From: Karthik Nayak Date: Thu, 22 Jan 2026 09:05:44 GMT Subject: Re: [PATCH v3 6/6] fetch: delay user information post committing of transaction Message-ID: In-Reply-To: <67808128-5f41-4ae5-b224-a76048cd82c8@gmail.com> Phillip Wood writes: > Hi Karthik > > On 20/01/2026 09:59, Karthik Nayak wrote: > >> +struct ref_update_display_info { >> + bool failed; >> + char success_code; >> + char fail_code; >> + const char *summary; >> + const char *fail_detail; >> + const char *success_detail; >> + const char *ref; >> + const char *remote; >> + struct object_id old_oid; >> + struct object_id new_oid; >> +}; > > I was expecting that we'd pass around a struct like > > struct ref_update_display_info_array { > size_t alloc, nr; > ref_update_display_info *info; > }; > > rather than passing a pointer, count pair as separate parameters. That > would also allow us to use ALLOC_GROW() rather than reallocating the > array each time we append to it which is rather inefficient. > > Thanks > > Phillip > That's fair, I was considering an array and didn't see the need, but using 'ALLOC_GROW()' does make it simpler, plus we'd totally remove the need for the double pointer. Will change. Thanks!