With the move to ats 10 and the cmake build the question came up if we should
continue to support the traffic_top feature to do remote logic to trafficserver
instances to get information.
We currently have a PR https://github.com/apache/trafficserver/pull/10814 to
remove the checking and need
I should add that the records.cfg is the only current way to load a cfg
file with a different name. The runroot and the env vars only allow us to
point to a different directory with the predefined names.
still +1 for this, however, there might be a functional break for some
people because of this
+1
On Fri, Nov 1, 2019 at 9:42 AM Sudheer Vinukonda
wrote:
> +1
>
>
> On Nov 1, 2019, at 6:52 AM, Leif Hedstrom wrote:
> >
> > Hi all,
> >
> > I’d like to remove all (most?) of the settings that lets you override
> configuration file names via records.config. We have other means to
> accomplis
I really think we should be doing a c and c++ API. Allen had a PR at one
point to move some source around to make this easier to do.
Jason
On Fri, May 17, 2019 at 10:55 AM Leif Hedstrom wrote:
>
>
> > On May 17, 2019, at 9:48 AM, Walt Karas
> wrote:
> >
> > But are there people who write plugi
+1
On Fri, May 10, 2019 at 6:46 PM Bryan Call wrote:
> I would like to propose that we move from the IRC to Slack for IM
> communication by June 1st. This is in response to ASF moving to Slack for
> Infra and other channels and spamming that has happened on the IRC over the
> last year.
>
> Thi
This sound more like something to add to the core autest code. ie an env
var that says don't run this test ( or some set of tests)
I would like to avoid if we can adding a condition for flow control to each
test as this becomes a mess really quickly
Jason
On Fri, Apr 12, 2019 at 10:38 AM Walt K
I was thinking of https://en.cppreference.com/w/cpp/thread/scoped_lock
Jason
On Thu, Mar 21, 2019 at 4:53 PM Walt Karas
wrote:
> What's the name of it in the Standard Lib?
>
> On Thu, Mar 21, 2019 at 4:45 PM Jason Kenny .invalid>
> wrote:
>
> > I am unclear why
I am unclear why it is needed. C++ 17 already has this in the std library.
these should be used instead of making our own
Jason
On Wed, Mar 20, 2019 at 4:17 PM Leif Hedstrom wrote:
> Don’t we have scoped mutexes already ?
>
> — Leif
>
> > On Mar 20, 2019, at 12:08, Walt Karas
> wrote:
> >
> >
We can modify the default config file to have this value set for all tests.
On Mon, Feb 4, 2019 at 6:43 PM Walt Karas wrote:
> I meant the output that is (optionally) generated by TSDebug() calls
> in trafficsever.
>
> On Mon, Feb 4, 2019 at 5:52 PM Jason Kenny
> wrote:
>
which debug output? autest or curl?
On Mon, Feb 4, 2019 at 4:29 PM Pushkar Pradhan
wrote:
> Walt,
> What is your proposal, is it to run all tests with debug output?
> This would require modifying all existing tests. That also brings up
> another question, what sections should be enabled?
> I.e.
Having more output will slow stuff down. However, that is not the main
issue.
I had pointed out before various testability issues with ATS. We have a
number of tests that have to wait for an amount of time for the test to
work. These delays are absolute, in that we have to wait the full time
given
I do plan to add a switch to allow one to error on condition testing. This
should be easy to do. However I do report which test are skipped and why (
and if that reason is not good enough one can help fix the test as most if
not all these should be reason that can be provided with the condition
cal
nd = 'python
> tcp_client.py' syntax.
> That would continue to pull the correct python executable right?
>
> On Thu, Nov 8, 2018 at 2:24 PM Jason Kenny
> wrote:
>
> > I am looking at cleaning up the logic to "setup" what is needed to run
> > autest agai
I am looking at cleaning up the logic to "setup" what is needed to run
autest against traffic server.
The current logic "works" well but when it does not. It tries to install
python 3.5 if it is not installed. microserver needs to have python 3.5 or
better to work currently. I would like to move t
Hi,
from the logs I think you want to look at the _sandbox directory to see if
the files there show anything. What is odd is that you are failing on
returns codes such as 52 and 35. This might be a sign of something else.
