scripts/checkCommits1by1.fsx: augment

[tehraninasab]

Improve this fsx script to also detect commits with bad CI
status. That is, if commit msg contains text "failing test"
then it would be allowed to have broken CI, else CI should be
green.

Fixes #95

[knocte]

Also "Compile F# scripts" step in GitHubCI is
removed because there is no error in
detectNotUsingGitPush1by1.fsx script
but compileFSharpScripts.fsx script can not
compile this file.

WHAT????

First, I don't understand why it would not be able to compile this file. And second, don't you think you should explain why it cannot compile this file? The word why in WorkflowGuidelines.md is referring to the ultimate why. Do not hide information to the next developer.

[knocte]

First, I don't understand why it would not be able to compile this file. And second, don't you think you should explain why it cannot compile this file? The word why in WorkflowGuidelines.md is referring to the ultimate why. Do not hide information to the next developer.

Third, do you think it would be acceptable that we remove the ability to check the correctness of ALL f# SCRIPTS just because ONE of them cannot be compiled?

[tehraninasab]

Also "Compile F# scripts" step in GitHubCI is
removed because there is no error in
detectNotUsingGitPush1by1.fsx script
but compileFSharpScripts.fsx script can not
compile this file.

WHAT????

First, I don't understand why it would not be able to compile this file. And second, don't you think you should explain why it cannot compile this file? The word why in WorkflowGuidelines.md is referring to the ultimate why. Do not hide information to the next developer.

I think it should be investigated and resolved in fsxc. What can I do about it in this PR?

[knocte]

I think it should be investigated and resolved in fsxc

If you think it's a fsxc bug, then:

1. State so in the commit message, there's no reason to hide your suspicions.
2. Or even go file a bug in fsx repo, and then reference the bug URL in your commit message so that the next developer understands why the script cannot be compiled, by reading your commit message.

However, this problem you're having is basically the same we're dealing with in #92 , so fix PR92 first with my guidance there (by doing the thing I initially told you to), and then after you do that, you would be able to apply the same thing here.

By the way when I say by doing the thing I initially told you to I don't intend to sound snarky, but think about it from my point of view: I tell you to do A, then you decide that for doing A you would need to do something that looks a bit "ugly" (putting a very long string in an F# file) so you try to fix the ugliness by trying other technique B slightly different than A: you suffer problems with technique B, and decide to solve the problem with a workaround C that deep down you know would be unacceptable by me (eliminating an entire step from CI?), instead of forgetting about B and really doing A as I asked you to.

[tehraninasab]

I think it should be investigated and resolved in fsxc

If you think it's a fsxc bug, then:

1. State so in the commit message, there's no reason to hide your suspicions.
2. Or even go file a bug in fsx repo, and then reference the bug URL in your commit message so that the next developer understands why the script cannot be compiled, by reading your commit message.

However, this problem you're having is basically the same we're dealing with in #92 , so fix PR92 first with my guidance there (by doing the thing I initially told you to), and then after you do that, you would be able to apply the same thing here.

By the way when I say by doing the thing I initially told you to I don't intend to sound snarky, but think about it from my point of view: I tell you to do A, then you decide that for doing A you would need to do something that looks a bit "ugly" (putting a very long string in an F# file) so you try to fix the ugliness by trying other technique B slightly different than A: you suffer problems with technique B, and decide to solve the problem with a workaround C that deep down you know would be unacceptable by me (eliminating an entire step from CI?), instead of forgetting about B and really doing A as I asked you to.

I didn't know the compile error is because of ResolutionFolder. I thought when you say the code could be greatly simplified by using FSharp.Data JSON parser instead of regex, it's not ok to put 500 lines of JSON string in the script and it's supposed to become shorter. Sorry.

[knocte]

Please rebase this PR 🙏 Thanks

[knocte]

I didn't know the compile error is because of ResolutionFolder.

I didn't know either, but you don't need to know; remember what I always say? Make it work, then after that, make it pretty. You were making it pretty (because 500 in-line string is ugly) before making it work.

[knocte]

Unfortunately there are conflicts, please do a rebase when you have time (low-priority task).

[tehraninasab]

Unfortunately there are conflicts, please do a rebase when you have time (low-priority task).

Done

[knocte]

@realmarv can you rebase this PR please

[knocte]

This PR needs to be rebased, when you have time. I've also realised that since the script would gain a new feature (not related to "detect not using gitPush1by1"), then we should rename the script. Maybe analyzeCommits1by1.fsx?

[knocte]

@realmarv nit: we have detected some -> it has been detected that some (as using "we" is a bit weird)

[tehraninasab]

Done

[tehraninasab]

This PR needs to be rebased, when you have time. I've also realised that since the script would gain a new feature (not related to "detect not using gitPush1by1"), then we should rename the script. Maybe analyzeCommits1by1.fsx?

Done