This was my attempt to attach the prior version history for Chronology to the new Chronology-Core and Chronology-Tests packages. It appears to break the update stream, so I will delete Chronology-Core-dtl.2 and Chronology-Tests-dtl.2 updates from the source.squeak.org repository.
Hopefully no one has tried to update their trunk images during this unfortunate experiment.
Sorry, Dave
On Mon, Feb 29, 2016 at 02:05:19AM +0000, commits@source.squeak.org wrote:
David T. Lewis uploaded a new version of Chronology-Core to project The Trunk: http://source.squeak.org/trunk/Chronology-Core-dtl.2.mcz
==================== Summary ====================
Name: Chronology-Core-dtl.2 Author: dtl Time: 28 February 2016, 8:24:12.86775 pm UUID: 0dc832c9-bf4c-4559-a705-c19f44392089 Ancestors: Chronology-Core-dtl.1, Kernel-tfel.1001
Re-save Chronology-Core-dtl.1 with Kernel-tfel.1001 added to ancestry to preserve version history from Kernel package.
=============== Diff against Chronology-Core-dtl.1 ===============
If I want to move a sub-package out of a larger package (for example move Kernel-Chronology out of package Kernel and into its own package called Chronology-Core), and if I want to preserve the version history for that sub-package, is there a way to do this?
I do not see any tool support for it. I tried manually updating the ancestry for the new package (Chronology-Core) and adding the ancestry from a suitable Kernel ancestor. This seems to produce a resonable version history in my local image, and all looked good when I saved the package to local package-cache.
After copying the locally saved MCZ's with added version history up to source.squeak.org, the result was a broken update stream. So ....
My manual hack is clearly not the right way to attack this problem. Is there another approach that might work?
I really do not want to lose version history on the entire Chronology package, so I am hoping that there is some simple thing that I may have overlooked.
Thanks, Dave
On Sun, Feb 28, 2016 at 09:17:42PM -0500, David T. Lewis wrote:
This was my attempt to attach the prior version history for Chronology to the new Chronology-Core and Chronology-Tests packages. It appears to break the update stream, so I will delete Chronology-Core-dtl.2 and Chronology-Tests-dtl.2 updates from the source.squeak.org repository.
Hopefully no one has tried to update their trunk images during this unfortunate experiment.
Sorry, Dave
On Mon, Feb 29, 2016 at 02:05:19AM +0000, commits@source.squeak.org wrote:
David T. Lewis uploaded a new version of Chronology-Core to project The Trunk: http://source.squeak.org/trunk/Chronology-Core-dtl.2.mcz
==================== Summary ====================
Name: Chronology-Core-dtl.2 Author: dtl Time: 28 February 2016, 8:24:12.86775 pm UUID: 0dc832c9-bf4c-4559-a705-c19f44392089 Ancestors: Chronology-Core-dtl.1, Kernel-tfel.1001
Re-save Chronology-Core-dtl.1 with Kernel-tfel.1001 added to ancestry to preserve version history from Kernel package.
=============== Diff against Chronology-Core-dtl.1 ===============
Seems like Adopt should work...
On Sun, Feb 28, 2016 at 8:32 PM, David T. Lewis lewis@mail.msen.com wrote:
If I want to move a sub-package out of a larger package (for example move Kernel-Chronology out of package Kernel and into its own package called Chronology-Core), and if I want to preserve the version history for that sub-package, is there a way to do this?
I do not see any tool support for it. I tried manually updating the ancestry for the new package (Chronology-Core) and adding the ancestry from a suitable Kernel ancestor. This seems to produce a resonable version history in my local image, and all looked good when I saved the package to local package-cache.
After copying the locally saved MCZ's with added version history up to source.squeak.org, the result was a broken update stream. So ....
My manual hack is clearly not the right way to attack this problem. Is there another approach that might work?
I really do not want to lose version history on the entire Chronology package, so I am hoping that there is some simple thing that I may have overlooked.
Thanks, Dave
On Sun, Feb 28, 2016 at 09:17:42PM -0500, David T. Lewis wrote:
This was my attempt to attach the prior version history for Chronology to the new Chronology-Core and Chronology-Tests packages. It appears to break the update stream, so I will delete Chronology-Core-dtl.2 and Chronology-Tests-dtl.2 updates from the source.squeak.org repository.
Hopefully no one has tried to update their trunk images during this unfortunate experiment.
