Re: [Qemu-devel] [PATCH] gdbstub: Fix memory leak

[Stuart Brady]

On Tue, Oct 18, 2011 at 06:18:11PM +0000, Blue Swirl wrote:

> Cool. Please include the spatch with the commit message.

Thanks, will do!

FWIW, it results in:

 51 files changed, 99 insertions(+), 136 deletions(-)

I wonder if that needs splitting up at all?

> IIRC I had also problems with identifiers, I could not get checking
> identifiers against CODING_STYLE to work.

FWIW, I got there in the end:

   @@ type T; T *E; @@
   -E = g_malloc(sizeof(*E))
   +E = g_new(T, 1)

   @@ type T; T *E; @@
   -E = g_malloc0(sizeof(*E))
   +E = g_new0(T, 1)

   @@ type T; T *E; expression N; @@
   -E = g_malloc(sizeof(*E) * N)
   +E = g_new(T, N)

   @@ type T; T *E; expression N; @@
   -E = g_malloc0(sizeof(*E) * N)
   +E = g_new0(T, N)

   @@ type T; T *E; @@
   -E = g_malloc(sizeof(T))
   +E = g_new(T, 1)

   @@ type T; T *E; @@
   -E = g_malloc0(sizeof(T))
   +E = g_new0(T, 1)

   @@ type T; T *E; expression N; @@
   -E = g_malloc(sizeof(T) * N)
   +E = g_new(T, N)

   @@ type T; T *E; expression N; @@
   -E = g_malloc0(sizeof(T) * N)
   +E = g_new0(T, N)

With this added, I get:

 246 files changed, 514 insertions(+), 557 deletions(-)

That includes the original 99 insertions and 136 deletions.

I think I should at least submit the patch to convert cases involving
casts as the first part of the patch series, and convert the sizeof(*E)
and sizeof(T) cases as the second and third parts.  Sound okay?

Cheers,
-- 
Stuart Brady