- Jason
On Mon, Oct 29, 2018 at 12:21 PM 毛建峰 wrote:
> add the logs file
There are a few items in my mind that could be fixed.
1) There is a massive complex build knot of depends in this code. Changing
it would clean this up
2) The current code is still hard to follow. the YAML change would make it
easier to understand, follow and maintain
3) independent of the "protoc
I am just odd... but I search on the commit message. ie the one value in
the top of the PR page. It has always worked for me.. except when ( which
we generally don't have) we have a lot of commits with the same message. In
these cases, we get a set of commits back and I can check which one I want
This method works.. it is a lot easier if used with namespaces as the "set"
of changed objects are wrapped in a versioned namespace. This is a lot
easier for a user to map to, then having a lot of different versioned
objects.
The end result is that the user will map to a space like ats_8 that is a
I think this is a great move. +1 all the way.
Jason
On Tue, Jun 5, 2018 at 10:19 AM, Leif Hedstrom wrote:
> Hi all,
>
> I’ve prepared a patch that moves as much as was easily moved out of
> proxy/*.{cc,h} to cmd/traffic_server, the PR is
>
> https://github.com/apache/trafficserver/pull/
community. making complex as it scares
people away to other projects
Jason
On Wed, May 23, 2018 at 4:25 PM, Leif Hedstrom wrote:
>
>
> > On May 23, 2018, at 3:20 PM, Jason Kenny
> wrote:
> >
> >>> 1) I don’t like the “src” top level directory, that implies th
>>1) I don’t like the “src” top level directory, that implies that all
source is there (which is clearly not the case)
We cannot move everything at once. We have to do stuff in small steps.
I would really like to get all the source under an src/ dir to solve a
number of stupid build problems we h
Is the concern bufferwritter or the use fo bufferwritter in TSDebug. I
agree the "extra" value is small for TSDebug. I feel the use of
bufferwritter is great within our code base.
-Jason
On Thu, Apr 19, 2018 at 5:42 PM, Bryan Call wrote:
> Replacing Debug()/TSDebug() with BufferWriter/bwformat
Hi all,
Derek pointed out an issue he has found in When.PortOpen(...) function that
he found is causing issues with some tests. The issue is that I made a
change to PortOpen() to check for the state that the given port is open,
via using netstat like logic. The previous revision of the code was to
I this we could a lot of value out of TBB. The tasking engine is much
better than what we have. However I would not suggest we just replace
everything, but instead bring it in and look at "low-hanging" fruit such as
the hash map that we can get value out of today. Larger projects as using
its advan
interesting to not the junction make will use TBB under the covers...
for example:
https://github.com/preshing/junction/blob/14c0b8f835d26899c22eb7e747ccdffb775c49f2/junction/extra/impl/MapAdapter_TBB.h
On Fri, Sep 29, 2017 at 5:13 PM, Bryan Call wrote:
> Here are some benchmark numbers I ran ac
Seriously... I am fine with one big PR, but it will be big. (adding one line,
removing 3 for each header) I think it is easier to do smaller batches
Jason
On Monday, July 31, 2017, 3:19:33 PM CDT, Jason Kenny
wrote:
So we should make a massive PR to fix all headers?
This is not some ploy to
trying to see if clang-tidy had anything that
supports this but I could not find it.
On Mon, Jul 31, 2017 at 12:08 PM Jason Kenny
wrote:
> I don't see the need for making more complex configure checks... this is a
> supported feature of all compilers that supports C++11.
> Jason
>
ng saved by
> include guards.
Then let's add a configure check to enforce that the compiler supports this and
make a sweep over the whole codebase. I would like to avoid the kind of
never-ending transition we have seen with this kind of thing in the past.
>
>
>
> On Mond
practice by most
modern C++ people. However, this may not be so. So I want to bring the
discussion on this again, for those that want to use inefficient #ifdef guards
over the more modern #prgama once replacement.
Jason
On Friday, July 7, 2017, 2:07:14 PM CDT, Jason Kenny
wrote:
Hi all,
I would
No AuTest is most likely not broken. It is showing something else is wrong.
