On Sun, Aug 25, 2019 at 02:57:48AM -0400, Jeff King wrote: > And I think this is actually a real bug in the current code! We keep a > pointer to the encoding string, which survives because of the history. > But that history is bounded, and we could have an indefinite number of > changed files in the middle. If I modify t9300 like this: Here are two patches. The first fixes the existing bug with "encoding", and the second uses the approach I suggested to fix the leak you noticed. The second one does carry a greater risk of regression than your patch, but I think it's worth it for the fact that it makes any other bugs (like the "encoding" one) more obvious. [1/2]: fast-import: duplicate parsed encoding string [2/2]: fast-import: duplicate into history rather than passing ownership fast-import.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) -Peff
