This is clearly my fault. I lost track of the plan. Going forward, I need to get a better handle on plans -- I've been to reactive, having discussions spread across launchpad, GitLab, IRC, and the mailing list.
Here's one proposal: instead of reverting the recent merge, we could cut a release off master before the merge. That would give us a less disruptive release (at the expense of a little complexity on the release branch).
Would this make everyone happy?
Sent from my iPhone
I asked you many times if you knew what you were doing, but I don't think you did wrong at any point, except maybe for not realizing you had changed plan. [Also, I'm a proponent of releasing more often, but that's a different debate.]
I'd like to proceed forward. I don't any good reason to undo any of the current changes.
If you want to keep supporting make-build and/or if Daniel wants to support it on the ECL side, that's possible (see my proposed reimplementation in comments to !34). I'm not aware of any other breakage.
Anton, can you run cl-test-grid with ASDF master?
—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Communism is feudality without chivalry. — Faré
On Sun, Nov 27, 2016 at 1:55 PM, Robert P. Goldman rpgoldman@sift.net wrote:
This is clearly my fault. I lost track of the plan. Going forward, I need to get a better handle on plans -- I've been to reactive, having discussions spread across launchpad, GitLab, IRC, and the mailing list.
Here's one proposal: instead of reverting the recent merge, we could cut a release off master before the merge. That would give us a less disruptive release (at the expense of a little complexity on the release branch).
Would this make everyone happy?
On 11/27/16 Nov 27 -1:11 PM, Faré wrote:
I asked you many times if you knew what you were doing, but I don't think you did wrong at any point, except maybe for not realizing you had changed plan. [Also, I'm a proponent of releasing more often, but that's a different debate.]
I'd like to proceed forward. I don't any good reason to undo any of the current changes.
Well, as my earlier message suggests, we don't have to undo anything: we can simply make a release out of the state before the removal of operation initargs. That would give everyone time to adjust.
I think the person who cares the most is Daniel, so Daniel, what do you say?
If you want to keep supporting make-build and/or if Daniel wants to support it on the ECL side, that's possible (see my proposed reimplementation in comments to !34). I'm not aware of any other breakage.
Anton, can you run cl-test-grid with ASDF master?
—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Communism is feudality without chivalry. — Faré
On Sun, Nov 27, 2016 at 1:55 PM, Robert P. Goldman rpgoldman@sift.net wrote:
This is clearly my fault. I lost track of the plan. Going forward, I need to get a better handle on plans -- I've been to reactive, having discussions spread across launchpad, GitLab, IRC, and the mailing list.
Here's one proposal: instead of reverting the recent merge, we could cut a release off master before the merge. That would give us a less disruptive release (at the expense of a little complexity on the release branch).
Would this make everyone happy?
I really don't see what that buys, but if so, please make it 3.1.8 rather than 3.2.0.
On Sun, Nov 27, 2016, 17:14 Robert Goldman rpgoldman@sift.net wrote:
On 11/27/16 Nov 27 -1:11 PM, Faré wrote:
I asked you many times if you knew what you were doing, but I don't think you did wrong at any point, except maybe for not realizing you had changed plan. [Also, I'm a proponent of releasing more often, but that's a different debate.]
I'd like to proceed forward. I don't any good reason to undo any of the current changes.
Well, as my earlier message suggests, we don't have to undo anything: we can simply make a release out of the state before the removal of operation initargs. That would give everyone time to adjust.
I think the person who cares the most is Daniel, so Daniel, what do you say?
If you want to keep supporting make-build and/or if Daniel wants to support it on the ECL side, that's possible (see my proposed reimplementation in comments to !34). I'm not aware of any other breakage.
Anton, can you run cl-test-grid with ASDF master?
—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics•
Communism is feudality without chivalry. — Faré
On Sun, Nov 27, 2016 at 1:55 PM, Robert P. Goldman rpgoldman@sift.net
wrote:
This is clearly my fault. I lost track of the plan. Going forward, I
need to get a better handle on plans -- I've been to reactive, having discussions spread across launchpad, GitLab, IRC, and the mailing list.
Here's one proposal: instead of reverting the recent merge, we could
cut a release off master before the merge. That would give us a less disruptive release (at the expense of a little complexity on the release branch).
Would this make everyone happy?
Robert Goldman writes:
On 11/27/16 Nov 27 -1:11 PM, Faré wrote:
I'd like to proceed forward. I don't any good reason to undo any of the current changes.
Well, as my earlier message suggests, we don't have to undo anything: we can simply make a release out of the state before the removal of operation initargs. That would give everyone time to adjust.
I think the person who cares the most is Daniel, so Daniel, what do you say?
I'm fine with having 3.1.7.26 (already added to ECL), so go ahead keeping the changes. What I want to say that its somewhat confusing to agree on something and having François do the opposite out of the blue (and seeing nothing wrong with that). But I suppose it's ASDF developers concern, not mine.
Best regards, Daniel
I'm fine with having 3.1.7.26 (already added to ECL), so go ahead keeping the changes.
Good.
What I want to say that its somewhat confusing to agree on something and having François do the opposite out of the blue (and seeing nothing wrong with that). But I suppose it's ASDF developers concern, not mine.
I acted after discussing with Robert.
What concern whatsoever do you have for ASDF internals?
If your actual concern is that you want make-build to be provided by ASDF itself rather than by some ECL extension, that is possible. We could have a version of the function in backward-interface (no less "working" than the previous bogus function), and just have it print a warning that it is deprecated when called. You will still have to migrate off of it before it gets removed, but then user code if any will keep running a bit longer before you migrate.
—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Once upon a time, there was a person whose life was so good there was no story to tell about it. — Foer
On 11/28/16 Nov 28 -12:24 AM, Daniel Kochmański wrote:
Robert Goldman writes:
On 11/27/16 Nov 27 -1:11 PM, Faré wrote:
I'd like to proceed forward. I don't any good reason to undo any of the current changes.
Well, as my earlier message suggests, we don't have to undo anything: we can simply make a release out of the state before the removal of operation initargs. That would give everyone time to adjust.
I think the person who cares the most is Daniel, so Daniel, what do you say?
I'm fine with having 3.1.7.26 (already added to ECL), so go ahead keeping the changes. What I want to say that its somewhat confusing to agree on something and having François do the opposite out of the blue (and seeing nothing wrong with that). But I suppose it's ASDF developers concern, not mine.
This really wasn't Faré's doing -- the mistake was only mine. The commit was done by Faré, because I was going AFK for a while. But it was completely my slip-up and not his.
Best, r
-
Daniel Kochmański -
Faré -
Robert Goldman -
Robert P. Goldman