS.
With this patch squid builds using clang-3.5 and libc++; the make
check still fails due to libcppunit having a different
On Wed, Apr 15, 2015 at 6:24 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 16/04/2015 4:11 a.m., Kinkie wrote:
Hi,
the following patch enables using libc++ in place of libstdc++.
This is necessary with recent clang++ on some platforms, such as
FreeBSD 9.3 and reportedly newer MacOS
Hi,
in order to do what you aim for you can either use an (externa) ICAP
service or an eCap adaptor (http://www.e-cap.org/).
On Thu, Jun 18, 2015 at 1:42 PM, James Hunter
james.hun...@jartechnologies.com wrote:
Hi,
I've been looking to inject special JavaScript code into every HTML page my
Hi,
judging from the latest builds (jenkins was stuck for a while) it seems
that there's serious issues with std::atomic in libc++-6.4 with clang.
Any objections to deprecating these OS/compiler combinations for trunk? gcc
works fine.
--
Kinkie
My 2c: I vote for reality; possibly with a shaming announce message; I
wouldn't even recommend logging the violation: there is nothing the
average admin can do about it. I would consider adding a shaming
comment in the release notes.
On Wed, Jun 24, 2015 at 10:12 PM, Alex Rousskov
+1 then
On Fri, Jul 3, 2015 at 7:20 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 4/07/2015 4:09 a.m., Kinkie wrote:
Hi,
my review (copy-paste from IRC)
kinkie SBuf::size_type urlLen = ...
kinkie what about using auto?
yadi auto seems to only work sometimes.
kinkie O_O
hi,
a quick googling would have brought you to
http://wiki.squid-cache.org/ConfigExamples/Intercept/IptablesPolicyRoute
Please check if the document can satisfy your needs.
On Thu, May 21, 2015 at 1:34 PM, johndunne imir...@alldunne.com wrote:
What IP tables rules are you using to redirect
Hi Alex,
it turns out that the itch is real, and this is actually going to be
useful in carrying forward the HttpHeader restructuring I'm doing in
lp:~kinkie/squid/coverity-fixes, to iterate over the list of header types
for various purposes (compacting, etc).
So I went ahead and just
On Wed, Aug 12, 2015 at 9:11 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/12/2015 09:36 AM, Kinkie wrote:
So I went ahead and just implemented it, starting from your observations.
It's a bit more complex that your proposal, as I took the chance to
implement
Hi,
Probably as a result of some too aggressive cleanup activity of mine,
some build farm nodes are to be rebuilt.
I'm taking the chance for some spring cleaning and a general update,
decommissioning fedora 20 in favor of 22 (there's an issue there with
libxml2 on clang), upgrading ICC to the
Hi all,
the attached patch (a merge request from
lp:~kinkie/squid/mempools-nozero) changes MEMPROXY_CLASS so that it doesn't
zero memory before releasing it to requestors. The constructors of all
classes using that feature were checked, so that they fully initialize all
data members and thus
On Thu, Aug 20, 2015 at 7:11 AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 20/08/2015 7:28 a.m., Kinkie wrote:
Hi all,
the attached patch (a merge request from
lp:~kinkie/squid/mempools-nozero) changes MEMPROXY_CLASS so that it
doesn't
zero memory before releasing it to requestors
This is in fact what I have tried to - sometimes not to the letter,
unfortunately, and I'm sorry for that.
On Fri, Aug 21, 2015 at 6:46 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/21/2015 04:37 AM, Kinkie wrote:
I changed that only when something in the vicinity changed
expensive, but what if these changes could be automated? It'd
be enough to run the same script on trunk and on the branch(es) and the
effort of doing a flag-day change would be next to negligible.
HERE would be a perfect POC for this way of acting.
What do you think guys?
Kinkie
AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 23/08/2015 7:21 p.m., Kinkie wrote:
Hi,
std::uniform_int_distribution is not available on any compiler in
centos-6.
We can either retire centos-6 or autoconf our way around the missing
functionality.
