Re: [Scons-dev] tools and testing

[William Blevins]

Ah, right. The PR didn't show the commit for some reason.

https://bitbucket.org/scons/scons/commits/0ba0c0371c240ef6bf302e4a6ec21b9e8d487dbe?at=default

On Tue, Jun 6, 2017 at 11:43 AM, Russel Winder  wrote:
> On Fri, 2017-04-21 at 18:55 -0400, William Blevins wrote:
>> Bill/Russel,
>>
>> I explicitly resolved Issue 1924 with PR 304.
>
> There doesn't seem to be a test in test/D/Issues/1924 to check for no
> regressions, did the test go elsewhere?
>
>> http://scons.tigris.org/issues/show_bug.cgi?id=1924
>> https://bitbucket.org/scons/scons/pull-requests/304/issue-1924-update
>> d-d-scanner-per-d/diff
>>
>> Issue 1924 and 2993 are explicitly related in code area. I know this
>> because the Issues references a PR that my fix replaced.
>> http://scons.tigris.org/issues/show_bug.cgi?id=2993
>
> Again no test in test/D/Issues/2993 unless the test was names something
> else.
>
> I'll see if I can suss this one out.
>
>
>> Since I resolved Issue 1924 at the D language definition level, it is
>> likely that Issue 2993 is resolved and I just didn't realized I fixed
>> more than 1 issue simultaneously. I think the 3rd example from the
>> CHANGES.txt is exactly what Issue 1924 is asking for...
>>
>> https://bitbucket.org/scons/scons/src/b8e7614a7a9240fb15e87b0a9e7f04a
>> 9056240df/src/CHANGES.txt?at=default=file-view-default
>> - Updated D language scanner support to latest: 2.071.1. (PR #1924)
>>   https://dlang.org/spec/module.html accessed 11 August 2016
>>   - Enhancements:
>> - Added support for selective imports: "import A : B, C;" -> A
>> - Added support for renamed imports. "import B = A;" -> A
>> - Supports valid combinations: "import A, B, CCC = C, DDD = D :
>> EEE = FFF;" -> A, B, C, D
>>
>> If I accidentally resolved that issue too, just add it to the
>> CHANGES.txt.
>>
>> V/R,
>> William
>>
>
> --
> Russel.
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Russel Winder]

On Fri, 2017-04-21 at 18:55 -0400, William Blevins wrote:
> Bill/Russel,
> 
> I explicitly resolved Issue 1924 with PR 304.

There doesn't seem to be a test in test/D/Issues/1924 to check for no
regressions, did the test go elsewhere?

> http://scons.tigris.org/issues/show_bug.cgi?id=1924
> https://bitbucket.org/scons/scons/pull-requests/304/issue-1924-update
> d-d-scanner-per-d/diff
> 
> Issue 1924 and 2993 are explicitly related in code area. I know this
> because the Issues references a PR that my fix replaced.
> http://scons.tigris.org/issues/show_bug.cgi?id=2993

Again no test in test/D/Issues/2993 unless the test was names something
else.

I'll see if I can suss this one out.


> Since I resolved Issue 1924 at the D language definition level, it is
> likely that Issue 2993 is resolved and I just didn't realized I fixed
> more than 1 issue simultaneously. I think the 3rd example from the
> CHANGES.txt is exactly what Issue 1924 is asking for...
> 
> https://bitbucket.org/scons/scons/src/b8e7614a7a9240fb15e87b0a9e7f04a
> 9056240df/src/CHANGES.txt?at=default=file-view-default
> - Updated D language scanner support to latest: 2.071.1. (PR #1924)
>   https://dlang.org/spec/module.html accessed 11 August 2016
>   - Enhancements:
> - Added support for selective imports: "import A : B, C;" -> A
> - Added support for renamed imports. "import B = A;" -> A
> - Supports valid combinations: "import A, B, CCC = C, DDD = D :
> EEE = FFF;" -> A, B, C, D
> 
> If I accidentally resolved that issue too, just add it to the
> CHANGES.txt.
> 
> V/R,
> William
> 

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[William Blevins]

It might be. I submitted a fix for 1924 as PR 304. That should have been
merged...

On Apr 21, 2017 2:17 PM, "Bill Deegan"  wrote:

:)
Sounds good.

