Release Regressions for 8.9

[jbclements]

There are also many new failures related to version number format checkings, see #4630 for a discussion of those.

URL to examine results on 8.8: https://pkg-build.racket-lang.org/
URL to examine results on 8.8.900: https://plt.cs.northwestern.edu/pkg-build/

Build Failures

• isl-spec Caused by new export of ==> from lang/htdp-intermediate-lambda https://plt.cs.northwestern.edu/pkg-build/server/built/fail/isl-spec.txt
• sci https://plt.cs.northwestern.edu/release-pkg-build/server/built/fail/sci.txt Seems to be caused by a difference in availability of gfortran shared library.

[samth]

The failure in isl-spec was likely caused by racket/htdp#193 by @mikesperber. Generally the policy is that we avoid new provided identifiers that cause failures in existing packages, so I think something should change here.

[mikesperber]

The failure in isl-spec was likely caused by racket/htdp#193 by @mikesperber.

Yes. But the identifier is the same as the one it conflicts with, so I don't understand the exact problem. @samth Where can I find the code that causes the problem?

[samth]

The file that no longer works is: https://github.com/dbp/isl-spec/blob/main/main.rkt
It breaks because it gets the ==> identifier from the quickcheck package which defines ==> here: https://github.com/ifigueroap/racket-quickcheck/blob/master/quickcheck/property.rkt#L14
The definition of ==> in htdp/intermediate-lambda is here: https://github.com/racket/deinprogramm/blob/master/deinprogramm-signature/deinprogramm/quickcheck/quickcheck.scm#L508

[mfelleisen]

I think I know where this package was used and Daniel @dpb can confirm that he won't use it again. But even if, it should be straightforward to resolve this issue with a change to this package.

[samth]

If @dbp can change the package to fix the problem now, that would be great and would resolve this particular problem.

[dbp]

I wasn't sure if not providing it was enough, so I stopped requiring it from quickcheck. As @mfelleisen said, this is not an actively used package, it was a one-time creation for a course (if I used it again, I'll deal with setting it up at that point).

[samth]

There were other identifiers that had the same issue, so I just make a PR to make it all pass: dbp/isl-spec#2

[dbp]

@samth 👍

[samth]

The sci failure seems to have gone away (for reasons I don't understand). There is currently an unexpected failure with envy but that seems to be about the timing of the different builds picking up lexi-lambda/envy@580610f. Hopefully no more failures crop up.

[jbclements]

It appears to me that I can check off both of these boxes, and I'm going to do so. @samth, or anyone else, feel free to uncheck these if you think this is not correct.

[jbclements]

Closing....