RHEL6.7 is supported by RH through
On Sun, Aug 23, 2015 at 7:55 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/21/2015 11:31 AM, Kinkie wrote:
On Fri, Aug 21, 2015 at 6:39 PM, Alex Rousskov wrote:
On 08/21/2015 04:37 AM, Kinkie wrote:
the attached patch does:
... a long list
On Sun, Aug 23, 2015 at 9:16 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/19/2015 11:11 PM, Amos Jeffries wrote:
* when adding a new constructor ensure the big-5 operators (ctor, dtor,
assign, emplace, emplace-assign) are all present. Even if they use =
default
IMO,
Hi all,
the attached script seems to do the job just fine.
--
Francesco
remove-here.sh
Description: Bourne shell script
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Hi,
this is the second of six patches borne out of the mempools-nozero effort.
This patch implements some cleanups (NULL-nullptr, spacing etc); it is the
least important of the whole series.
--
Francesco
mempools-nozero-cleanups.patch
Description: Binary data
Hi all,
this patch is the first (of six) aiming at decomposing into more
digestible chunks the already-posted (and mostly audited) mempools-nozero
branch.
The attached patch fixes the constructors of all MEMPROXY_CLASS users so
that they completely initialize objects, avoiding reliance on
Hi,
part 4 (of 5) patches from the mempools-nozero branch: debug
This patch transforms Debug::OutStream into a MEMPROXY_CLASS instead of
rolling its own new and delete operators. The biggest changes come from
pulling in various Makefile.am files the necessary dependencies.
I've taken the chance
Oh, and +1 from me to the version with those alterations.
Alterations implemented and merged as r14250
--
Francesco
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Hi,
this patch:
- removes unused wordlistDup
- implements proper constructors for wordlist, making it MEMPROXY_CLASS-safe
- makes its destructor private to force clients to use wordlistDestroy
instead
- implements wordlistChopHead to support the one case in the whole codebase
where a wordlist
Hi,
this is the final patch in the mempools-nozero series (for now).
It requires the already-merged part 1 (fix constructors) and at least part
of part 3 (wordlist).
It:
- implements a pass-through method in Mem::AllocatorProxy to enable
specifying to the allocator the zeroing policy
- extends
Hi,
this is part 2 of the coverity-fixes merge: SBuf hash.
This code allows to have std::unordered_maps where the key is a SBuf, in
case-sensitive and case-insensitive variants. The hashing is the same as
used in the current Squid hash implementation (from lib/hash.cc) in order
not to upset the
Hi,
this is the first of a series of patches stemming from the
coverity-fixes effort.
The aim of this first submission is to implement WholeEnum and EnumRange
and supporting unit tests.
Big-context patch attached
--
Francesco
coverity-fixes-enumiterator.patch
Description: Binary data
Merged with requested changes as r14244.
On Mon, Aug 24, 2015 at 7:10 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 25/08/2015 3:36 a.m., Kinkie wrote:
Hi,
part 4 (of 5) patches from the mempools-nozero branch: debug
This patch transforms Debug::OutStream into a MEMPROXY_CLASS
I'll answer the other points separately, but
* the iterator_type being used in static_cast before the typedef _looks_
very dangerous (even if its fine and works).
- I would place typedef up the top in their own protected area to avoid
a) depending on implicit compiler behaviour and b) future
On Mon, Aug 24, 2015 at 8:37 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/24/2015 12:08 PM, Amos Jeffries wrote:
On 25/08/2015 4:49 a.m., Kinkie wrote:
+ * behavior is undefined if the iterator
+ * is incremented (or decremented) outside the range representing valid
Hi,
the attached patch turns FtpServer::EarlyErrorKind from a C enum into
c++11 strongly-typed enum. The code did not require any change, just a bit
of search-and-replace.
