Remove #willJumpIfFalse

[jecisc]

#willJumpIfFalse is already a method present un Debbuging-Core in Pharo.

This method creates a clash with the one already in Pharo and need to be removed.

[jecisc]

The error does not seems related to my PR

[adri09070]

Was it recently added to Debugging-Core? Because, if Pharo let me create it, it means it didn't exist in Debugger-Core, right?

[jecisc]

It has been there for a while I think. At least some months.

Pharo does not prevent to override existing methods, that’s what is happening each time we update a method. So extensions can replace existing methods but we end up in a dirty state

[StevenCostiou]

How did you detect the clash?
Because I can only see the Sindarin protocol and not the original method.

There is only one usage, in the Flashback Decompiler.
I did not know that package.

[StevenCostiou]

Errors are unrelated but are scary: "could not resolve baseline of Sindarin"?

[adri09070]

Errors are unrelated but are scary: "could not resolve baseline of Sindarin"?

it's ok, it will be resolved once #61 is merged, I think

[jecisc]

How did you detect the clash? Because I can only see the Sindarin protocol and not the original method.

There is only one usage, in the Flashback Decompiler. I did not know that package.

I detected the clash while doing a "recalculate dirty packages" on the Pharo repository. We currently have two clashes and this one is one of them.

Here is the version in the Pharo repository: https://github.com/pharo-project/pharo/blob/Pharo12/src/Debugging-Core/InstructionStream.extension.st#L130

I prioritized keeping the one that is in Debbugging-Core but if you think there is a better solution I'm also open to it. I'd just like to avoid clashes of methods in the base image of Pharo

[StevenCostiou]

No you are right, the original one should be kept.
I just did not know how to detect that kind of case.

I will fix the baseline problem then merge this PR.

[StevenCostiou]

eeeh just a question: that is going to Pharo 11 ?

[adri09070]

Well, tests don't pass because of the removed method. It doesn't find the other one

[jecisc]

It should not break P11 I think but if we can keep it for p12 it would be nice since the release is for this week :)

[jecisc]

Well, tests don't pass because of the removed method. It doesn't find the other one

Hum... This is weird because it's present in Pharo (cf link I provided just before)

[StevenCostiou]

I think the other PR I just merged went to P11 :/
How do I make it go to P12?

[adri09070]

I think the other PR I just merged went to P11 :/
How do I make it go to P12?

No, it's not. It has been merged into master.
To merge it into Pharo11, we need to merge master into the Pharo-11 branch.

To merge it into Pharo12, we should create a branch Pharo-12, merge master into Pharo-12, and change the baseline of NewTools to load the Pharo12 branch

[StevenCostiou]

Ah! Nice, so no impact? We can continue merging on master?

[adri09070]

Ah! Nice, so no impact? We can continue merging on master?

yes we can

[jecisc]

My guess is that the problem is this:

• The Pharo image get bootstraped and loads the Debugging-Core package with the Pharo version of #willJumpIfFalse
• Later in the bootstrap Sindarin is loaded and install its own version of #willJumpIfFalse
• The CI here uses this image and unload Sindarin to reload the new version, its version of #willJumpIfFalse is removed but the Pharo original version is not installed back.

So, merging this should be fine.

[adri09070]

It looks correct.

What I can do is merge this into the master branch as it has no direct effect on Pharo. (Pharo11 and 12 loads the Pharo-11 branch).

And then I can run another CI on the master branch and it should be green