No description provided.
No need to log start time if it's already being done on every line. #1427
pull rebroad wants to merge 1 commits into bitcoin:master from rebroad:StartTimeFix changing 1 files +2 −0-
rebroad commented at 11:31 AM on June 6, 2012: contributor
-
csyangchen commented at 12:19 PM on June 6, 2012: none
I think the old one is preferred if I want to parse Startup time from log file, not caring whether or not the log file tracks timestamp.
PS: I think the timestamp format itself is a little bit redundant. Taking a significant fraction of the log file size.
-
rebroad commented at 1:44 PM on June 6, 2012: contributor
@csyangchen if one is wanting to parse start up times from the log file with timestamps turned on, one wouldn't use the startup time line anyway, but the first line with "Bitcoin version", or the line after "Default data directory". The "startup time" line is entirely redundant with timestamps enabled.
-
csyangchen commented at 4:11 PM on June 6, 2012: none
@rebroad I agree with the point you made. However, a redundant line in debug won't hurt, just not worthy of a separate pull, IMO.
Previously (appeared at least in v0.6.0, fixed by now ) there is a duplicated timestamp output in main.cpp, on the recv message: "timestamp timestamp received: ...". That, I think, is a duplication worthy of fix, considering frequency of the recv message.
If we actually want a start time message, I think we should also provide a stop time message, so that users can parse the debug file and calculate the session time, when timestamp is not provided.
-
Diapolo commented at 9:01 AM on June 7, 2012: none
@csyangchen The duplicate timestamps are fixed, I created a patch for this a few weeks ago. I also think this pull is unneeded.
-
luke-jr commented at 10:16 PM on June 8, 2012: member
It's not duplicate if you have logtimestamps disabled. Just because someone might be interested in a few cases where times matter, doesn't mean they want every line timestamped.
- luke-jr referenced this in commit 19f16ad66f on Jun 22, 2012
-
rebroad commented at 8:35 PM on July 2, 2012: contributor
@luke-jr everything takes so long to compile on my laptop, as due to the way git works, it causes all file modification times to be changed, so make tried to recompile everything. If I could work out a way for make to use checksums instead of modification times, then my compiles wouldn't take so long, and I'd have tested this. thanks for the fix!
-
Diapolo commented at 9:49 PM on July 2, 2012: none
I still think this one is no needed, your other current pulls that re-work debug-messages and stuff make far more sense.
-
jgarzik commented at 4:50 PM on August 1, 2012: contributor
ACK, for ever-so-slightly enhanced privacy
-
BitcoinPullTester commented at 1:40 AM on August 10, 2012: none
Automatic sanity-testing: FAILED MERGE, see http://jenkins.bluematt.me/pull-tester/8617052cb72f4ac96ce5f8dbf024d7cbf41f0351 for test log.
This pull does not merge cleanly onto current master
-
No need to log start time if it's already being done on every line. 016178132e
- jgarzik referenced this in commit 38e8f282d2 on Sep 4, 2012
- jgarzik merged this on Sep 4, 2012
- jgarzik closed this on Sep 4, 2012
- lateminer referenced this in commit 06abba7a14 on Jan 22, 2019
- DrahtBot locked this on Sep 8, 2021