[squeak-dev] Re: [Seaside] Someone to tackle failing tests in Squeak?

David T. Lewis lewis at mail.msen.com
Thu Nov 13 02:49:06 UTC 2014

On Tue, Nov 11, 2014 at 08:11:33AM +0100, Johan Brichau wrote:
> Hi David,
> > On 11 Nov 2014, at 01:44, David T. Lewis <lewis at mail.msen.com> wrote:
> > 
> > Is there a Metacello configuration for Seaside that I should be using to get
> > me closer to the correct configuration?
> Looks like Paul gave you the necessary instructions. Mind that the failing tests are in latest version of Grease, which is loaded for Seaside 3.1.x (stable) and Seaside 3.2.0 (which is in development).
> Btw, the Metacello instructions are also mentioned on http://seaside.st/download/pharo <http://seaside.st/download/pharo>
> Let me know if you need any help.

Hi Johan,

I am able to load Grease and Seaside 3.2.0 from the Metacello configurations,
which work very well. There is one glitch in loading Seaside, because
WASecureKeyGenerator class>>canSeed tries to use GRPharoPlaform subclass,
does not exist. But that seems to be related to the same issue that is
causing the original unit test failures that you are asking about.

The failures in WAExternalFileLibraryTest are related to GRPlatform.
GRPlatform is providing a range of compatibility methods that are
vectored through its concrete subclasses, which appear to include
implementations for the various flavors of Pharo, as well as for a
number of versions of Gemstone.

I do not see any implementation of GRPlatform for Squeak. Presumably
this might be a class called GRSqueakPlatform, but I cannot find this
in any of the repositories, and the "GRPlatform current" idiom does
not answer anything appropriate for generic Squeak.

Hopefully I am just confused (by far the most likely explanation), but
it seems that either I am not loading the right packages, or (worse)
I am trying to "debug" software that nobody has gotten around to writing

Are these Travis CI failures actual failures, in the sense that they
are tests that used to work, but now are failing due to some new
problem? Or are they just an indication that GRSqueakPlatform does
not exist, and we need to find someone to implement it?


More information about the Squeak-dev mailing list