Rename the --driver global option for consistency with forge naming

[0x2b3bfa0]

After #1073 (comment), the --driver option should be renamed to whichever name we choose¹ to describe forges.

Footnotes

1. To my mind, --forge would be the mot juste. 🤷🏼‍♂️ ↩

[0x2b3bfa0]

“What generic term for ‘GitHub/GitLab/Bitbucket’ do you think users will understand best?”

Top 5 choices of 9 based on an internal ad populum poll of @iterative members:

term	votes	percentage
git provider	16	46%
git host	12	34%
forge	3	9%
git forge	2	6%
git site	1	3%

[dacbd]

unclear, discuss what you want to call GitHub Gitlab and bb sure, but functionally this option selects the correct lib to make and interrupt API calls and responses from said term; I believe that very much meets the definition of driver in the context of cml

[0x2b3bfa0]

functionally this option selects the correct lib to make and interrupt API calls and responses from said term

Sure, but that's an implementation detail. 🙃

I believe that very much meets the definition of driver in the context of cml

That's how it's being named in CML internals, indeed, but from a user experience¹ standpoint it's rather meaningless.

Footnotes

1. Not to mention developer experience, but that's for another day. ↩

[0x2b3bfa0]

Scope clarification: this issue is mostly about renaming --driver to something else that's more descriptive and bears an accurate domain-specific meaning; I don't mind if it's --forge or --provider as per the poll above.

[dacbd]

I disagree that this flag should be renamed.

As far as I read in your linked thread is about the wording in the description and docs, not the flag itself.

[0x2b3bfa0]

After #1272, this issue is about choosing an adequate word to describe forges and using it to replace driver in both code and documentation.