Re: [PATCH 1/5] pack-objects: plug leak in `read_stdin_packs()`
- From
Taylor Blau <me@ttaylorr.com>
- Date
- Mar 25, 2026, 23:03 UTC
- Message-ID
- <acRpruzeYfivaCs7@nand.local>
- In-Reply-To
- <acI_pTWTcJN6QaK1@pks.im>
On Tue, Mar 24, 2026 at 08:39:17AM +0100, Patrick Steinhardt wrote:
Show 16 quoted lines
> On Thu, Mar 19, 2026 at 06:24:15PM -0400, Taylor Blau wrote: > > The `read_stdin_packs()` function added originally via 339bce27f4f > > (builtin/pack-objects.c: add '--stdin-packs' option, 2021-02-22) > > declares a `rev_info` struct but neglects to call `release_revisions()` > > on it before returning, creating a leak. > > > > The related change in 97ec43247c0 (pack-objects: declare 'rev_info' for > > '--stdin-packs' earlier, 2025-06-23) carried forward this oversight and > > did not address it. > > > > Ensure that we call `release_revisions()` appropriately to prevent a > > leak from this function. > > Would be curious to learn why none of our tests fail with this. The fix > looks obviously correct though, and there are no other early exits that > might need fixing here.
I believe it's because the only memory we allocate here is in revs->pending, but we copy it to old_pending in prepare_revision_walk() and all object_array_clear() on it.
Since our traversal doesn't use any other fields of rev_info that would cause it to allocate memory, we don't see any leaks in our tests.
I'll rewords the commit message to clarify that this is preventing the *potential* of a leak, not an actual leak.
Thanks, Taylor