On Mon, Oct 30, 2023 at 05:54:01AM -0400, Jeff King wrote: > On Mon, Oct 30, 2023 at 10:45:56AM +0100, Michal Suchánek wrote: > > > > Yes, this is a problem. Those messages are streamed out as we receive > > > them from the server, and before we get any status report back. I don't > > > think we'd want to buffer them, as they can be arbitrarily large (and > > > may even be progress reports that are meant to be shown in real-time). > > > > Not when the user asked to not show progress. > > Sure, if you make the buffering behavior dependent on the presence of > "-q". I guess that is not that hard to do, but it does increase the > complexity of the code (you have both a streaming and a buffering code > path). > > Anyway, I am not all that convinced this is a fruitful path versus just > fixing the server side. But if you want to look into writing a patch, go > for it. I will try to bug the forge upstream first. Thanks Michal