Hi Junio, On Thu, 28 Jan 2021, Junio C Hamano wrote: > Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: > > > Worse than the `die()` is probably the allocated memory; IIRC there is no > > way to release the memory allocated by `setup_revisions()` unless one > > performs a revwalk. > > Hmph, do you mean the singleton instances of "struct object" and > those specific types that contain it? I am concerned when I see what `add_rev_cmdline()` does, for example. Ciao, Dscho
