torsten.grust
2020-11-5 09:07:11

Nope, no other editors involved. It’s a tiny group of students who are affected. We’re on it.


torsten.grust
2020-11-5 09:16:29

Good morning, Matt. I’ve seen your commit fc53f2998c422dcbf7f49a754529c9949c69e0b7 which appears to be related to this issue. We’d be happy to test that. Is a current build available somewhere? Thanks!


mflatt
2020-11-5 14:00:47

A build will arrive https://www.cs.utah.edu/plt/snapshots/, but it’s still an hour or so away from finishing. I do not expect that commit to fix the bug, since the commit is just a fancier (more correct) version of the s #; change that I suggested. It might allow. progress to a different error message and give us new information, but even that is far from certain.


torsten.grust
2020-11-5 14:03:07

I am grateful for the immediate attention, Matt. We’ll report back what the student sees.


gknauth
2020-11-5 15:19:10

Just thinking out loud, I wonder if this is a specific Windows 10 in Germany problem, having to do with file encodings, or default OS assumptions about encodings. I work with files from all over the world (I’m at AccuWeather), my work desktop is Win10, and I used to have file encoding nightmares all the time. I coded around them (put in handlers and defaults), but that was in Scala. Just wondering if it could be that. But then why would things be new to 7.8 (assuming 7.7 was Ok)? Something low-level that changed? Do Racket unit tests have a variety of files in different encodings, running in different OS environments with different language/country assumptions? That’s what I sometimes what I had to do in my work environment.


gknauth
2020-11-5 15:20:25

Preparing a proper German sentence: Racket must in Germany succeed!


torsten.grust
2020-11-5 15:21:48

@gknauth: it appears the bug is quite old (appears in 7.7 and even in 7.4).


torsten.grust
2020-11-5 15:22:36

We’ve had the student restart her machine in all sorts of language settings/locales. No effect.


torsten.grust
2020-11-5 15:23:42

Mind you, she’s in the absolute minority. We have 700 students runnning/loading/saving the same files without any problem whatsoever. (Yup, all located in Germany :wink:.)


gknauth
2020-11-5 15:24:19

Those keybindings are very emacsy and readline friendly, so I never noticed it was a problem. But I agree C-y (yank) for paste is not what non-Emacs people expect. On Linux, C-a / C-e for beginning/end of line are probably natural for people used to readline.


torsten.grust
2020-11-5 15:24:50

Your points are well taken, however. File encoding will haunt me until retirement. (I am a database person when it comes to research.)


gknauth
2020-11-5 15:25:19

Having custom keybindings copied from a template might be nice.


gknauth
2020-11-5 15:29:24

I’m glad to hear the >700 students are not having a problem. I’m sad the one student is having a problem. Since it’s only her, now I wonder, it is her, or her machine? Can she open her file on someone else’s machine? Can a different user on her machine open a file? It sounds like some weird default that’s either attached to her user profile or that of the machine.


torsten.grust
2020-11-5 15:33:12

The issue already arises when she tries to save a newly created file (newly created on her machine). This is not about strangely encoded umlauts, or similar, it appears.


mflatt
2020-11-5 15:34:06

The snapshot is ready, BTW. For the most direct test, your student will want to download the one labeled “BC”: https://www.cs.utah.edu/plt/snapshots/current/installers/racket-7.9.0.3-x86_64-win32-bc.exe If you instead use the first Windows option https://www.cs.utah.edu/plt/snapshots/current/installers/racket-7.9.0.3-x86_64-win32-cs.exe, that’s Racket CS (i.e., the snapshot has already switched over to CS as the default). It would be interesting if BC vs. CS makes a difference, but I don’t expect a difference in this case.


torsten.grust
2020-11-5 15:34:30

We’re on it.


spdegabrielle
2020-11-5 18:18:14

• To use command line tools (raco) you will also need to run the command sudo xattr -r -d com.apple.quarantine /Applications/Racket\ v7.9/ Mac users! Is this no longer true?


samth
2020-11-5 18:20:12

It should no longer be true, but where did that advice come from?


george.privon
2020-11-5 18:22:39

i don’t remember having to do that and don’t see it in my zsh history (but the latter isn’t very stringent since it seems i’ve forgotten to make the history saving unlimited)



samth
2020-11-5 18:40:45

Sure but where did it come from? I didn’t think it was required for 7.8 but maybe you can check?



sorawee
2020-11-5 19:21:15

@spdegabrielle has been updating it to keep the version number up-to-date


sorawee
2020-11-5 19:21:26

That’s how 7.7 becomes 7.8


samth
2020-11-5 19:23:45

sure, but was it recommended on the mailing list or something?


mflatt
2020-11-5 19:28:49

Originated with the v7.5 release announcement?


spdegabrielle
2020-11-5 19:34:00

I remove it.


samth
2020-11-5 19:34:22

Don’t remove it if it’s still needed — my question is just whether we’ve checked


mflatt
2020-11-5 19:38:05

It’s not needed anymore (just double checked), since distributions are notarized.


spdegabrielle
2020-11-5 19:38:35

I don’t think it is still needed. I downloaded 7.9(cs) today, changed $PATH, and ran raco setup -l quickscript-extra and MacOS did not complain. So it looks like it is no longer needed. I can roll back if necessary.


george.privon
2020-11-5 20:23:31

I can confirm. I just installed 7.9cs and was able to run raco pkg migrate 7.8 without issue (and without running that xattr command)


capfredf
2020-11-5 20:33:04

I think this instruction is still needed if users want to try out pre-built snapshots.


capfredf
2020-11-5 20:34:40

And on my machine, I don’t have to use sudo with this command.


spdegabrielle
2020-11-5 23:28:17

@capfredf you are correct!


spdegabrielle
2020-11-5 23:38:04

corrected https://github.com/racket/racket/wiki/Set-your-PATH-environment-variable to note that the command is needed if you are using a snapshot.