what happened to the "tools/regression" directory - and the stuff in it?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
27 messages Options
12
Reply | Threaded
Open this post in threaded view
|

what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
The tools in the tools/regression directory are critical to the testing of the serialization library on one's local machine.  What happened to these.  

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Rene Rivera-2
On Wed, Mar 18, 2015 at 3:07 PM, Robert Ramey <[hidden email]> wrote:

> The tools in the tools/regression directory are critical to the testing of
> the serialization library on one's local machine.  What happened to these.
>

They went here.. <https://github.com/boostorg/regression>.

What is critical and how?


--
-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams - http://robot-dreams.net
-- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
Rene Rivera-2 wrote
On Wed, Mar 18, 2015 at 3:07 PM, Robert Ramey <[hidden email]> wrote:

> The tools in the tools/regression directory are critical to the testing of
> the serialization library on one's local machine.  What happened to these.
>

They went here.. <https://github.com/boostorg/regression>.
What is critical and how?
I used them to generate at massive table which detailed all results by build variant.  So they were critical for me.  Since no one else complains I presume that I'm the only one that use them.

So my question is - how is one supposed to invoke this functionality in the current setup?

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Rene Rivera-2
On Wed, Mar 18, 2015 at 3:24 PM, Robert Ramey <[hidden email]> wrote:

> Rene Rivera-2 wrote
> > What is critical and how?
>
> I used them to generate at massive table which detailed all results by
> build
> variant.  So they were critical for me.  Since no one else complains I
> presume that I'm the only one that use them.
>

Probably :-)

So my question is - how is one supposed to invoke this functionality in the
> current setup?
>

I don't know.. Since you have told me what the functionality is
specifically. What did you used to build/run? And how?



--
-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams - http://robot-dreams.net
-- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
Rene Rivera-2 wrote
So my question is - how is one supposed to invoke this functionality in the
> current setup?

I don't know.. Since you have told me what the functionality is
specifically. What did you used to build/run? And how?
I've used the procedure described here http://www.boost.org/doc/libs/1_57_0/tools/regression/doc/library_status.html

Obviously something equivalent to this is necessary for every library author.  What do other library authors use?  What do you recommend for testing on one's own machine?

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Steven Watanabe-4
AMDG

On 03/18/2015 02:44 PM, Robert Ramey wrote:
>
> I've used the procedure described here
> http://www.boost.org/doc/libs/1_57_0/tools/regression/doc/library_status.html
>
> Obviously something equivalent to this is necessary for every library
> author.  What do other library authors use?  What do you recommend for
> testing on one's own machine?
>

I just use b2 | less

In Christ,
Steven Watanabe


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Rene Rivera-2
In reply to this post by Robert Ramey
On Wed, Mar 18, 2015 at 3:44 PM, Robert Ramey <[hidden email]> wrote:

> Rene Rivera-2 wrote
> > So my question is - how is one supposed to invoke this functionality in
> > the
> >> current setup?
> >
> > I don't know.. Since you have told me what the functionality is
> > specifically. What did you used to build/run? And how?
>
> I've used the procedure described here
>
> http://www.boost.org/doc/libs/1_57_0/tools/regression/doc/library_status.html


OK. I'll see about updating that to where the code for that went to. And
making sure it's working as best as I can. Won't be soon though.

Also, note that the code for that tool is now at <
https://github.com/boostorg/regression/tree/develop/library_status>. And
IIRC it's updated to build if you clone/get the regression repo. But I
never tried using the resulting executable. And process_jam_log is at <
https://github.com/boostorg/regression/tree/develop/testing>.

Obviously something equivalent to this is necessary for every library
> author.  What do other library authors use?  What do you recommend for
> testing on one's own machine?
>

Like Steven I just use b2 directly and read the output. But then I have a
very simple library to deal with. So I'm not a good example.


--
-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams - http://robot-dreams.net
-- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
In reply to this post by Steven Watanabe-4
Steven Watanabe-4 wrote
I just use b2 | less
LOL -  for 250 tests with multiple compilers/configurations?  and re-run b2 every time you want to go back and see a previous result?

Seriously - how do other people deal with this?

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Andrey Semashev-2
On Wednesday 18 March 2015 14:09:41 Robert Ramey wrote:
> Steven Watanabe-4 wrote
>
> > I just use b2 | less
>
> LOL -  for 250 tests with multiple compilers/configurations?  and re-run b2
> every time you want to go back and see a previous result?
>
> Seriously - how do other people deal with this?

I don't have 250 tests, and b2 >1.txt works just fine for me. I run tests on
one compiler (typically, gcc) before pushing to develop and let the testers do
the rest. If I suspect the change can affect compatibility with a specific
compiler (typically, msvc) I may run the tests on that compiler as well. That
is a rare case, though.


