Smalltalk › Pharo › Pharo Smalltalk Developers

[update] #10487

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

25 messages Options

Mariano Martinez Peck

Re: [update] #10487

On Mon, Oct 19, 2009 at 2:59 PM, Michael Roberts <[hidden email]> wrote:

I think we should report the pharo tests the same way we report the
pharo-core tests. Unless someone disagrees I will look to split out
the wiki into a number of test areas. I agree we can not necessarily
fix all the tests owned by the packages but we should publish a
baseline. This will be important for inclusion of new packages, a
prerequisite being not introducing new failures.

+ 1

thanks
Mike

On Mon, Oct 19, 2009 at 5:31 PM, Adrian Lienhard <[hidden email]> wrote:
> We have to check the packages that are part of the Pharo image before
> we release 1.0. Currently we have 40 failures and 55 errors, which is
> simply not good enough. There is code loaded that can not work
> (considering there are 38 undeclared references...).
>
> Adrian
>
> On Oct 19, 2009, at 17:08 , Damien Cassou wrote:
>
>> 2009/10/19 Mariano Martinez Peck <[hidden email]>:
>>> What I propose is to identify the red tests of Pharo dev and look
>>> for the
>>> package and ask the maintainers (if there is anyone) to fix it. And
>>> the
>>> regarding tests, we can tag them as 1.1
>>
>> Ok. Can you have a look at that please? The problem is that we will
>> also have to maintain these tests.
>>
>>
>>>>> Moreover, if packages are included in Pharo dev it is because it
>>>>> is a
>>>>> Pharo decision. So, if pharo thinks that, for example, ocompletion
>>>>> must be in Pharo, so, I think that ocompletion tests must be green.
>>>>
>>>> But I don't think that the package lists (web and dev images) are
>>>> blessed by the Pharo board.
>>>>
>>> And I am not agree with this also. I think Pharo board, and even the
>>> community has to choose the Pharo packages. Please, take in mind
>>> that the
>>> final users will use and see Pharo, NOT PHARO CORE.
>>
>> I confirm. The Pharo user community has the last word on what gets
>> integrated into images I generate, not me. The list of packages is at:
>> http://code.google.com/p/pharo/wiki/Packages.
>>
>> --
>> Damien Cassou
>> http://damiencassou.seasidehosting.st
>>
>> "Lambdas are relegated to relative obscurity until Java makes them
>> popular by not having them." James Iry
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Dale

Re: [update] #10487

In reply to this post by Stéphane Ducasse

----- "Stéphane Ducasse" <[hidden email]> wrote:

| ideally we should have a miniimage and a metacello configuration files and
| people can pick what they want. + an universe based on metacello.
|
| Stef
|

Stef,

I like the idea of a Universe based on metacello ... A central repository where configurations can be published will be really useful, otherwise google is your only friend:)... A user could either load the target project or load the configuration of the project to view or modify to fit their needs ...

Dale

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Stéphane Ducasse

Re: [update] #10487

In reply to this post by Lukas Renggli

Lukas

What I would like ideally is that I get a miniimage + a configuration
+ a **FAST** package
loading so that we could load all the tests each time we integrate
then we could have the tests not in the image.

I think that there are different scenarios (dev my app vs changing the
system)
and I can tell you that they are change like cleaning image segment
for which I would
love to have tests because I would be more confortable with fixing it.

Stef

>>> Personally I would like to download an image *without* any tests, no
>>> matter if this is a core or a dev image. And have included a
>>> detailed
>>> report on the results of running all the tests in that particular
>>> image.
>>
>> But why, apart from the tiny image disk usage reduction. The test
>> sometimes show how to use the, most times underdocumented, package.
>
> IMHO, that's a myth. Did a single test case ever in your entire life
> help you to understand a piece of code? It never happened to me so far
> :-)
>
> The rational behind my opinion is that I do not care about the tests
> of the other packages when I develop software. Did you ever (other
> than by accident) run the Collection or Compiler tests when you
> developed some of your code?
>
> Btw, the same is true for the Monticello working copies. I frankly do
> not care what version of Collection or Compiler is loaded in my image.
> Sometimes I just drop all this information before I load my code. This
> greatly reduces clutter.
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Lukas Renggli

Re: [update] #10487

> I think that there are different scenarios (dev my app vs changing the
> system)
> and I can tell you that they are change like cleaning image segment
> for which I would
> love to have tests because I would be more confortable with fixing it.

Yeah definitely, I completely agree. The tests are very valuable and
should be run regularly. It is just for me as a normal user/developer
(of Seaside) is not really interested in running the collection tests
all the time. I am more interested in what tests pass (to chose the
right image, to know the limitations) than in the tests themselves.

Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Stéphane Ducasse

Re: [update] #10487

yeap!
100% in sync.

On Oct 19, 2009, at 9:57 PM, Lukas Renggli wrote:

>> I think that there are different scenarios (dev my app vs changing
>> the
>> system)
>> and I can tell you that they are change like cleaning image segment
>> for which I would
>> love to have tests because I would be more confortable with fixing
>> it.
>
> Yeah definitely, I completely agree. The tests are very valuable and
> should be run regularly. It is just for me as a normal user/developer
> (of Seaside) is not really interested in running the collection tests
> all the time. I am more interested in what tests pass (to chose the
> right image, to know the limitations) than in the tests themselves.
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project