On Mon, Aug 28, 2023 at 10:03:46PM +0200, Theo Buehler wrote: > On Mon, Aug 28, 2023 at 09:18:40PM +0200, Marc Espie wrote: > > On Mon, Aug 28, 2023 at 08:42:50PM +0200, Marc Espie wrote: > > > Turns out the exec in cmd_exec.c has absolutely zero reason to be > > > different from what engine does. > > > > > > This small patch moves a bit of code around, so that all execv() consumers > > > benefit from the same optimisation (namely, no extra shell when not > > > needed). > > > > > > The only visible change should be that now, VAR!=cmd *will* display > > > some relevant information if exec fails. > > > > > > This is a fairly trivial change, I don't expect any fallout. > > > > > > (still need to run it through tests) > > > > Better with the patch (thx miod@) > > > > Index: cmd_exec.c > > [...] > > Please drop this: > > > +static void retry_with_temp_file(bool, char **); > > Otherwise ok once you're happy with your testing. > > Oh, I thought I had removed the code along with the call.
Yeah, I don't intend to commit that part with the patch anyhow. Gonna run a build or two just to be sure.