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Steven Watanabe-4
In reply to this post by Robert Ramey
AMDG

On 03/18/2015 03:09 PM, Robert Ramey wrote:
> Steven Watanabe-4 wrote
>> I just use b2 | less
>
> LOL -  for 250 tests with multiple compilers/configurations?  and re-run b2
> every time you want to go back and see a previous result?
>
> Seriously - how do other people deal with this?
>

I was completely serious.  None of my libraries
have more than 100 tests, and I don't commit
changes at all until the tests are completely
clean on all the toolsets I'm testing with.
There's a lot of output, but I can ignore most
of it and just search for "failed."  When I'm
debugging I generally only run one toolset at
a time.  When I run multiple toolsets, it's almost
always a last check that everything's okay before
I commit.

In Christ,
Steven Watanabe


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
In reply to this post by Rene Rivera-2
Rene Rivera-2 wrote
Also, note that the code for that tool is now at <
https://github.com/boostorg/regression/tree/develop/library_status>. And
IIRC it's updated to build if you clone/get the regression repo. But I
never tried using the resulting executable. And process_jam_log is at <
https://github.com/boostorg/regression/tree/develop/testing>.

Obviously something equivalent to this is necessary for every library
> author.  What do other library authors use?  What do you recommend for
> testing on one's own machine?
>

Like Steven I just use b2 directly and read the output. But then I have a
very simple library to deal with. So I'm not a good example.
In my source tree there was a folder icon named regression and I just clicked it and the folder appeared in the right place in my modular-boost tree.  Which I suppose is OK - except that there's another corner of git that I have to figure out how to use - this is a burden to me.

When I first saw this I presumed that the boost test python scripts had been enhanced to provide this sort of facility in one's local environment.  Now I'm poking around to see how all this works.  There is a directory with 15K lines of *.py code it !!!

Robert Ramey

Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Michael Caisse-3
In reply to this post by Robert Ramey
On 03/18/2015 02:09 PM, Robert Ramey wrote:

> Steven Watanabe-4 wrote
>> I just use b2 | less
>
> LOL -  for 250 tests with multiple compilers/configurations?  and re-run b2
> every time you want to go back and see a previous result?
>
> Seriously - how do other people deal with this?
>
> Robert Ramey
>
>

We have 230 top level tests in Spirit. I use:

   b2 > out.txt

a simple:

   grep failed out.txt | wc -l

tells me how many failed. I can then go and look at details in the output.

michael

--
Michael Caisse
ciere consulting
ciere.com

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Michael Caisse-3
On 03/19/2015 12:30 PM, Michael Caisse wrote:

> On 03/18/2015 02:09 PM, Robert Ramey wrote:
>> Steven Watanabe-4 wrote
>>> I just use b2 | less
>>
>> LOL -  for 250 tests with multiple compilers/configurations?  and
>> re-run b2
>> every time you want to go back and see a previous result?
>>
>> Seriously - how do other people deal with this?
>>
>> Robert Ramey
>>
>>
>
> We have 230 top level tests in Spirit. I use:
>
>    b2 > out.txt
>
> a simple:
>
>    grep failed out.txt | wc -l


Well....

    grep "^\*\*passed" out.txt | wc -l

will get the passed ones. I have actually have some post processing that
I do for the automated Bamboo builds to show failures differently. The
above expression wont actually work for what you want. Adjust the grep
expression to meet your needs.  (o;




_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

John Maddock-2

>>
>> a simple:
>>
>>    grep failed out.txt | wc -l
>
>
> Well....
>
>    grep "^\*\*passed" out.txt | wc -l
>
> will get the passed ones. I have actually have some post processing
> that I do for the automated Bamboo builds to show failures
> differently. The above expression wont actually work for what you
> want. Adjust the grep expression to meet your needs.  (o;
>

grepping for "failed" shows the "failed as expected" messages as well.

I grep for

\.\.failed

which works for me.

John.

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Edward Diener-3
In reply to this post by Rene Rivera-2
On 3/18/2015 5:18 PM, Rene Rivera wrote:

> On Wed, Mar 18, 2015 at 3:44 PM, Robert Ramey <[hidden email]> wrote:
>
>> Rene Rivera-2 wrote
>>> So my question is - how is one supposed to invoke this functionality in
>>> the
>>>> current setup?
>>>
>>> I don't know.. Since you have told me what the functionality is
>>> specifically. What did you used to build/run? And how?
>>
>> I've used the procedure described here
>>
>> http://www.boost.org/doc/libs/1_57_0/tools/regression/doc/library_status.html
>
>
> OK. I'll see about updating that to where the code for that went to. And
> making sure it's working as best as I can. Won't be soon though.

