Hi all,
first I must apologize that I'm totally and hopelessly behind in reading the mails on the lists. I couldn't find answers to the following questions, so please be kind if I just overlooked them.
And I'm sending this to the foundation list because I think this is a harvesting issue...
Why are a lot of the (browser, Undeclared references) query messages broken by the introduction of SystemNavigation (I'm using image 3.6a update 5247)? And why are there 9 implementors of #systemNavigation all over the system? (I guess that is what the class extension discussion was about, I read that much)
And the deprecated call should allow you to proceed with the deprecated version until it is fixed and not just cut you off. Boy, we went down this road with modules, didn't we?
Michael
Hi michael
On Friday, June 6, 2003, at 08:31 AM, Michael Rueger wrote:
Hi all,
first I must apologize that I'm totally and hopelessly behind in reading the mails on the lists. I couldn't find answers to the following questions, so please be kind if I just overlooked them.
And I'm sending this to the foundation list because I think this is a harvesting issue...
Why are a lot of the (browser, Undeclared references) query messages broken by the introduction of SystemNavigation (I'm using image 3.6a update 5247)? And why are there 9 implementors of #systemNavigation all over the system? (I guess that is what the class extension discussion was about, I read that much)
The idea is that by having systemNavigation is class like stringHolder all the subclass avoid to duplicate SystemNavigation new and still be able to override it. Now I do not want to put systemNavigation in Object (sure we could use a class extension) but for the moment I'm waiting to have a better package management (even if avi will do the right remark that monticello is working and DVS as well). So if ***you*** reviewers, guides tell us we are really stupid and we should have used DVS I will say why not but for the moment we do not have a decomposition of the system in terms of packages.
And the deprecated call should allow you to proceed with the deprecated version until it is fixed and not just cut you off.
We know we fixed that as soon as we realized that it was stupid. The PROBLEM is that not all the changesets (where this is fixed) have been harvested. I said it, said it and said it again but there is nothing I can do except apologize to have been stupid at first.
By the way michael cleaning the system is not that fun and I ask regurlarly for design idea because I do not think that I have all the answers (the past proved it). Now just participate, read the propositions, browse the changesets, ....help us to push this out of the door.
Stef
Boy, we went down this road with modules, didn't we?
Michael
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
Hi Stephane,
By the way michael cleaning the system is not that fun and I ask
I know, remember how long I've been struggeling with the network rewrite :-)
regurlarly for design idea because I do not think that I have all the answers (the past proved it).
There was no offense intended, you guys are doing good work. My main concern was that changes that break too much, especially in the dev tools, shouldn't get into the image.
Let's see how I'll mess up the system with my rewrite ;-)
Michael
Hi michael
The funny aspect of cleaning the kernel was to be sure the system would not break. So normally it should not, you only see the problem of using a stupid deprecated mechanism at the beginning. Now you should get notified and process smoothly.
The most fun clean was to move the current changeset, from Smalltalk to ChangeSet class. I still did not get feedback on this one but this was fun to violate temporary encapsulation to be able to pass the current instance around.
On Friday, June 6, 2003, at 08:26 PM, Michael Rueger wrote:
Hi Stephane,
By the way michael cleaning the system is not that fun and I ask
I know, remember how long I've been struggeling with the network rewrite :-)
I imagine that this is even worse, with all these protocols. By the way are you developing tests for the clean you are doing. We are trying but as I'm the only really active now I got stuck and continue to think about the next clean.
regurlarly for design idea because I do not think that I have all the answers (the past proved it).
There was no offense intended, you guys are doing good work.
I did not take like that at all. My message was that any important changes need feedback. Often we have the right answer/idea but just talking with somebody or reading an email really helps.
My main concern was that changes that break too much, especially in the dev tools, shouldn't get into the image.
Let's see how I'll mess up the system with my rewrite ;-)
Michael
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
squeakfoundation@lists.squeakfoundation.org
-
Michael Rueger -
Stephane Ducasse