--
Francesco
=== modified file 'src/servers/FtpServer.cc'
--- src/servers/FtpServer.cc 2015-08-19 10:18:02 +
+++
In fact, this all started when I was told that the kernel currently
shipping in rawhide has a problem where netinet/in.h and linux/in.h are
incompatible, and this breaks our build when netfilter is enabled, because
both get indirectly inlcluded.
IMO rawhide could be : a second-class, nice-to-have
this.
On Fri, Jun 26, 2015 at 7:31 AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 26/06/2015 5:36 a.m., Kinkie wrote:
Hi,
judging from the latest builds (jenkins was stuck for a while) it seems
that there's serious issues with std::atomic in libc++-6.4 with clang.
Any objections
merged as r14183
On Tue, Jul 28, 2015 at 12:33 AM, Amos Jeffries squ...@treenet.co.nz
wrote:
+1.
Amos
--
Francesco
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
in
the headersTable (I'd go for bools for speed, I believe we can afford to
waste 100-ish bytes per flag), and possibly switching some enums to c++11
strongly-typed enums (enum class foo) in some places.
Code has been casually run-tested, seems to be OK.
On Thu, Jul 30, 2015 at 9:44 PM, Kinkie gkin
On Wed, Aug 5, 2015 at 4:08 AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 5/08/2015 9:08 a.m., Kinkie wrote:
On Tue, Aug 4, 2015 at 2:58 PM, Amos Jeffries wrote:
On 4/08/2015 11:22 p.m., Kinkie wrote:
Hi all,
the attached patch is a build- and run-tested merge proposal for next
On Wed, Jul 29, 2015 at 11:31 PM, Amos Jeffries squ...@treenet.co.nz
wrote:
On 30/07/2015 9:10 a.m., Kinkie wrote:
Hi all,
I'm starting to work on refactoring HttpHeader to use LookupTable, and
boy that code is a mess..
Since it's going to take significant effort, I'd like to get
Hi,
the preliminary polygraph results for using a std::unordered_map with a
perfect hash (generated by gperf) to do RegisteredHeader - HttpHdr lookups
are in, and are quite disappointing.
The hash is as effective as it can be: two table lookups and simple
calculations. Yet when using it squid
On Thu, Aug 6, 2015 at 12:38 AM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/05/2015 09:33 AM, Kinkie wrote:
I crossed the topic of enumerating over iterations.
You mean iterating enum[eration]s :-)
I have my thinking cap backwards ;)
We kind-of follow (C-ish)best
the new API for the HttpHeader::parse method.
BTW: there is an issue in the profiler cachemgr report - it segfaults for
me on mac.
I'll try to check it out on Linux in the next few days.
--
Kinkie
profiler.patch
Description: Binary data
___
squid-dev
ctor and dtor ..
More to come as I cross them.
(work in progress at lp:~kinkie/squid/coverity-fixes)
--
Francesco
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Merged as revision 14210 with some minor added polish (ostream for
HttpHdrCcType and Http::HdrType).
On Wed, Aug 5, 2015 at 3:58 PM, Kinkie gkin...@gmail.com wrote:
On Wed, Aug 5, 2015 at 4:08 AM, Amos Jeffries squ...@treenet.co.nz
wrote:
On 5/08/2015 9:08 a.m., Kinkie wrote:
On Tue, Aug
I like this.
On Wed, Aug 5, 2015 at 6:24 PM, Amos Jeffries squ...@treenet.co.nz wrote:
Adds a PackableStream class which provides std::ostream semantics for
writing data to a cachemgr report.
FYI: This follows on from discussiosn back in 2011 regarding how to
restructure the cachemgr
is a reworked hash_string, so it shouldn't be any less
effective than that for our workloads.
--
Francesco
# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: kin...@squid-cache.org-20150727174103-o1cq5vr952njx2ho
# target_branch: file:///home/kinkie/squid/workspace/trunk
Hi,
It's ripped from pur own hash.cc
Accord in to documentation I've found on Stackexchange, it's hen sanctioned
way to do it. Only alternative is to pass the Hasher template argument
every time.