Sorry, Dave
On Mon, Feb 29, 2016 at 02:05:19AM +0000, commits@source.squeak.org wrote:
David T. Lewis uploaded a new version of Chronology-Core to project The Trunk: http://source.squeak.org/trunk/Chronology-Core-dtl.2.mcz
==================== Summary ====================
Name: Chronology-Core-dtl.2 Author: dtl Time: 28 February 2016, 8:24:12.86775 pm UUID: 0dc832c9-bf4c-4559-a705-c19f44392089 Ancestors: Chronology-Core-dtl.1, Kernel-tfel.1001
Re-save Chronology-Core-dtl.1 with Kernel-tfel.1001 added to ancestry to preserve version history from Kernel package.
=============== Diff against Chronology-Core-dtl.1 ===============
I know how to adopt a different ancestor within a package, but I don't see any obvious way to adopt an ancestor from a different package.
Don't laugh, but what I actually did was use an inspector to find the ancestry for the version that I wanted to adopt from Kernel, saved its ancestors array with "Smalltalk at: #FOO put ancestors", then in an inspector for the version in Chronology-Core, I did "ancestors := ancestors, FOO". After saving this to my package cache, I had an MCZ with plausible looking ancestry. I did the same for Chronology-Tests.
I then copied the new MCZs to trunk, and tried updating an image. The process seemed to be stuck running "forever", though I am not sure why. I tried it a couple of times with the same result.
At that point I quickly decided to delete the new MCZs before I really screwed something up badly in trunk. I figured that it was time to ask for help before I really made a mess of things.
Dave
Seems like Adopt should work...
On Sun, Feb 28, 2016 at 8:32 PM, David T. Lewis lewis@mail.msen.com wrote:
If I want to move a sub-package out of a larger package (for example move Kernel-Chronology out of package Kernel and into its own package called Chronology-Core), and if I want to preserve the version history for that sub-package, is there a way to do this?
I do not see any tool support for it. I tried manually updating the ancestry for the new package (Chronology-Core) and adding the ancestry from a suitable Kernel ancestor. This seems to produce a resonable version history in my local image, and all looked good when I saved the package to local package-cache.
After copying the locally saved MCZ's with added version history up to source.squeak.org, the result was a broken update stream. So ....
My manual hack is clearly not the right way to attack this problem. Is there another approach that might work?
I really do not want to lose version history on the entire Chronology package, so I am hoping that there is some simple thing that I may have overlooked.
Thanks, Dave
On Sun, Feb 28, 2016 at 09:17:42PM -0500, David T. Lewis wrote:
This was my attempt to attach the prior version history for Chronology to the new Chronology-Core and Chronology-Tests packages. It appears to break the update stream, so I will delete Chronology-Core-dtl.2 and Chronology-Tests-dtl.2 updates from the source.squeak.org repository.
Hopefully no one has tried to update their trunk images during this unfortunate experiment.
Sorry, Dave
On Mon, Feb 29, 2016 at 02:05:19AM +0000, commits@source.squeak.org wrote:
David T. Lewis uploaded a new version of Chronology-Core to project
The Trunk:
http://source.squeak.org/trunk/Chronology-Core-dtl.2.mcz
==================== Summary ====================
Name: Chronology-Core-dtl.2 Author: dtl Time: 28 February 2016, 8:24:12.86775 pm UUID: 0dc832c9-bf4c-4559-a705-c19f44392089 Ancestors: Chronology-Core-dtl.1, Kernel-tfel.1001
Re-save Chronology-Core-dtl.1 with Kernel-tfel.1001 added to ancestry
to preserve version history from Kernel package.
=============== Diff against Chronology-Core-dtl.1 ===============
Hi Dave, why did attaching teh history break the update stream?
Incidentlally, the update stream was _already_ broken before, and is still is, due to DateAndTime>>#< using an underscore assignment...
On Sun, Feb 28, 2016 at 8:17 PM, David T. Lewis lewis@mail.msen.com wrote:
This was my attempt to attach the prior version history for Chronology to the new Chronology-Core and Chronology-Tests packages. It appears to break the update stream, so I will delete Chronology-Core-dtl.2 and Chronology-Tests-dtl.2 updates from the source.squeak.org repository.
Hopefully no one has tried to update their trunk images during this unfortunate experiment.
