Hi,

On Sat, Oct 4, 2014 at 8:34 AM, Thierry Goubier <thierry.goubier@gmail.com> wrote:

Le 04/10/2014 08:14, Tudor Girba a ��crit :

Hi Thierry,

On Sat, Oct 4, 2014 at 12:03 AM, Thierry Goubier
<thierry.goubier@gmail.com <mailto:thierry.goubier@gmail.com>> wrote:

�� [...]
�� And now, to make integration easier, we pile more stuff on top of
�� the existing, bound to be removed, infrastructure: Glamour, Rubric,
�� etc... And both Glamour and Spec don't make it easy to solve Morphic
�� bugs.

�� I value your ambition a lot :) But I also feel that it stresses the
�� Rmod team, and it leaks over the community.

I believe there is a confusion here. GT and Glamour are not maintained
by the RMoD team. GT is�� an external project and it is maintained in its
separate repository. That is why it stresses RMoD less and it is a way
of scaling.

I agree that the GT Tools are a bit different. Still, TxText integration will have to work on all users of all forms of text, including GT (and additionally GT).

Yes. That will become a priority once we address the current issues with GTInspector/Playground.

��

And now you are starting to have the issue of handling slices on GT from Pharo: do people have to register on Moose, resynchronize on GT on Moose, push their slices there, wait for Moose to push a slice to integrate on Pharo, wait for Pharo to integrate that slice?

Something like that. Only there is no wait time to integrate in Moose as GT is a standalone project, and Moose simply integrates the latest development version of all its dependencies. So, there is not that much difference to integrating in Pharo. Once a new stable version of GT is available, it will automatically be available in Pharo.

��

Also, since more than one year I am working on fixing various Morphic
problems, and I can tell you that GT provides a significant productivity
boost in that specific area. In any case, you should also note that
these tools are under development since several years (the overall
project started 5 years ago) and they are being actively used since 2
years in Moose. I believe we never had that much focus on the tools before.

I found that stepping through the opening sequence of GT to debug MorphTreeMorph a pain and had to find another way to understand the issue.

Can you be more specific? What did you try to achieve and did not manage?

��

I do have specific tools I can use in such circumstances that you don't have.

What kind of tools? I would be interested in learning about them and the analysis use cases they support.

��

I agree that Morphic has improved due to the fact GT is exposing it in different ways (and you solving bugs in there), and that you are integrating Alain's work (Rubric) in advance. I'm not worried by GT per se (see above for the contribution things), but by the fact core elements (Athens, TxText, NativeBoost, git) are a long way from stabilising.

I think the way is not that long. But, I am an optimist. Let's see :)

Cheers,

Doru

��

Thierry

Cheers,
Doru

--
www.tudorgirba.com <http://www.tudorgirba.com>

"Every thing has its own flow"

www.tudorgirba.com

"Every thing has its own flow"