--
Francesco
On Jul 28, 2015 1:23 AM, Alex Rousskov rouss...@measurement-factory.com
wrote:
On
HI,
the low-hanging fruits from Coverity's analysis have been picked, now
working on somewhat more complex fixes.
The attached patch takes a hint from two benign coverity defects to:
- rework the berEncodeLoginData error handling path to be more readable
- refactor Digest auth's field lookup
On Sun, Jul 26, 2015 at 10:23 PM, Amos Jeffries squ...@treenet.co.nz
wrote:
On 27/07/2015 4:48 a.m., Kinkie wrote:
HI,
the low-hanging fruits from Coverity's analysis have been picked, now
working on somewhat more complex fixes.
The attached patch takes a hint from two benign coverity
+1. Excellent idea.
On Sat, Jul 25, 2015 at 1:57 AM, Amos Jeffries squ...@treenet.co.nz wrote:
Cache.log produced at level ALL,9 are very verbose, and tracking down
what specific section,level details to log for a shorter trace without
lost details can sometimes be tricky and time consuming.
Hi,
this patch fixes another batch of coverity warnings.
Mostly harmless, but they are good practice.
--
Francesco
# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: kin...@squid-cache.org-20150721145709-q49cty3dm84vk67q
# target_branch: file:///home/kinkie/squid/workspace
to reduce the false positives in the scenarios hihglighted by
Amos, OR
2c. rework the existing features (if possible) to allow the early
detection which is in my understanding the most touted benefit of the
proposed solution
Am I missing something obvious?
Kinkie
Hi all,
this patch, part of my renewed push to use std:: containers in place
of our hand-rolled ones, migrates basic_ncsa_auth to
std::unordered_map. It's been build- and run-tested successfully.
--
Francesco
=== modified file 'helpers/basic_auth/NCSA/basic_ncsa_auth.cc'
---
On Wed, Oct 28, 2015 at 3:16 PM, Amos Jeffries <squ...@treenet.co.nz> wrote:
> On 29/10/2015 2:05 a.m., Kinkie wrote:
>> Hi all,
>>this is the final patch of my current drive on mempools-nozero (aka
>> convert explicit pool users to MEMPROXY_CLASS).
>> Build
Hi,
I'd like to pick some low-hanging SourceFormat fruits. this patch
renames cache_manager.cc to CacheManager.cc.
--
Francesco
# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: kin...@squid-cache.org-20151109110536-tyi0em1kvypffv1x
# target_branch: ../trunk
#
One step at a time.. :)
On Mon, Nov 9, 2015 at 5:28 PM, Amos Jeffries <squ...@treenet.co.nz> wrote:
> On 10/11/2015 4:40 a.m., Kinkie wrote:
>> Hi,
>> I'd like to pick some low-hanging SourceFormat fruits. this patch
>> renames cache_manager.cc to CacheManager
.
--
Kinkie
coverity-fixes-v1.patch
Description: Binary data
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
Hi all,
this patch saves and restores errno around calls to strftime in
debugLogTime, preventing harmless errors in these calls from clobbering
errno, thus messing up xstrerror() calls within debugs().
--
Francesco
=== modified file 'src/debug.cc'
--- src/debug.cc 2015-01-13 07:25:36
On Sun, Jul 12, 2015 at 2:34 AM, Amos Jeffries squ...@treenet.co.nz wrote:
On 12/07/2015 5:42 a.m., Kinkie wrote:
Hi all,
this patch saves and restores errno around calls to strftime in
debugLogTime, preventing harmless errors in these calls from clobbering
errno, thus messing up
On Sun, Jul 12, 2015 at 10:20 AM, Amos Jeffries squ...@treenet.co.nz
wrote:
On 12/07/2015 7:03 p.m., Kinkie wrote:
I tried it. Unfortunately it fails for those cases of debugs() which do
not
use xstrerror, e.g.
../../src/esi/Element.h: In member function 'virtual bool
ESIElement
Meh I mucked up there. THat wont work due to same reason for not using
xstrerror().
