Difference between revisions of "Timestamp"
From WA2IAC Wiki
m (→Generating Timestamps) |
m |
||
| Line 9: | Line 9: | ||
There are many ways to do this, "this" being "program_being_debugged". | There are many ways to do this, "this" being "program_being_debugged". | ||
| − | + | Here are the methods explored below: | |
*shell | *shell | ||
| Line 18: | Line 18: | ||
Of course there are some other ways to do this: write a 'C' program (don't know if 'ts' is written in C or something else), try various other languages. However, the scope of this project was "quick & dirty" with the emphasis on quick. Nonetheless, I managed to carve out the time necessary to test the methods above. | Of course there are some other ways to do this: write a 'C' program (don't know if 'ts' is written in C or something else), try various other languages. However, the scope of this project was "quick & dirty" with the emphasis on quick. Nonetheless, I managed to carve out the time necessary to test the methods above. | ||
| + | |||
| + | Note that when debugging, we often want to 'tail' the output to watch what's going on in realtime, and possibly use our monkey minds to attempt to correlate with external events. Like so: | ||
| + | |||
| + | <pre> | ||
| + | program_being_debugged 2>&1 | timestamp_generator | tail -f | ||
| + | </pre> | ||
| + | |||
| + | In this case, it's important to minimize buffering along the pipeline so that output gets to our optical sensors as quickly as possible. This can be done with the "unbuffer" command, which requires installation of the 'expect' package (expect-dev on some distros). | ||
==Results== | ==Results== | ||
| Line 27: | Line 35: | ||
</pre> | </pre> | ||
| − | + | Summary: | |
| + | Perhaps surprisingly, the awk and perl versions performed the best. The difference between the perl and awk versions were slight. The perl version, at most, used 10% more memory and resources than the awk version. Very impressive for the "kitchen sink" language. | ||
| + | The tests were performed on a machine under Ubuntu 14.10. Here are the results: | ||
| Line 35: | Line 45: | ||
| − | |||
| − | |||
| − | ==daemontools== | + | ===shell=== |
| + | bash version | ||
| + | <pre> | ||
| + | 188.36 user | ||
| + | 779.17 system | ||
| + | 13:25.67 elapsed | ||
| + | 120% CPU | ||
| + | (0avgtext+0avgdata 6072maxresident)k | ||
| + | 0inputs+0outputs | ||
| + | (0major+28033858minor)pagefaults | ||
| + | 0 swaps | ||
| + | </pre> | ||
| + | |||
| + | ===awk=== | ||
| + | awk version | ||
| + | <pre> | ||
| + | 6.49 user | ||
| + | 12.23 system | ||
| + | 0:12.42 elapsed | ||
| + | 150 % CPU | ||
| + | (0avgtext+0avgdata 6072maxresident)k | ||
| + | 0inputs+0outputs pagefaults | ||
| + | 0 swaps | ||
| + | </pre> | ||
| + | |||
| + | ===daemontools=== | ||
| + | |||
| + | daemontools | ||
| + | <pre> | ||
| + | 6.03 user | ||
| + | 17.68 system | ||
| + | 0:13.12 elapsed 180%CPU (0avgtext+0avgdata 6080maxresident)k | ||
| + | 0inputs+0outputs | ||
| + | (0major+139066minor)pagefaults | ||
| + | 0swaps | ||
| + | </pre> | ||
| + | |||
| + | ===ts=== | ||
| + | ts version | ||
| + | <pre> | ||
| + | 16.81 user | ||
| + | 14.19 system | ||
| + | 0:16.22 elapsed | ||
| + | 191 %CPU | ||
| + | (0avgtext+0avgdata 7864maxresident)k | ||
| + | 0inputs+0outputs | ||
| + | (0major+210354minor)pagefaults | ||
| + | 0 swaps | ||
| + | </pre> | ||
| + | |||
| + | ===perl=== | ||
| + | |||
| + | perl | ||
| + | <pre> | ||
| + | 6.78 user | ||
| + | 11.80 system | ||
| + | 0:12.20 elapsed | ||
| + | 152% CPU | ||
| + | (0avgtext+0avgdata 7868maxresident)k | ||
| + | 0inputs+0outputs | ||
| + | 0major+141587minor)pagefaults | ||
| + | 0swaps | ||
| + | </pre> | ||
| + | |||
| + | ==Random== | ||
| + | |||
| + | ===Shell History=== | ||
| + | |||
| + | Shell History Timestamping | ||
| − | = | + | HISTTIMEFORMAT takes format string of strftime. Check out the strftime manual to choose and construct the timestamp that suit your taste. My favorite is “%F %T “. |
| + | export HISTTIMEFORMAT="%F %T " | ||
| − | + | Execute history again and you will see the effect on the spot, bare in mind that the timestamp for command lines that executed at previous sessions may not valid, as the time was not tracked. | |
