[squeak-dev] The Trunk: CollectionsTests-dtl.350.mcz

Chris Muller asqueaker at gmail.com
Thu Mar 4 03:14:40 UTC 2021


> > > On Mar 2, 2021, at 3:59 AM, Christoph Thiede <
> christoph.thiede at student.hpi.uni-potsdam.de> wrote:
> > >
> > > ???Hi Dave,
> > >
> > > could there be something wrong with the ancestry of this version? The
> > > notification email does not contain any visible diff and when updating
> my
> > > image today, I got an MCEmptyDiffyVersion warning.
>

Dave's comment explicitly acknowledged the omission of 349.  I'm curious
why you would assume something's wrong.  It's one of several Warnings, but
it's not an Error.


> Yes, it looks like I made a mistake here. My intention was to bypass
> CollectionsTests-jar.349 entirely by letting CollectionsTests-ul.348
> be the sole ancestor of CollectionsTests-dtl.350 with no changes between
> 348 and 350. This was the wrong thing to do because many people already
> had 349 loaded in their image.
>

That doesn't make it wrong in the least.  You had good reason to exclude it
from the ancestry.  I personally would've simply deleted 349 from the
repository and not even made 350.  The coherence and correctness of the
permanent ancestry far outweighs the temporary in-RAM state of a
developers' image, which isn't confusing if one pays the ancestry the
attention it deserves.


> I do not think that I can fix the 350 version without causing additional
> problems for the update stream. It should  not cause a problem for the
> update stream in general, but for everyone who had already updated to
> 349, a manual update of CollectionTests is probably necessary.
>
> I apologize for the error :-(


I put that Warning message in years ago for no more than debugging an issue
at that time.  It should be removed.  Not only is there no need to
apologize, but thanks for doing the right thing.

 - Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20210303/033333d1/attachment.html>


More information about the Squeak-dev mailing list