Update commit_and_push_changes function to add a default commit message

[ziadhany]

• Add a default print logger to prevent exceptions
• Bump minecode_pipelines version to 0.0.1b24

[keshav-space]

@ziadhany See comment below.
Also why use print? why not simply pass self.log while calling commit_and_push_changes? Spraying custom logging mechanisms in pipeline functions defeats the whole purpose of having a centrally controlled logger managed by the pipeline, which helps in proper dissemination of pipeline logs to the run instance and UI.

[keshav-space]

Can you explain why we need to compute commit_message here? when we already take care of missing commit message in commit_changes function.

[ziadhany]

@keshav-space The main error comes from If a pipeline fails before updating a checkpoint, when a new pipeline runs, it starts and tries to commit some data that was already pushed. This raises the error:
Nothing to commit, working tree clean This disrupts the pipeline and causes it to fail

One way to handle this is to use commit_and_push_changes, since it pushes and commits at the same time and can catch GitCommandError.

Why add a print logger by default?

• This prevents pipeline disruption and avoids unreadable errors such as 'NoneType' object is not callable.' Since the logger is optional, using a print statement with a False return provides a simple and safe fallback.

Why add commit_and_push_changes?

• To avoid duplicating commit messages for every ecosystem (e.g., CRAN, Swift, Composer).

[keshav-space]

@ziadhany

The main error comes from If a pipeline fails before updating a checkpoint, when a new pipeline runs, it starts and tries to commit some data that was already pushed. This raises the error: Nothing to commit, working tree clean This disrupts the pipeline and causes it to fail

One way to handle this is to use commit_and_push_changes, since it pushes and commits at the same time and can catch GitCommandError.

Please use commit_and_push_changes we already handle empty commit scenario see #1888, no special error handling is needed on call side.

Why add a print logger by default?

This prevents pipeline disruption and avoids unreadable errors such as 'NoneType' object is not callable.' Since the logger is optional, using a print statement with a False return provides a simple and safe fallback.

Let's not use print as default logger and if needed we should instead make logger required field that would be much simpler.

Why add commit_and_push_changes?

To avoid duplicating commit messages for every ecosystem (e.g., CRAN, Swift, Composer).

I don't get you here.
Can you explain why the below fallback commit is not sufficient for CRAN, Swift and Composer?

scancode.io/scanpipe/pipes/federatedcode.py

Lines 229 to 242 in 23b94ad

	if not commit_message:
	author_name = settings.FEDERATEDCODE_GIT_SERVICE_NAME
	author_email = settings.FEDERATEDCODE_GIT_SERVICE_EMAIL
	purls = "\n".join(purls)
	commit_message = f"""\
	Add {mine_type} results for:
	{purls}
	Tool: {tool_name}@v{tool_version}
	Reference: https://{settings.ALLOWED_HOSTS[0]}
	Signed-off-by: {author_name} <{author_email}>
	"""

And why would we need to compute default commit message in commit_and_push_changes when we already have default commit message in downstream commit_changes function?

[ziadhany]

@keshav-space Ok, I get your point. I didn’t notice we already have a default commit message, and I will close this PR.