On Sun, Feb 20 2022, Johannes Schindelin wrote: > Hi Junio, > > On Sat, 19 Feb 2022, Junio C Hamano wrote: > >> Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: >> >> > I notice that you did not take this into `seen` yet. I find that a little >> > sad because it would potentially have helped others to figure out the >> > failure in the latest `seen`: >> > https://github.com/git/git/runs/5255378056?check_suite_focus=true#step:5:162 >> > >> > Essentially, a recent patch introduces hard-coded SHA-1 hashes in t3007.3. >> >> I saw the thread, I saw a few patches were commented on, and a few >> were left unanswered, but one was replied by the original submitter >> with a "Good catch!", making me expect the topic to be discussed or >> rerolled to become ready relatively soon. > > Yes, I have local changes, but I had really hoped that this patch series > would get a chance to prove its point by example, i.e. by offering the > improved output for the failures in `seen`. I hoped that because I think > that those improvements speak for themselves when you see them. I think it's a good idea to get wider expose in "seen", "next" etc. for topics where the bottleneck is lack of feedback due to lack of wider exposure. But in this case I've pointed out both direction & UX issues to you that you haven't addressed. Both what I sent a reminder of yesterday in [1], and more relevant to what you're discussing here a reply [2] where I looked & tested your new output v.s. the old, and found that on test failures it: * Replaced summary output with a much more verbose version. * Turned the GitHub UI from usable (but sometimes hard to find the needle in the haystack) to *extremely slow*. Seemingly because the browser was asked to make sense of~30k lines of output, with some of it hidden dynamically by JavaScript. 1. https://lore.kernel.org/git/220220.86bkz1d7hm.gmgdl@xxxxxxxxxxxxxxxxxxx/ 2. https://lore.kernel.org/git/220126.86sftbfjl4.gmgdl@xxxxxxxxxxxxxxxxxxx/
