Re: [PATCH 3/3] odb: properly close sources before freeing them
- From
Eric Sunshine <sunshine@sunshineco.com>
- Date
- Dec 6, 2025, 11:43 UTC
- Message-ID
- <CAPig+cQNKQt=kMaNYNWAPAfGej-mhLUR_BXS4J58JjVUtG7VKw@mail.gmail.com>
- In-Reply-To
- <aTQVt4zgMbsX_6tD@pks.im>
On Sat, Dec 6, 2025 at 6:38 AM Patrick Steinhardt <ps@pks.im> wrote:
Show 27 quoted lines
> On Fri, Dec 05, 2025 at 06:14:22PM -0500, Eric Sunshine wrote: > > On Fri, Dec 5, 2025 at 6:36 AM Patrick Steinhardt <ps@pks.im> wrote: > > > In the next commit we are about to move the packfile store into the ODB > > > source so that we have one store per source. This will lead to a memory > > > leak in the following commit when reading data from a submodule via > > > git-grep(1): > > > [...] > > > Signed-off-by: Patrick Steinhardt <ps@pks.im> > > > > Considering that this is patch [3/3], to what does "In the next > > commit..." refer? > > Good catch! I split this out of another, bigger, patch series. But as > I've started to hit the leak in a different patch series, as well, I > decided to split it out into a smaller patch series. > > I've queued the following change locally, but will refrain from sending > out a new version for now. > > 3: a06d0716c3 ! 3: c9ca233c29 odb: properly close sources before freeing them > @@ Commit message > In the next commit we are about to move the packfile store into the ODB > - source so that we have one store per source. This will lead to a memory > - leak in the following commit when reading data from a submodule via > - git-grep(1): > + source so that we have one store per source. This can lead to a memory > + leak when reading data from a submodule via git-grep(1):
I would think that you would also want to drop the "In the next commit we are about to..." bit (considering, again, that this is patch [3/3]).