# Incremental GC and pushing back the 3.13.0 release

**URL:** <https://discuss.python.org/t/incremental-gc-and-pushing-back-the-3-13-0-release/65285>\
**Category:** Core Development\
**Tags:** release\
**Created:** [September 28, 2024, 12:43am UTC](https://discuss.python.org/t/incremental-gc-and-pushing-back-the-3-13-0-release/65285 "2024-09-28T00:43:56Z")\
**Posts on this page:** 1\
**Showing post:** 1

<div class="post-metadata">

**Author:** ![thomas](https://sea2.discourse-cdn.com/flex002/user_avatar/discuss.python.org/thomas/32/5492_2.png) [@thomas](https://discuss.python.org/u/thomas)\
**Post date:** [September 28, 2024, 12:43am UTC](https://discuss.python.org/t/incremental-gc-and-pushing-back-the-3-13-0-release/65285/1 "2024-09-28T00:43:56Z")

</div>

I’m a little \[1\] concerned with the impact of the incremental GC change in 3.13, which recently showed up ([Incremental GC means that Sphinx is significantly slower in 3.13 than in 3.12 · Issue #124567 · python/cpython · GitHub](https://github.com/python/cpython/issues/124567)). It’s not clear that the incremental GC provides significant improvements (although the smaller pauses are probably desirable), it clearly has slightly more overhead in common cases, and we’re still discovering pathological cases. I don’t think we should release 3.13.0 with the incremental GC.

Obviously taking it out at this stage is a drastic change, and not something I want to do without a new release candidate. There’s also a bunch of other fixes that went in since rc2, so I really want rc3 anyway. Which means pushing back the release to make room for the new release candidate.

What I would like to do:

1. Roll back the incremental GC change.
2. Release 3.13.0rc3 this Monday, September 30th.
3. Release 3.13.0 final a week later, on Monday, October 7th (probably starting the release process on Sunday).

Since I am the release manager for 3.13, I think I will just do what I want to do 😛 .

Rolling back the incremental GC sounds drastic but after talking to a few people about this we think it’s not that big a deal. The code hasn’t changed much since, and we’re rolling back to known old code, which is less risky than attempting to work around the pathological cases in the new code. A week is not a long time for a new release candidate, but it should be enough to find any obvious issues with the rollback (and the other changes that have gone into 3.13 since rc2).

(I don’t think the incremental GC needs to be rolled back in main (3.14) at this point, but it does feel like we need a bit more of a value proposition for the change. I’m glad I can leave that up to @hugovk 😂)

Obviously a shift in schedule like this is annoying for redistributors, so we’re already reaching out to some of the usual folks for this.

* * *

1. somewhat of an understatement

---

_[View the full topic](https://discuss.python.org/t/incremental-gc-and-pushing-back-the-3-13-0-release/65285)._