Jason
On Sunday, July 23, 2017, 1:59:32 PM CDT, James Peach wrote:
Hi all,
I am unable to get autest tests to run succesfully on my Fedora26 box. I’m
running the ci/regression script, which builds and runs all the test
I am making a PR for this
Jason
On Sunday, July 9, 2017, 4:03:44 PM CDT, Leif Hedstrom wrote:
> On Jul 9, 2017, at 8:00 AM, Alan Carroll
> wrote:
>
> I have PR 2254 which is a conversion of IpMapTest from our internal
> regression test / TestBox to Catch. Overall I think Catch is clearly be
Hi all,
I would suggest we start using
#pragma once
instead of the classic header guard of
#ifndef MYHEADER_H
#define MYHEADER_H
...#endif //MYHEADER_H
The reason for this is that #pragma once:
1) is supported on all the compilers and all the platform combination we use
for quite some time now.
Hi all,
There was a talk about using gtest or something else for unit testing. I had
suggested Catch instead, however, I think everyone without looking or trying
anything out said gtest. So to help get a real test written I have various PR
coming with unit tests. I had my interns write tests in
Friday, June 16, 2017, 9:18:34 AM CDT, Jason Kenny
wrote:
If you are using Visual Studio Code. It should just find the setup.cfg file
that controls the formatting.
Jason
On Thursday, June 15, 2017, 6:21:07 PM CDT, Masaori Koshiba
wrote:
Hi trafficserver-dev,
Now, all python code of ATS are
If you are using Visual Studio Code. It should just find the setup.cfg file
that controls the formatting.
Jason
On Thursday, June 15, 2017, 6:21:07 PM CDT, Masaori Koshiba
wrote:
Hi trafficserver-dev,
Now, all python code of ATS are formatted by autopep8 after some
discussions on some PRs (#2
location where
other code exists for supporting testing... which is currently in tests/tools
Jason
On Friday, June 2, 2017 03:19:52 PM CDT, Jason Kenny
wrote:
I am fine with tests/plugins.
We just need to make sure the code is documented to say what it tests. Some
cases will be
James,
to help clarify what we are doing. There currently is an issue in which autest
cannot correct test certain builds of ATS. The issue comes from the need to
point to a build and copy/hard link bits in to a sandbox directory to run the
test in a controlled environment. Todo this we use TS_RO
I am fine with tests/plugins.
We just need to make sure the code is documented to say what it tests. Some
cases will be one-off plugins while other will be reusable
Jason
On Friday, June 2, 2017, 3:17:52 PM CDT, Steven R. Feltner
wrote:
+1 for tests/plugins. It should cut down on the confusio
I still think catch will work better.. it is lighter and easier to make tests
for no binaries needed
On Thursday, May 25, 2017 02:23:02 PM CDT, Leif Hedstrom
wrote:
> On May 24, 2017, at 10:17 PM, James Peach wrote:
>
>
>
> On 24 May 2017, at 15:00, Alan Carroll wrote:
I am not against this.
I would only point out that from experience in the past I found many C++
developers like the catch library better. https://github.com/philsquared/Catch
Jason
On Wednesday, May 24, 2017 04:38:13 PM CDT, Bryan Call
wrote:I am in favor of using GTest fo
>>The tests drag in traffic_server because of the link-time dependencies
>>of the systems they are testing, not because of the test framework.
This is exactly right. I would go on by saying how I want to clean this mess
upThe main point we have to understand is that the current build logic
I would suggest for unit testing.
https://github.com/philsquared/Catch
This has been nice to use. This does not require use to sync a bunch of
binaries to get it to work.
I think most the of the plugins will test better with AuTest. There will still
be a need to unit test APIs in trafficserver or
.lua.cfg
Jason
On Monday, February 27, 2017, 9:13:33 AM CST, Alan Carroll
wrote:.luaconf
I have to say this is an odd argument.
Anyone that does real world code in enterprise environments knows that you have
to migrate between versions. Config value that are renamed have to allow for
time migrate to a new value. Hard switches will upset customers very quickly.
They need time to be w
Hi all,
I am about ready to make a pull request for the updated testing system to
replace the existing TSQA.