Sorry, Dave
On Mon, Feb 29, 2016 at 02:05:19AM +0000, commits@source.squeak.org wrote:
David T. Lewis uploaded a new version of Chronology-Core to project The Trunk: http://source.squeak.org/trunk/Chronology-Core-dtl.2.mcz
==================== Summary ====================
Name: Chronology-Core-dtl.2 Author: dtl Time: 28 February 2016, 8:24:12.86775 pm UUID: 0dc832c9-bf4c-4559-a705-c19f44392089 Ancestors: Chronology-Core-dtl.1, Kernel-tfel.1001
Re-save Chronology-Core-dtl.1 with Kernel-tfel.1001 added to ancestry to preserve version history from Kernel package.
=============== Diff against Chronology-Core-dtl.1 ===============
On 29.02.2016, at 16:55, Chris Muller asqueaker@gmail.com wrote:
Incidentally, the update stream was _already_ broken before, and is still is, due to DateAndTime>>#< using an underscore assignment…
What? Didn’t I fix that in Kernel-bf.998?
Btw this came from me enabling the syntaxHighlightingAsYouTypeLeftArrowAssignment preference assuming that it would show assignment arrows but store as colon-equals. Maybe we should get rid of that, or fix it to do just that.
- Bert -
On Tue, Mar 1, 2016 at 8:17 AM, Bert Freudenberg bert@freudenbergs.de wrote:
On 29.02.2016, at 16:55, Chris Muller asqueaker@gmail.com wrote:
Incidentally, the update stream was _already_ broken before, and is still is, due to DateAndTime>>#< using an underscore assignment…
What? Didn’t I fix that in Kernel-bf.998?
I'm referring to the update stream being broken. A stock Squeak5.0-15113 cannot currently be updated to latest trunk because of that underscore. I wanted to make sure Dave wasn't confusing that with his change.
Btw this came from me enabling the preference assuming that it would show assignment arrows but store as colon-equals. Maybe we should get rid of that, or fix it to do just that.
+1
I'm referring to the update stream being broken. A stock Squeak5.0-15113 cannot currently be updated to latest trunk because of that underscore. I wanted to make sure Dave wasn't confusing that with his change.
Correction: Okay, I didn't realize we allowUnderscoreAssignments by default. Really, still?
When I make my trunk image, I have been loading my own prefs to set my updateUrl to trunk for myself, but it also includes setting this off, and so that's why it was blowing up for me.
On 01.03.2016, at 21:15, Chris Muller asqueaker@gmail.com wrote:
I'm referring to the update stream being broken. A stock Squeak5.0-15113 cannot currently be updated to latest trunk because of that underscore. I wanted to make sure Dave wasn't confusing that with his change.
Correction: Okay, I didn't realize we allowUnderscoreAssignments by default. Really, still?
When I make my trunk image, I have been loading my own prefs to set my updateUrl to trunk for myself, but it also includes setting this off, and so that's why it was blowing up for me.
Yes we should change that pref.
And it looks like the differ replaces underscore assignments by ansi assignments. Maybe that’s why I missed one line when publishing my fix ...
Not entirely sure how to fix it since various subsequent versions of Kernel have been committed, and then the code has been moved to Chronology. Well, we at least should change the offending line in Chronology now. It’s in DateAndTime>><.
Btw I’ve been thinking about how to get a “proper” ancestry of Chronology and I guess the right thing would be to write a script that goes through all Kernel package versions, carves out a Kernel-Chronology package from it, and commits that. It could filter versions where no definition in Kernel-Chronology changed. The latest of those packages could then be adopted as ancestor of Chronology-Core.
I don’t think we ever have done that kind of history surgery in Monticello, but it is something I’ve heard people do in git, for example.
- Bert -
... and so I tried to load my personal preferences AFTER updating, but the removal of InputSensor has rendered all .prefs files unloadable...
On Tue, Mar 1, 2016 at 2:15 PM, Chris Muller asqueaker@gmail.com wrote:
I'm referring to the update stream being broken. A stock Squeak5.0-15113 cannot currently be updated to latest trunk because of that underscore. I wanted to make sure Dave wasn't confusing that with his change.
Correction: Okay, I didn't realize we allowUnderscoreAssignments by default. Really, still?
When I make my trunk image, I have been loading my own prefs to set my updateUrl to trunk for myself, but it also includes setting this off, and so that's why it was blowing up for me.
squeak-dev@lists.squeakfoundation.org
-
Bert Freudenberg -
Chris Muller -
David T. Lewis