On Thu, Jan 19, 2012 at 3:51 PM, Don Clugston <dclugston at googlemail.com>wrote:

> And stage 3 is a bit random. We apparently have a few projects which
> must be tested (dsimcha's, for example)
> but I think most things that are tested are open source, and could
> therefore be automated.
>

Maybe SciD and Dstats should be used for automated testing of DMD.  While neither was written to be a compiler test suite, they empirically are ridiculously good at finding compiler bugs, probably because of the coding style which heavily uses the newest/most advanced features of D.  Both are open source and on Github.  Dstats has good unittest coverage and SciD's has improved lately, too.  We might also consider other projects like these (open source, good unittest coverage, empirically seems to frequently expose compiler bugs) to add to the automated testing system.  Does anyone want to suggest one?

In principle anyone could do the testing I do.  I just do it because I'm familiar with the relevant codebases and it would be very hard to break these libs into discrete test cases to add to the real DMD test suite as opposed to testing them as-is.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/dmd-internals/attachments/20120119/d82d04e4/attachment.html>

On 19 January 2012 22:01, David Simcha <dsimcha at gmail.com> wrote:
> On Thu, Jan 19, 2012 at 3:51 PM, Don Clugston <dclugston at googlemail.com> wrote:
>>
>> And stage 3 is a bit random. We apparently have a few projects which
>> must be tested (dsimcha's, for example)
>> but I think most things that are tested are open source, and could
>> therefore be automated.
>
>
> Maybe SciD and Dstats should be used for automated testing of DMD.? While neither was written to be a compiler test suite, they empirically are ridiculously good at finding compiler bugs, probably because of the coding style which heavily uses the newest/most advanced features of D.? Both are open source and on Github.? Dstats has good unittest coverage and SciD's has improved lately, too.? We might also consider other projects like these (open source, good unittest coverage, empirically seems to frequently expose compiler bugs) to add to the automated testing system.? Does anyone want to suggest one?
>
> In principle anyone could do the testing I do.? I just do it because I'm familiar with the relevant codebases and it would be very hard to break these libs into discrete test cases to add to the real DMD test suite as opposed to testing them as-is.

I imagine:
Maintain a list of projects. Where to get it, how to build it.

Download, build with latest version. pass -> green light
if fails, build with previous version.  pass -> red light.
Otherwise, assume project is broken. Ignore and move onto next project.

Are we still going to use trello?? I set up a card for my current phobos activity.? Andrei noted that he's dropping it...

Just so I can voice my opinion, it has potential, and seeing as how Walter doesn't do much work on Phobos, do we need Walter's approval to use it for Phobos only?? If nothing else, I see it being a useful tool for managing the proposed improvements.? Every few weeks someone says "what's next for approval in phobos, anybody know?"? If there were a board to manage that, it would be useful.

-Steve



>________________________________
> From: Steve Schveighoffer <schveiguy at yahoo.com>
>To: Discuss the phobos library for D <phobos at puremagic.com>
>Sent: Thursday, January 19, 2012 8:50 AM
>Subject: Re: [phobos] [dmd-internals] Planning software?
> 
>https://trello.com/schveiguy
>
>
>
>----- Original Message -----
>> From: Andrei Alexandrescu <andrei at erdani.com>
>> To: Discuss the phobos library for D <phobos at puremagic.com>
>> Cc:
>> Sent: Tuesday, January 17, 2012 3:16 PM
>> Subject: Re: [phobos] [dmd-internals] Planning software?
>> 
>> On 1/17/12 11:34 AM, Brad Roberts wrote:
>>>? IMHO, we've got plenty of tools in our mix we just don't use them
>> well.
>>> 
>>>? Bugzilla already has several dimensions of scheduling, including severity,
>> priority, and milestones.? Additionally, it's
>>>? probably way overdue to apply more structure to the products/components.? I
>> know it's something Don has advocated before.
>>> 
>>>? Wiki is a good document collaboration tool.
>>> 
>>>? The DLang website is a good communication tool, as are the
>> newsgroups/mailing lists.
>>> 
>>>? Between bugzilla for fine grained categorization of
>> issues/enhancements/whatever and a page on the website to state high
>>>? level goals, I'm not sure what else we really need.
>>> 
>>>? I absolutely agree that we need to have a better roadmap and to execute on
>> it, just not that another tool is needed or
>>>? useful.
>> 
>> I understand your concern (expanded in your other messages). Let's, however, not derail the discussion into a false choice. Yes, we do have tools we don't use. That doesn't mean we should stop looking at _other_ tools that may be more adequate tools for our needs.
>> 
>> So the proof that we don't use bugzilla votes or priorities does not prove e.g. trello.com would not be useful.
>> 
>> I think we have two good proposals. One is trello.com, the other is github's new tools. The github chain does not seem very mature, but integration and potential are definitely important.
>> 
>> Could you guys make trello.com accounts so we find each other there? Let's give it a whirl.
>> 
>> 
>> Thanks,
>> 
>> Andrei
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos
>> 
>_______________________________________________
>phobos mailing list
>phobos at puremagic.com
>http://lists.puremagic.com/mailman/listinfo/phobos
>
>
>

On Thursday, January 19, 2012 15:54:04 Steve Schveighoffer wrote:
> Are we still going to use trello? I set up a card for my current phobos activity. Andrei noted that he's dropping it...

It's looking like the general response to Andrei's push was to ask why bugzilla isn't good enough, which has resulted in the discussion focusing entirely on tools rather than the organizational issues that the tools would be used to solve. And after a heated discussion on the issue, Andrei appears to have given up.

I haven't looked at trello yet, but I don't see any reason not to give it a whirl if it seems like it could be useful.

- Jonathan M Davis

On 1/19/12 5:54 PM, Steve Schveighoffer wrote:
> Just so I can voice my opinion, it has potential, and seeing as how Walter doesn't do much work on Phobos, do we need Walter's approval to use it for Phobos only?  If nothing else, I see it being a useful tool for managing the proposed improvements.  Every few weeks someone says "what's next for approval in phobos, anybody know?"  If there were a board to manage that, it would be useful.

I think that's a good idea, though the impact will be smaller. Are the Phobos majors up for trying this?

Andrei

On Thursday, January 19, 2012 18:05:03 Andrei Alexandrescu wrote:
> I think that's a good idea, though the impact will be smaller. Are the Phobos majors up for trying this?

https://trello.com/jmdavis

- Jonathan M Davis

Don Clugston, el 19 de enero a las 21:51 me escribiste:
> Here's some concrete suggestions:

I think this is a very concrete and achievable plan.

> A beta is released. We wait for m days for regression reports.
> Any regressions result in another beta. After p days without further
> reports, a release is made.

I would only add just a very small suggestion. Numbering betas, tagging them and maybe even adding milestones to them. This way bug reports for it can be created to make easier to organize how to fix them, and keep track of them. This would also make betas more visible, which will mean that more people will test them.

-- 
Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
----------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------
El techo de mi cuarto lleno de galaxias