On Tue, Sep 8, 2015 at 6:56 PM, Usman Bhatti usman.bhatti@gmail.com wrote:
I've committed the packages with correct suffixes and updated configurations of Rubric and GlamourCore.
However, the packages are not loading because Roassal2 stable loads a different version of GlamourCore.
So, we have:
ConfigurationOfGlamour loading ConfigurationOfGlamourCore stable (3.1.3.Moose51) and ConfigurationOfRoassal2 loading ConfigurationOfGlamourCore 3.1.4.
The first one should load the correct packages but the second one takes precedence and hence forces to load older versions of the packages. I'll have a look later.
A wild guess:
The correct packages were loaded before because GlamourCore loaded Rubric-AndreiChis.207 and when another configuration tried to load an older version ( Rubric-AlainPlantec.206), it did not happen because that is just old version of the same package.
With suffix added (Rubric-Moose51-usmanbhatti.207), Metacello might not consider the package to be the ancestor of Rubric-AlainPlantec.206 and hence it proceeds with loading the package thus overwriting fixes with an old version.
In the past, I used loadDirective to debug package resolution mechanism in Metacello thru this command to see the actual list of packages loading:
(ConfigurationOfGlamour project version: #stable) record loadDirective
But I didn't work now (I do not get any package names actually loaded). Is there something changed recently in Metacello?
On Mon, Sep 7, 2015 at 3:39 PM, Tudor Girba tudor@tudorgirba.com wrote:
Hi,
On Mon, Sep 7, 2015 at 2:33 PM, Usman Bhatti usman.bhatti@gmail.com wrote:
Hi,
On Fri, Sep 4, 2015 at 5:49 PM, Tudor Girba tudor@tudorgirba.com wrote:
Hi,
Thanks. But, I meant to say that the packages should have a suffix. For example: Glamour-Morphic-Renderer-AndreiChis.324 should ideally be Glamour-Morphic-Renderer-Moose51-AndreiChis.324
I see.
Otherwise, we can also create a conflicting package with the one from the main repo of Glamour. For example, right now, we have in the Glamour repo: Glamour-Morphic-Renderer-SeanDeNigris.324
Luckily, in this case, the version was committed by Sean, but this could well have been committed by Andrei, and then we would have had a mess when using the local package cache (because we would have had 2 different implementations using the same file name).
But I do not understand what problem can be caused by two packages having same names coming from two different repos?
Imagine you have in the same folder a: moose51.image moose60.image package-cache/
In Moose 5.1 you load through Gofer PackageA.XYZ.70, and in Moose 6.0 you load also through Gofer a package with the same name but from a different repo. You will get problems in the Moose 6.0 image because the second time you load, it will be cached in package-cache.
So, all versions committed in the Moose51 repo, should have a Moose51 suffix.
Is that Ok? Could you republish the two packages with the Moose51 suffix?
Ok. I'll take care of that.
Thanks!
Doru
regards.
Cheers, Doru
Cheers, Doru
On Fri, Sep 4, 2015 at 4:12 PM, Usman Bhatti usman.bhatti@gmail.com wrote:
On Fri, Sep 4, 2015 at 3:10 PM, Tudor Girba tudor@tudorgirba.com wrote:
Hi,
Just a note: please still prefix the versions with a Moose51 suffix because we should not confuse them with the ones from the regular repository.
I created the two versions:
ConfigurationOfGlamourCore>>version313Moose51: (new #stable) ConfigurationOfRubric>>version1214Moose51:
they are loading correctly in the latest build and the bug is fixed.
regards.
Cheers, Doru
On Fri, Sep 4, 2015 at 3:02 PM, Usman Bhatti usman.bhatti@gmail.com wrote:
> > > On Fri, Sep 4, 2015 at 2:41 PM, Andrei Chis < > chisvasileandrei@gmail.com> wrote: > >> >> >> On Fri, Sep 4, 2015 at 2:21 PM, Usman Bhatti < >> usman.bhatti@gmail.com> wrote: >> >>> >>> >>> On Fri, Sep 4, 2015 at 12:41 PM, Tudor Girba <tudor@tudorgirba.com >>> > wrote: >>> >>>> As I mentioned when we released 5.1, we should create a separate >>>> repo for committing code packages with fixes for 5.1. The configuration >>>> remains in the same place and we will create a new version that will point >>>> to the new packages. >>>> >>>> @Usman: could you create the moose/moose51 repo? >>>> >>> >>> MCHttpRepository >>> location: 'http://smalltalkhub.com/mc/Moose/Moose51/main' >>> user: '' >>> password: '' >>> >>> @Andrei, can you commit the fix(es) in the repo? I see how to >>> correct the problem in Glamour renderer but I do not understand what checks >>> you would wish to introduce in Rubric for default theme colors. >>> >>> >> >> I committed the two fixes. >> Can you update the configurations? >> > > Yes. I'm doing it now. > > >> >> Cheers, >> Andrei >> >> >> >>> >>>> Doru >>>> >>>> -- >>>> www.tudorgirba.com >>>> >>>> "Every thing has its own flow" >>>> >>>> On 04 Sep 2015, at 12:34, Usman Bhatti usman.bhatti@gmail.com >>>> wrote: >>>> >>>> >>>> >>>> On Fri, Sep 4, 2015 at 11:49 AM, stephan stephan@stack.nl >>>> wrote: >>>> >>>>> >>>>> >>>>> On 04-09-15 11:42, Usman Bhatti wrote: >>>>> >>>>>> A new version only will not solve the problem. The fix also >>>>>> needs to be committed somewhere and if I commit the fix it in the main >>>>>> branch (see screenshot below), I need to merge all the the changes done for >>>>>> Moose 6.0 (and update other packages that these packages depend on) and >>>>>> hence break my tools in Moose 5.1. >>>>>> >>>>> Why? The only thing you are depending on is that #bleedingEdge >>>>> returns a version for Moose 6, isn't it? >>>> >>>> >>>> No. We are using the stable version of Moose (release Moose 5.1) >>>> for Pharo 4.0 and not the bleeding edge. >>>> >>>>> >>>>> >>>>> Stephan >>>>> _______________________________________________ >>>>> Moose-dev mailing list >>>>> Moose-dev@iam.unibe.ch >>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>>>> >>>> >>>> _______________________________________________ >>>> Moose-dev mailing list >>>> Moose-dev@iam.unibe.ch >>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>>> >>>> >>>> _______________________________________________ >>>> Moose-dev mailing list >>>> Moose-dev@iam.unibe.ch >>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>>> >>>> >>> >>> _______________________________________________ >>> Moose-dev mailing list >>> Moose-dev@iam.unibe.ch >>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>> >>> >> >> _______________________________________________ >> Moose-dev mailing list >> Moose-dev@iam.unibe.ch >> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >> >> > > _______________________________________________ > Moose-dev mailing list > Moose-dev@iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev > >
-- www.tudorgirba.com
"Every thing has its own flow"
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
-- www.tudorgirba.com
"Every thing has its own flow"
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
-- www.tudorgirba.com
"Every thing has its own flow"
Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev