feat: support checkpoints for one-step replay debugger

[violetyao]

What does this PR do?

As part of the one-step replay debugger command, this PR checks for the presence of checkpoints. If they are present, this PR uploads them before we execute the Apex Test(s). Checkpoints have a limit of 5 per org/user, if the limit is exceeded, the command execution would be canceled and a message indicating this limit violation would be shown to the user.

What issues does this PR fix or reference?

@W-7757402@

Functionality Before

The one-step replay debugger command did not support checkpoints.

Functionality After

The one-step replay debugger command uploads checkpoints if they are present.

If there are more than five checkpoints, the command executed is canceled and a message is reported to the user.

[AnanyaJha]

Small comment, otherwise it looks good!

[AnanyaJha]

What's the behavior if you don't have any checkpoints? Do we still need to run this command? Is it possible to check whether checkpoints are present beforehand?

[violetyao]

What's the behavior if you don't have any checkpoints? Do we still need to run this command? Is it possible to check whether checkpoints are present beforehand?

Thank you for reviewing, Ananya! We discussed in one of our daily syncs and decided to close @W-7748717@ because the debug & run command should actually not give warnings (should just run through the test) if there is no checkpoint/breakpoint.

We use SFDX: Update Checkpoints in Org to check if there is any checkpoint available and to upload checkpoints if they exist. If there is no checkpoint, the behavior of the debug & run button is the same with directly running tests using the green run button, except that those SFDX: Update Checkpoints in Org messages are printed. I've attached a screenshot below!

[AnanyaJha]

What's the behavior if you don't have any checkpoints? Do we still need to run this command? Is it possible to check whether checkpoints are present beforehand?

Thank you for reviewing, Ananya! We discussed in one of our daily syncs and decided to close @W-7748717@ because the debug & run command should actually not give warnings (should just run through the test) if there is no checkpoint/breakpoint.

We use SFDX: Update Checkpoints in Org to check if there is any checkpoint available and to upload checkpoints if they exist. If there is no checkpoint, the behavior of the debug & run button is the same with directly running tests using the green run button, except that those SFDX: Update Checkpoints in Org messages are printed. I've attached a screenshot below!

Awesome thank you for reminding me! Is it possible for us to exit out of the Uploading Checkpoints process once we discover that there aren't any checkpoints so we can avoid printing the rest of the messages as well?

[violetyao]

Great point! I agree that the messages are confusing if there is no checkpoint available. I've added a hasOneOrMoreActiveCheckpoints function and attached a screenshot of the outputs if there is no checkpoint available.

Let me know what you think! If the change looks good to you, let's have Sonal review the no_enabled_checkpoints message wording.

[AnanyaJha]

Great point! I agree that the messages are confusing if there is no checkpoint available. I've added a hasOneOrMoreActiveCheckpoints function and attached a screenshot of the outputs if there is no checkpoint available.

Let me know what you think! If the change looks good to you, let's have Sonal review the no_enabled_checkpoints message wording.

Do you think the user needs to be notified that they don't have any checkpoints at all? @jag-sfdc thoughts? If anything, I think it should be a warning modal, not an error one.

[violetyao]

Great point! I agree that the messages are confusing if there is no checkpoint available. I've added a hasOneOrMoreActiveCheckpoints function and attached a screenshot of the outputs if there is no checkpoint available.

Let me know what you think! If the change looks good to you, let's have Sonal review the no_enabled_checkpoints message wording.

Do you think the user needs to be notified that they don't have any checkpoints at all? @jag-sfdc thoughts? If anything, I think it should be a warning modal, not an error one.

I'm leaning towards still giving users warnings if they don't have any checkpoints at all. It might be confusing to users when they try todebug & run and expect that tests stop at checkpoints. I agree that a warning modal is better than an error. Let us know what you think Jag!