Tinderbox test logs are dominated by lists of files, generated by unzip -o firefox-3.7a6pre.en-US.win32.zip and especially unzip -o firefox-3.7a6pre.en-US.win32.tests.zip This makes the logs take a long time to load, and creates false matches when searching through the log for interesting bits. Please add -q or -qq to make unzip quiet.
Personally, I'd like to WONTFIX this one. I've been saved a few times by having the list of files a slave unzips available in the logs. More data is more better I think. Is there a way we can reduce false matches? What are you searching on?
How has it helped you? Would you be able to get the same information by downloading the zip file yourself, in that relatively rare case? I'm usually searching on the name of a test or some other short string.
Sample run for a random job "Rev3 Fedora 12 mozilla-central debug test mochitest-other on 2010/06/28 13:14:56" jhford-wifi:~ jhford$ wget http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1277756096.1277757695.31503.gz\&fulltext=1 -O sample.log jhford-wifi:~ jhford$ cat sample.log | wc -l 155221 jhford-wifi:~ jhford$ grep "^ inflating:" sample.log | wc -l 19930 jhford-wifi:~ jhford$ grep "^firefox/" sample.log | wc -l 240 If this is a common problem, try doing something like: curl $URL | grep -v '^ inflating:' | grep -v '^firefox/' > your-processed-log As always, we can filter out data that we do have much easier than we can recreate data that we don't have.