BUT Before I do I would like some others to give it this a run locally on
your
setup. It would be very help for the effort to improve out testing if I
could get some help and other ey
I think the only real thing I could add to this is that it would be nice if we
could use the same internal C++ class as part of the C++api, vs wrapping
another class around the C API.
Jason
From: Leif Hedstrom
To: dev@trafficserver.apache.org
Sent: Saturday, June 11, 2016 4:56 PM
Subj
Have to remember the joy of e-mail at yahoo Let me try to get a good link to
catch
https://github.com/philsquared/Catch
Jason From: Jason Kenny
To: "dev@trafficserver.apache.org"
Sent: Friday, April 22, 2016 9:56 AM
Subject: Re: Unit testing for plugin
I am planning on
I am planning on proposing the use of philsquared/Catch
|
|
|
| ||
|
|
|
||
philsquared/Catch
Catch - A modern, C++-native, header-only, framework for unit-tests, TDD and
BDD | |
|
|
for unit testing as part of my work on fixing up testing ATS within Ya
Yep that did the trick
Jason
- Original Message -
From: Jason Kenny
To: "dev@trafficserver.apache.org" ; Alan Carroll
Sent: Thursday, January 28, 2016 11:34 AM
Subject: Re: clang format
Ahh.. I did not use the magic version. Let me try to get this on my system
correctl
anuary 28, 2016 11:24 AM, Jason Kenny
wrote:
Hi,
I was updating a patch by running clang-format. However when I ran this I
noticed that I have a lot of files that are now different (50+ files). Should
this be happening. I ran this on fedora 23. I would have thought I should only
see changes in
Hi,
I was updating a patch by running clang-format. However when I ran this I
noticed that I have a lot of files that are now different (50+ files). Should
this be happening. I ran this on fedora 23. I would have thought I should only
see changes in the one file I modified, ie taffic_via.cc.
Original Message -
From: Yongming Zhao
To: dev@trafficserver.apache.org; Jason Kenny
Cc: Alan Carroll
Sent: Wednesday, January 27, 2016 11:42 AM
Subject: Re: Proposal for how to update source code layout.
this thread is a big evolution change in TS if it happen, glad to see the
discuss right now o
I think what you are missing is that the decoupling is more complex than a few
file tweaks. There is code in the wrong place. Instead of trying to move
everything around in iocore,proxy,lib,mgmt and cmd, which will be only
confusing when looking at different version of ATS, is to make a clean
s
more about it. However for me the issue is that having
options to only makes everything better and stronger. At the moment I think the
current setup is not doing this.
Jason
From: Phil Sorber
To: dev@trafficserver.apache.org; Jason Kenny
Sent: Tuesday, January
we have a clear understand of why this is a needed pre-step to a number of
other wanted and needed improvements to ATS.
Jason
From: James Peach
To: dev@trafficserver.apache.org; Jason Kenny
Sent: Tuesday, January 26, 2016 12:47 PM
Subject: Re: Proposal fo
correctly I really believe we have to clean up the
source layout. I want to get this done first and as quick as possible to reduce
the pain caused by the source moves that would upset existing ( and future)
pull requests.
Jason
From: Phil Sorber
To: dev@trafficserver.apache.org; Jason Kenny
Hi everyone,
Second try as the first attempt seems to be a blob of unformated text.
I want to see about moving forward on the source clean up I believe we all
agreed to the last ATS summate. What I would like to get general agreement on
what I plan to do to the source layout so it is clear w
Hi everyone, I want to see about moving forward on the source clean up I
believe we all agreed to the last ATS summate. What I would like to get general
agreement on what I plan to do to the source layout so it is clear why and
where everything will go before I start making lots of small pull re
Hi,
I have not built with asan with clang on the mac. I have built for RHEL 6 and
7. I had to configure with these options.
CXXFLAGS -fno-omit-frame-pointer -fsanitize=address -static-libasan\
CFLAGS="-fno-omit-frame-pointer -fsanitize=address
-static-libasan"\--disable-freelist\LUA_LDFLAGS="-fno
59 matches
Mail list logo
