nickbnf
Crashed on search in log files with 5 million lines
@yurivict could you upload your file somewhere and provide search pattern that caused that crash? Might be some FreeBSD issue.
I selected 5 sample lines from the log file. Please run this shell script:
echo "[16:05:30] Site:1Gfey7..fcdp Try to get updates from: Peer:59.78.171.105 Left: [<Peer:74.82.60.33 >, <Peer:36.149.3.27>, <Peer:110.185.40.144>, <Peer:82.132.224.203>, <Peer:217.20.210.113>, <Peer:74.82.160.65 >, <Peer:207.229.128.38>, <Peer:101.229.178.5>, <Peer:171.15.127.81>, <Peer:220.118.79.25>, <Peer:154.16.127.99>, <Peer:131.21.181.186>, <Peer:119.4.121.73>, <Peer:221.125.176.145>, <Peer:219.147.140.204>, <Peer:74.82.60.164 >, <Peer:113.57.33.71>, <Peer:36.51.179.142>, <Peer:221.130.50.206>, <Peer:212.169.111.134>, <Peer:123.139.86.70>, <Peer:201.185.39.103>, <Peer:131.219.222.194>]" > fake.log
echo "[16:05:30] FileServer Conn# 5 174.182.60.33 [?] > Connecting..." >> fake.log
echo "[16:05:44] Site:1CpQfw..GP2o Announced types ['onion'] in mode startup to 4 trackers in 10.001s, errors: [], slow: ['1.74s http://tracker.opentrackr.org:1337/announce', '1.36s http://explodie.org:6969/announce', '10s+ http://tracker1.wasabii.com.tw:6969/announce']:" >> fake.log
echo "[16:05:46] Site:138R53..NXLP Http tracker tracker1.wasabii.com.tw:6969/announce error: 0x06: TTL expired" >> fake.log
echo "[16:05:55] Site:1BLogC..AGg8 183.14.133.178:15441 Getting connection error: SOCKS5Error: 0x06: TTL expired in Peer.py line 83 > ConnectionServer.py line 152 (connection_error: 1, hash_failed: 0)" >> fake.log
(for i in `seq 100000`; do cat fake.log fake.log fake.log fake.log fake.log fake.log fake.log fake.log fake.log fake.log; done) > fake5M.log
Then search for a common pattern there: TTL. It stars very fast, but then slows down progressively, and never finishes.
As far as I understand the slowdown is mostly due to matches overview (the area near vertical scrollbar that shows approximate matching line positions). It updates on UI thread and processes all matches on every update thus getting slower during search. Refactoring this behavior is on my todo list for next klogg milestone (this is my fork, see issue 18). I'll do PR for glogg if code changes are not very extreme.
Unfortunately I can't reproduce the crash with either glogg or klogg on my linux box. Did you build glogg from source? Was it 32- or 64-bit version? Could you reproduce the crash under debugger to get call stack?
I couldn't reproduce crashes when I tried today either. Unfortunately, I deleted the original file. But this slowdown is pretty bad as it is.