On Fri, Apr 21, 2017 at 10:17 AM, Russel Winder 
wrote:

> On Fri, 2017-04-21 at 09:36 -0700, Bill Deegan wrote:
> > Is this bug resolved?
> >
> > http://scons.tigris.org/issues/show_bug.cgi?id=2993
> >
> >
>
> I suspect not given the name of the reporter. :-)
>
> I'd have to spend a bit of time properly researching the report,
> creating a test, and then putting a solution in place. I can give that
> a go in a couple of weeks. It is though nothing to do with the current
> pull request, it has to be a separate piece of work.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>

___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Bill Deegan]

:)
Sounds good.

On Fri, Apr 21, 2017 at 10:17 AM, Russel Winder 
wrote:

> On Fri, 2017-04-21 at 09:36 -0700, Bill Deegan wrote:
> > Is this bug resolved?
> >
> > http://scons.tigris.org/issues/show_bug.cgi?id=2993
> >
> >
>
> I suspect not given the name of the reporter. :-)
>
> I'd have to spend a bit of time properly researching the report,
> creating a test, and then putting a solution in place. I can give that
> a go in a couple of weeks. It is though nothing to do with the current
> pull request, it has to be a separate piece of work.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Russel Winder]

On Fri, 2017-04-21 at 09:36 -0700, Bill Deegan wrote:
> Is this bug resolved?
> 
> http://scons.tigris.org/issues/show_bug.cgi?id=2993
> 
> 

I suspect not given the name of the reporter. :-)

I'd have to spend a bit of time properly researching the report,
creating a test, and then putting a solution in place. I can give that
a go in a couple of weeks. It is though nothing to do with the current
pull request, it has to be a separate piece of work.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Bill Deegan]

Is this bug resolved?

http://scons.tigris.org/issues/show_bug.cgi?id=2993



On Fri, Apr 21, 2017 at 2:29 AM, Russel Winder  wrote:

> On Wed, 2017-04-19 at 10:34 -0700, Bill Deegan wrote:
> > Plenty of existing end to end tests function with mocked out external
> > tools.
> > Any reason this can't be done for your pull request?
>
> The current pull request doesn't need such a thing as it doesn't
> include the dub tool. I believe the tests in place do cover the new
> builder. I am just doing the documentation side of things – the D tool
> XML files were inadequate, so as requested I am getting them 'ship
> shape' as part of this pull request.
>
> I'll address the issue of executable discovery in the D tools as a
> separate pull request, to avoid interference with the addition of the
> new builder.
>
> When I get to the dub pull request, I hope to have tests in place that
> cover all this – I am just getting onto doing proper tests for the dub
> tool. All the email exchanges have been very helpful on that front.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Russel Winder]

On Wed, 2017-04-19 at 10:34 -0700, Bill Deegan wrote:
> Plenty of existing end to end tests function with mocked out external
> tools.
> Any reason this can't be done for your pull request?

The current pull request doesn't need such a thing as it doesn't
include the dub tool. I believe the tests in place do cover the new
builder. I am just doing the documentation side of things – the D tool
XML files were inadequate, so as requested I am getting them 'ship
shape' as part of this pull request.

I'll address the issue of executable discovery in the D tools as a
separate pull request, to avoid interference with the addition of the
new builder.

When I get to the dub pull request, I hope to have tests in place that
cover all this – I am just getting onto doing proper tests for the dub
tool. All the email exchanges have been very helpful on that front.

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Bill Deegan]

Plenty of existing end to end tests function with mocked out external tools.
Any reason this can't be done for your pull request?

On Tue, Apr 18, 2017 at 10:18 PM, Russel Winder 
wrote:

> On Tue, 2017-04-18 at 20:48 -0700, Bill Deegan wrote:
> > Would it be possible to mock the dub tool for a unit test?
> > (Or even for a E2E test?)
>
> I had sort of arrived at having to do that anyway to avoid the need for
> the dub program on the test server. I think we have to assume the Dub
> tool always does exactly what is says it does and so ignore it for all
> the tests. If I extract the subprocess call into a separate Python
> function, that enables mocking. For testing the basic stuff this should
> be OK. I have created a special case of the unit-threaded package
> because it needs a special builder for writing suite files. Most of
> that is already amenable to unit test, but not end-to-end unless dub is
> available.
>
> I will set about putting these tests in place, but given the SCons test
> infrastructure a full end-to-end system test may have to be left to
> what happens in the field.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Russel Winder]

On Tue, 2017-04-18 at 20:48 -0700, Bill Deegan wrote:
> Would it be possible to mock the dub tool for a unit test?
> (Or even for a E2E test?)

