pytorch / pytorch/vision

[Feedback] Warn Developers Regarding Major Code Changes

Open
#5,710 6 comments 5 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
17.9k
Forks
7.3k
Avg merge
1d 15h
Merged PRs (30d)
13

Description

🚀 The feature

Flag parts of the repo that are undergoing major changes (breaking changes between two releases) + have a bot post on PRs that touch these high traffic files.

Motivation, pitch

TL;DR: torchvision moves pretty fast and external developers want to know if they are developing on something that may change very soon.

Imagine you are an interested developer and you would love to contribute to torchvision. You scan through the issues and you see an interesting one you can help tackle! You follow some directions, get right into the task, and a few days later, you have a PR. You iterate based on CI results and you're pretty proud of your feature.

You get a reviewer to look through and hopefully approve your awesome change. BUT that's not what happens. Instead, the reviewer lets you know that the code you're touching is actually migrating to some other structure and the work you did on your PR has become irrelevant. What a downer. You wish you knew this earlier, before you spent so much time on the PR.

The current system can be quite discouraging for an external developer. To promote contributions, torchvision should notify developers of major code changes in some way. One way is to flag those paths that are likely changing soon and having a bot post on PRs touching those paths with a warning message. This way, developers have more clarity on what they should work on.

Alternatives

Other alternatives are for torchvision to update existing issues/feature requests with up-to-date content so that developers can fix an issue knowing that their work would not be in vain.

Additional context

This feedback was gathered from talking to current external contributors.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are identified in the issue. Start by mapping how torchvision tracks changes across releases and how pull-request automation is configured; done means major-change paths can be flagged and PRs touching them receive a warning.

Written by the indexing model from the issue text.

Assessment

Domain
developer-experience
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.