Use a const int xerrno = errno; first, then inside the debugs( ...
xstrerr(xerrno) ...
From my reading of the docs, there is no guarantee that errno be int; only
that it is an integral type.
I'm using const
Looks like a sensible plan to me.
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
On Fri, Jul 10, 2015 at 3:40 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 11/07/2015 1:21 a.m., Kinkie wrote:
On Fri, Jul 10, 2015 at 2:39 PM, Amos Jeffries squ...@treenet.co.nz
wrote:
On 10/07/2015 3:18 a.m., Kinkie wrote:
Hi,
I'm going through the issues identified by Coverity
Hi,
the attached patch implements this suggestion renaming to SBufHashCommon_
(it seems to be a common pattern) and adding a comment.
On Mon, Aug 24, 2015 at 8:51 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/24/2015 11:13 AM, Kinkie wrote:
this is part 2
there may be the need of small cleanups (e.g.
ssl/certificate_db.cc) but these should be viable for a regular patch.
--
Kinkie
remove-here.sh
Description: Bourne shell script
here-removal.patch.gz
Description: GNU Zip compressed data
___
squid-dev
On Thu, Aug 27, 2015 at 9:28 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/27/2015 08:58 AM, Kinkie wrote:
this patch turns StoreEntry into MEMPROXY_CLASS, getting rid of the
default constructors.
In the process, it turns hash_link into a class, giving it a default
On Fri, Aug 28, 2015 at 3:02 AM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/27/2015 03:04 AM, Kinkie wrote:
+ * zeroth, first, second, onePastLast, fourth
Please make onePastLast last.
I'm trying to highlight the fact
On Sun, Aug 30, 2015 at 12:07 AM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/29/2015 11:11 AM, Kinkie wrote:
I fail to understand the reason for this section of code in
bootstrap.sh
# do not bundle with the huge standard license text
The reason is explained
On Mon, Aug 24, 2015 at 3:14 AM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/23/2015 04:01 PM, Kinkie wrote:
Adding comment: will reallocate if ilen happens to equal pool size
I would recommend not adding a comment at all if no comment and the
above low-level
termination.
I've tried checing the patch, and noticed nothing obvious in it that could
cause the behavior.
Kinkie
___
squid-dev mailing list
squid-dev@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-dev
kkjuOn Aug 28, 2015 7:24 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/28/2015 09:02 AM, Kinkie wrote:
On Thu, Aug 27, 2015 at 9:28 PM, Alex Rousskov wrote:
On 08/27/2015 08:58 AM, Kinkie wrote:
this patch turns StoreEntry into MEMPROXY_CLASS, getting rid
Hi,
the attached patch sets up a new autoconf test to detect whether the
compiler supports c++11's override keyword, and defines a
generally-available OVERRIDE preprocessor macro which expands to
override if support is present and nothing if it isn't.
This patch is necessary to support icc
Hi,
implemented, r14262.
On Wed, Aug 26, 2015 at 4:18 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 27/08/2015 12:05 a.m., Kinkie wrote:
Hi,
the attached patch sets up a new autoconf test to detect whether the
compiler supports c++11's override keyword, and defines a
generally
On Wed, Aug 26, 2015 at 9:44 PM, Amos Jeffries squ...@treenet.co.nz wrote:
On 27/08/2015 6:19 a.m., Alex Rousskov wrote:
On 08/26/2015 11:04 AM, Kinkie wrote:
v2 patch attached.
FYI: If you have missed my last email on this thread, it is archived at
http://lists.squid-cache.org
On Mon, Aug 24, 2015 at 9:41 PM, Alex Rousskov
rouss...@measurement-factory.com wrote:
On 08/24/2015 01:21 PM, Kinkie wrote:
On Mon, Aug 24, 2015 at 8:37 PM, Alex Rousskov wrote:
On 08/24/2015 12:08 PM, Amos Jeffries wrote:
On 25/08/2015 4:49 a.m., Kinkie wrote
1 - 100 of 243 matches
Mail list logo