Skip to content

Response to issue #1232: "Close garbage PRs" #1242

@davidhedlund

Description

@davidhedlund

I’m unable to post directly to #1232 for some reason, but I’d like to address some of the concerns raised and clarify my approach regarding merging.

the following are mistake-riddled, content-free PRs

I respectfully disagree. Each pull request was carefully prepared and thoroughly reviewed. I have systematically compared every relevant driver, controller, and related issues, tracking all changes in a detailed spreadsheet. My goal has always been to improve the repository’s accuracy and consistency.

that thrash commit history for no sensible reason.

All changes and justifications are clearly documented in the "Conversation" tab of each PR. I have made multiple revisions in response to feedback and new findings—particularly because the controller driver documentation was previously incomplete. As I discovered undocumented behaviors, I updated both the code and the documentation accordingly.

the contributor who opened them has already done enough damage to the project

I acknowledge my earlier mistake in PR #1062 from December 2023, for which I have already apologized. At that time, all changes were fully reverted and the repository was restored. It’s important to note that there was no documented policy regarding these changes at the time, which contributed to the misunderstanding.

Regarding the current PRs, I recognize that the commit history may appear less tidy than ideal. This was primarily due to unclear or missing documentation, which led me to discover and address multiple issues on a daily basis—often requiring frequent updates to each controller file throughout the day. To keep things organized, I found it more convenient to maintain a change log in the PR’s Conversation thread, documenting each update as it was made.

Since then, I have made significant efforts to improve the documentation, as can be seen in my recent contributions. I am committed to learning from these experiences and ensuring that similar issues do not occur in the future.

these should be closed #1140 #1156 #1158 #1159 #1160 #1168 #1170 #1171 #1185 #1186 #1191

They should not be closed. Each PR listed has been carefully reviewed and improved through thorough testing against all relevant drivers and controllers, with significant time and personal resources invested (approximately 80 working hours and $240 for the controllers) to ensure accuracy. I have continuously refined these PRs based on feedback and new information, and where documentation was lacking, I have contributed updates to help future contributors avoid similar issues.

Summary of specific PRs (for full details, please see the Conversation tab on each PR):

Half of the PRs adds support for multiple input_device, input_device_display_name, input_vendor_id, and input_product_id variables.

Only 2 of the 11 PRs, #1160 and #1185, could be considered "content-free," as they simply remove comments (as we agreed on earlier) and sort the variables to improve consistency and make it easier to compare with other autoconfigs.

I understand if you’re frustrated that the PRs are still open. This may be related to concerns about RetroArch’s backward compatibility—Rob may not have approved them yet to ensure that older versions of RetroArch continue to function properly across all platforms.

Best

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions