markrmiller commented on pull request #365: URL: https://github.com/apache/lucene/pull/365#issuecomment-940786135
On Tue, Oct 12, 2021 at 2:45 AM Uwe Schindler ***@***.***> wrote: > It was pretty surprising this issue hid out for so long. I've run it on 5 > towers, a mini desktop box, a couple raspberry pi's, across Ubuntu, > Arch/Manjaro, x86, arm, bash, zsh, a couple others have ran it. Shell > scripts ... > > The problem with gradle is that it enables "colorful" output if your > console supports it. E.g., if you SSH to a remote machine, gradle often > isn't colorful. But on local terminals it often is. Not sure what makes it > think that the terminal supports colorful terminal escapes when you pipe > the gradle output to somewhere else. > That sounds reasonable, my runs on other machines are usually driven by Zeppelin using Java to call out to to a shell ssh command. It's still seems strange I had no problem on my primary machine, nor did Mike D seem to, but perhaps easier for 2 cases to have an env that circumvents it. Also, I occasionally do run it over ssh in colorful mode over the last couple months because I use WezTerm these days. No problem, just curious I tend to agree and with @dweiss <https://github.com/dweiss>: Just print > the classpath to a file. > > Don't forget to also fix the same issue in Solr's benchmark module! > I'm actually moving to cheating to do this for Lucene at this point. Initially, I just intended to hack together something that could resemble the Luceneutil standard search benchmark, so that I could easily make some observations. But I realized that if I properly structure this, the similar experience and environment will benefit Solr, and the Lucene end will cost me very little. That does end up meaning that everything I do that is or can be common will be in sync, even if one might step forward before the other. Even just using one part of the newer data gen API in the silly sample benchmark I added here led me to do a much-needed, comprehensive clean up and improvement pass on that data gen API. That will all go into Solr. I am putting in the nonembryonic documentation needed for Solr this week as well, and also structuring that so I can hit both, get some advantage out of that, and eat the minimal additional cost. FYI, I also am likely to rework this pull request setup a bit soon. I will have a repo that I update more frequently that will less often merge to my markrmiller repo, which will update this pr. I only recently learned that while Apache has unlimited action minutes, they do have limited throughput for the whole organization and I don't want to eat them here. Mark > -- - Mark http://about.me/markrmiller -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
