On 01/19/2011 11:37 PM, Norbert Hartl wrote:
On 19.01.2011, at 19:01, Dale Henrichs wrote:
I think that the a good working model is to store the ConfigurationOfXXX in the project repository along with the project mcz files.
When a new version is released the configuration should be copied to MetacelloRepository for the Pharo and Squeak community. I have a GemSource MetacelloRepository where I put copies of configurations that have been ported to GemStone .... when a new version for GemStone is available.
I would not recommend that configurations be removed from MetacelloRepository as that could break existing configurations that expect to find the configuration there. So if it is found that there are out-of-date configurations in MetacelloRepository, a new version of the configuration should be copied into MetacelloRepository.
Now that Metacello is more established, I don't think that it is critical to require that all configurations be copied to MetacelloRepository.
It still is useful to have a sort of clearinghouse for configurations and until a better solution comes along it serves that purpose. BTW, I think Stef has plans to provide better solutions for Pharo...
I understand what you are saying and I partially agree. But it just doesn't feel right and this is easy to show. In this example of ConfigurationOfXMLSupport there was not one caring person like you are that does everything and has on overview. There is a Configuration in MetacelloRepository than it was decided to keep the main Configuration file in XMLSupport repository. Development took place in the XMLSupport repository. In the meantime someone changed the Configuration in MetacelloRepsository probably for integration purposes. Now the Configuration has been copied from XMLSupport and has overwritten the changes that have been made to the existing Configuration. That is what you get if you branch off things.
I don't know how to solve this properly because I didn't have time to follow the metacello discussions. But if there is a possibility to specify the stable for pharo 1.1.1 that doesn't change even if the packages evolves I can't see much reason to have a writable version of the Configuration in MetacelloRepository.
Norbert
I understand that there are potential problems using distributed version control, but one of the reasons that I chose to use code for specifying Metacello configurations is thatit made it possible to use Monticello merge facilities to reconcile changes mad in different branches ... This was a breakdown in procedure ... the intent of MetacelloRepository is not to create branches of configurations...if I recall, there was a change in the developers that coincided with the creation of metacello configurations ... I would hope that this case be classed as a learning experience rather than a condemnation of process. I agree that there could be a better way, but I think that the better way involves tools support and until we get the tools we must resort to process ... Dale
On 01/19/2011 06:12 AM, laurent laffont wrote:
On Wed, Jan 19, 2011 at 2:49 PM, Norbert Hartl<norbert@hartl.name <mailto:norbert@hartl.name>> wrote:
On 19.01.2011, at 14:30, laurent laffont wrote:
On Wed, Jan 19, 2011 at 1:37 PM, Norbert Hartl<norbert@hartl.name <mailto:norbert@hartl.name>> wrote:
It depends where you look at. ConfigurationOfXMLSupport exists in MetacelloRepository as well as in XMLSupport. The latter one being the official one, the former one being very outdated. My sugesstion would be to remove ConfigurationOfXMLSupport from MetacelloRepository to lower the confusion.
Personnally I always look in MetacelloRepository, this should be the reference IMHO.
Laurent,
I think it is up to the maintainers where to put those files. And in the XMLSupport case the decision was made deliberately not to put it in MetacelloRepository but in XMLSupport. The only thing I want to avoid is having multiple different files floating around that are edited in an inconsistent way.
I thought that all ConfigurationOfXXX working for Pharo should be put in MetacelloRepository so we have a central place to look at.
But I may be wrong. Mariano ? Dale ?
Laurent
Norbert
Norbert
On 19.01.2011, at 13:18, Tudor Girba wrote:
The last version 1.1.6 is marked as #release.
Cheers, Doru
On 19 Jan 2011, at 12:26, laurent laffont wrote:
I put ConfigurationOfXMLSupport version 1.0 because blessing was #release. But I agree to change.
Laurent
On Wed, Jan 19, 2011 at 12:15 PM, Marcus Denker <marcus.denker@inria.fr<mailto:marcus.denker@inria.fr>> wrote:
On Jan 19, 2011, at 11:43 AM, Tudor Girba wrote:
Indeed, this is a problem for Moose in general. We depend on XMLSupport, but now we cannot update it in PharoDev. I would also strongly support the idea of removing XMLSupport from PharoDev.
What would be important: we need the latest version in Pharo 1.2... else how can we ever have a version where the tests are green? (not of XML, but in general)
Marcus
Cheers, Doru
On 19 Jan 2011, at 11:32, Fabrizio Perin wrote:
Hi,
I have a problem with the XMLSupport version from Pharo 1.2.
The problem is that I need to work with one of the last version of XMLSupport but in the pharo image 1.2 is loaded an old version of XMLSupport (I mean a version from Jan 2010). I did try to load a newer version of XMLSupport using ConfigurationOfXMLSupport but some errors make this operation impossible. So either Pharo-dev 1.2 load a newer (possibly the last) version of XMLSupport by default or Pharo-dev 1.2 should not load XMLSupport at all.
I think that doesn't make sense to use by default such an old version, also considering that the last XMLSupport has nice and useful features like the XMLPluggableElementFactory.
By avoiding to load XMLSupport in Pharo-dev 1.2 by default you let people free to use the version that they like.
Thanks,
Fabrizio
-- www.tudorgirba.com<http://www.tudorgirba.com/>
"We cannot reach the flow of things unless we let go."
-- Marcus Denker -- http://www.marcusdenker.de <http://www.marcusdenker.de/> INRIA Lille -- Nord Europe. Team RMoD.
-- www.tudorgirba.com<http://www.tudorgirba.com/>
"If you interrupt the barber while he is cutting your hair, you will end up with a messy haircut."