I had sort of arrived at having to do that anyway to avoid the need for
the dub program on the test server. I think we have to assume the Dub
tool always does exactly what is says it does and so ignore it for all
the tests. If I extract the subprocess call into a separate Python
function, that enables mocking. For testing the basic stuff this should
be OK. I have created a special case of the unit-threaded package
because it needs a special builder for writing suite files. Most of
that is already amenable to unit test, but not end-to-end unless dub is
available.

I will set about putting these tests in place, but given the SCons test
infrastructure a full end-to-end system test may have to be left to
what happens in the field. 

-- 
Russel.
=
Dr Russel Winder  t: +44 20 7585 2200   voip: sip:russel.win...@ekiga.net
41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc
Description: This is a digitally signed message part
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Bill Deegan]

Would it be possible to mock the dub tool for a unit test?
(Or even for a E2E test?)


On Tue, Apr 18, 2017 at 8:13 PM, Russel Winder  wrote:

> On Tue, 2017-04-18 at 09:24 -0700, Bill Deegan wrote:
> > Unitesting is there to make sure that no unintention breakage occurs.
> > I'm not willing to take a major piece of code into SCons core without
> > tests.
> >
> > That would violate SCons's practices.
>
> Excellent position, I have no problems with that. For tools I think the
> end-to-end tests are essential, and the SCons test framework very
> usable.
>
> The question is though that if the _Library type is not exposed is
> there actually a test obligation if all the publicly exposed things
> (the builders) are exercised by end-to-end tests.
>
> Having said this I am still working on a test for fetching a package
> from a local area and installing on a local area so it fits in the test
> framework, but it doesn't really emulate what happens in a real case,
> so it is not clear to me what is actually being tested, and whether the
> test is actually useful. My worry is that such a test simply tests
> whether Dub does what it says it does, which is nothing to do with
> SCons and the dub tool per se.
>
> On the other hand there are now function in the dub tool that could use
> a bit of testing for confidence sake. I am trying to work out how best
> to do that with an end to end perspective.
>
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev

Re: [Scons-dev] tools and testing

[Bill Deegan]

Unitesting is there to make sure that no unintention breakage occurs.
I'm not willing to take a major piece of code into SCons core without tests.

That would violate SCons's practices.

-Bill

On Tue, Apr 18, 2017 at 1:17 AM, Russel Winder  wrote:

> On Mon, 2017-04-17 at 08:26 -0700, Bill Deegan wrote:
> > Russel,
> >
> > Seems like pieces of it (including a new Node type) should be unit
> > testable.
> > Also, Can Dub pull from a local repository, or is there a stable, but
> > small
> > web accessible repository?
>
> I agree the _Library class should be amenable to some unittesting,
> though it is not clear how much would be actually useful and how much
> would just be testing to the sake of having tests. Influenced by Cope,
> I am less enamoured of unit tests than I used to be but totally for
> end-to-end tests. I will though be adding some tests for Dub now that
> it is working for the ApproxGC project.
>
> Dub can pull from a local repository as well as the central one, I
> shall have to find out if the central one can be excluded, I think it
> can. In this case it would be possible to construct a local repository
> and then run an end-to-end test which is good.
>
> This leaves only the question of which project to construct a package
> for. I am thinking that this could be another tool in the D armoury:
> build the local code and upload a package to the Dub repository.
> --
> Russel.
> 
> =
> Dr Russel Winder  t: +44 20 7585 2200   voip:
> sip:russel.win...@ekiga.net
> 41 Buckmaster Roadm: +44 7770 465 077   xmpp: rus...@winder.org.uk
> London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
>
> ___
> Scons-dev mailing list
> Scons-dev@scons.org
> https://pairlist2.pair.net/mailman/listinfo/scons-dev
>
>
___
Scons-dev mailing list
Scons-dev@scons.org
https://pairlist2.pair.net/mailman/listinfo/scons-dev