Thanks !

I agree with Robert on this even if I had not used local regression
testing but just executed b2 manually many times with different
compilers for my libraries. I test on quite a few compilers and
different versions largely on Windows and less so on Linux. Having an
automated procedure that can test all the compilers I want to use in
their various configurations at once, and then show me the regression
tests results locally, seems much better than what I do manually, even
if though I have my own command files both on Windows and Linux to make
things easier for myself.

>
> Also, note that the code for that tool is now at <
> https://github.com/boostorg/regression/tree/develop/library_status>. And
> IIRC it's updated to build if you clone/get the regression repo. But I
> never tried using the resulting executable. And process_jam_log is at <
> https://github.com/boostorg/regression/tree/develop/testing>.
>
> Obviously something equivalent to this is necessary for every library
>> author.  What do other library authors use?  What do you recommend for
>> testing on one's own machine?
>>
>
> Like Steven I just use b2 directly and read the output. But then I have a
> very simple library to deal with. So I'm not a good example.
>
>



_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
In reply to this post by Robert Ramey
thanks for all the helpful grep scripts.

are you guys arguing that this is more convenient than the following:

http://www.boost.org/doc/libs/1_54_0/tools/regression/doc/library_status.html

In any case, It's not a huge deal for me as I'm the only one using it I'll just move it to my own repository.

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Steven Watanabe-4
AMDG

On 03/19/2015 06:57 PM, Robert Ramey wrote:
> thanks for all the helpful grep scripts.
>
> are you guys arguing that this is more convenient than the following:
>
> http://www.boost.org/doc/libs/1_54_0/tools/regression/doc/library_status.html
>

  I'm not claiming that my method is better, but
it works well enough, and I don't think that
using this tool would give any significant benefit
for me.  If it helps you, that's great.

> In any case, It's not a huge deal for me as I'm the only one using it I'll
> just move it to my own repository.
>

In Christ,
Steven Watanabe


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Rene Rivera-2
In reply to this post by Robert Ramey
On Wed, Mar 18, 2015 at 4:52 PM, Robert Ramey <[hidden email]> wrote:

> Rene Rivera-2 wrote
> > Also, note that the code for that tool is now at <
> > https://github.com/boostorg/regression/tree/develop/library_status>. And
> > IIRC it's updated to build if you clone/get the regression repo. But I
> > never tried using the resulting executable. And process_jam_log is at <
> > https://github.com/boostorg/regression/tree/develop/testing>.
> >
> > Obviously something equivalent to this is necessary for every library
> >> author.  What do other library authors use?  What do you recommend for
> >> testing on one's own machine?
> >>
> >
> > Like Steven I just use b2 directly and read the output. But then I have a
> > very simple library to deal with. So I'm not a good example.
>
> In my source tree there was a folder icon named regression and I just
> clicked it and the folder appeared in the right place in my modular-boost
> tree.  Which I suppose is OK - except that there's another corner of git
> that I have to figure out how to use - this is a burden to me.
>
> When I first saw this I presumed that the boost test python scripts had
> been
> enhanced to provide this sort of facility in one's local environment.  Now
> I'm poking around to see how all this works.  There is a directory with 15K
> lines of *.py code it !!!
>

Note.. I've been cleaning up the regression tree to reduce that sort of
thing. But of course things take a long time to accomplish as I can't
interrupt testing and the release process. Anyway.. I do want to make your
job as easy as possible in this regard so I do plan to make the
library_status table generation easy to use. But..

I was planning on working on this a bit this weekend and it's not going to
happen. My dev machine is having problems and looks like it's close to
imminent death. So I'm doing backups, ordering a new machine, and then
setting up that new machine from the backups. So It's going to be a much
longer time before I get to this again :-(

--
-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams - http://robot-dreams.net
-- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Robert Ramey
Rene Rivera-2 wrote
So It's going to be a much longer time before I get to this again :-(
I've got my own copy and no one else uses it so it's really a non-problem for me.  Don't spend any measurable time on this.

Robert Ramey
Reply | Threaded
Open this post in threaded view
|

Re: what happened to the "tools/regression" directory - and the stuff in it?

Edward Diener-3
On 3/20/2015 2:07 PM, Robert Ramey wrote:
> Rene Rivera-2 wrote
>> So It's going to be a much longer time before I get to this again :-(
>
> I've got my own copy and no one else uses it so it's really a non-problem
> for me.  Don't spend any measurable time on this.

Could you share your copy via Git ?



_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
12