# [PATCH] commit-graph: fix start_delayed_progress() leak

3 messages from 2025-06-04 to 2025-06-04. Participants: Lidong Yan via GitGitGadget, Patrick Steinhardt, lidongyan.
Thread: https://gitlist.dev/t/63579

## Lidong Yan via GitGitGadget, 2025-06-04 03:11

Subject: [PATCH] commit-graph: fix start_delayed_progress() leak
Message-ID: <pull.1986.git.git.1749006675784.gitgitgadget@gmail.com>
URL: https://gitlist.dev/e/pull.1986.git.git.1749006675784.gitgitgadget%40gmail.com

```
From: Lidong Yan <502024330056@smail.nju.edu.cn>

In commit-graph.c:graph_write(), if read_one_commit() failed,
progress allocated in start_delayed_progress() will leak. Add
stop_progress() before goto cleanup.

Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
---
    commit-graph: fix start_delayed_progress() leak
    
    In commit-graph.c:graph_write(), if read_one_commit() failed, progress
    allocated in start_delayed_progress() will leak. Add stop_progress()
    before goto cleanup.

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1986%2Fbrandb97%2Ffix-graph-write-leak-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1986/brandb97/fix-graph-write-leak-v1
Pull-Request: https://github.com/git/git/pull/1986

 builtin/commit-graph.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/builtin/commit-graph.c b/builtin/commit-graph.c
index a783a86e797..ee48980248f 100644
--- a/builtin/commit-graph.c
+++ b/builtin/commit-graph.c
@@ -311,6 +311,7 @@ static int graph_write(int argc, const char **argv, const char *prefix,
 		while (strbuf_getline(&buf, stdin) != EOF) {
 			if (read_one_commit(&commits, progress, buf.buf)) {
 				result = 1;
+				stop_progress(&progress);
 				goto cleanup;
 			}
 		}

base-commit: 7014b55638da979331baf8dc31c4e1d697cf2d67
-- 
gitgitgadget

```

## Patrick Steinhardt, 2025-06-04 07:42

Subject: Re: [PATCH] commit-graph: fix start_delayed_progress() leak
Message-ID: <aD_46Qxh9oVj-P3U@pks.im>
URL: https://gitlist.dev/e/aD_46Qxh9oVj-P3U%40pks.im
In-Reply-To: <pull.1986.git.git.1749006675784.gitgitgadget@gmail.com>

```
On Wed, Jun 04, 2025 at 03:11:15AM +0000, Lidong Yan via GitGitGadget wrote:
> From: Lidong Yan <502024330056@smail.nju.edu.cn>
> 
> In commit-graph.c:graph_write(), if read_one_commit() failed,
> progress allocated in start_delayed_progress() will leak. Add
> stop_progress() before goto cleanup.
> 
> Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>

Nit: it might make sense to send multiple patches that are related, like
your memory leak fixes, in the same patch series. That makes it a bit
easier for reviewers to group together related reviews.

> diff --git a/builtin/commit-graph.c b/builtin/commit-graph.c
> index a783a86e797..ee48980248f 100644
> --- a/builtin/commit-graph.c
> +++ b/builtin/commit-graph.c
> @@ -311,6 +311,7 @@ static int graph_write(int argc, const char **argv, const char *prefix,
>  		while (strbuf_getline(&buf, stdin) != EOF) {
>  			if (read_one_commit(&commits, progress, buf.buf)) {
>  				result = 1;
> +				stop_progress(&progress);

This function calls `stop_progress_msg()`, which knows to exit in case
`*progress` is a NULL pointer. We thus don't have to guard this line
with `if (progress)`. So the patch looks good to me, thanks!

Patrick

```

## lidongyan, 2025-06-04 07:47

Subject: Re: [PATCH] commit-graph: fix start_delayed_progress() leak
Message-ID: <CB88AA24-6D26-4BED-B430-62A453E6B9D6@smail.nju.edu.cn>
URL: https://gitlist.dev/e/CB88AA24-6D26-4BED-B430-62A453E6B9D6%40smail.nju.edu.cn
In-Reply-To: <aD_46Qxh9oVj-P3U@pks.im>

```
2025年6月4日 15:42，Patrick Steinhardt <ps@pks.im> 写道：
> 
> On Wed, Jun 04, 2025 at 03:11:15AM +0000, Lidong Yan via GitGitGadget wrote:
>> From: Lidong Yan <502024330056@smail.nju.edu.cn>
>> 
>> In commit-graph.c:graph_write(), if read_one_commit() failed,
>> progress allocated in start_delayed_progress() will leak. Add
>> stop_progress() before goto cleanup.
>> 
>> Signed-off-by: Lidong Yan <502024330056@smail.nju.edu.cn>
> 
> Nit: it might make sense to send multiple patches that are related, like
> your memory leak fixes, in the same patch series. That makes it a bit
> easier for reviewers to group together related reviews.

Got it, though I think this is the last time I send leak-fix patches.

> 
>> diff --git a/builtin/commit-graph.c b/builtin/commit-graph.c
>> index a783a86e797..ee48980248f 100644
>> --- a/builtin/commit-graph.c
>> +++ b/builtin/commit-graph.c
>> @@ -311,6 +311,7 @@ static int graph_write(int argc, const char **argv, const char *prefix,
>> while (strbuf_getline(&buf, stdin) != EOF) {
>> if (read_one_commit(&commits, progress, buf.buf)) {
>> result = 1;
>> + stop_progress(&progress);
> 
> This function calls `stop_progress_msg()`, which knows to exit in case
> `*progress` is a NULL pointer. We thus don't have to guard this line
> with `if (progress)`. So the patch looks good to me, thanks!
> 
> Patrick
> 

Thanks,
Lidong


```
