search

ravthan / all-eyes

Automatically exported from code.google.com/p/all-eyes
0 stars 0 forks source link

The size of the timestamps in the message ids is not consistent across languages #95

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago 
The timestamp is the first part of the message id.  The log shows timestamps of 
length 13, 14, and length 15. This leads me to believe the size of the 
timestamps in the message ids is not consistent across languages.

Original issue reported on code.google.com by tbrt....@gmail.com on 20 Nov 2012 at 4:32

GoogleCodeExporter commented 8 years ago 
for filemon and filedescr, i was trying to follow a pattern of 10 character 
milliseconds, and 5 char random number.  But I think sometimes the random 
number is ranging from 3 to 5 characters.  Need to investigate.

I am following from wiki:

  "A string, combination of up to 15 digits, append char '-', append count (1 digit 
  to 6 digits). The 15 digits are from a combination of a high resolution timer that 
  may not corresponds to a real time. The count is the message counter of the 
  monitor process." 

I have a question, if we can just use the 10 digit millisecond?

Original comment by heyyoul...@gmail.com on 20 Nov 2012 at 6:56

GoogleCodeExporter commented 8 years ago 
Fixed in selfmon to be 16 characters for messageID(timestamp); now it is in 
microseconds.

r958

Original comment by ravt...@gmail.com on 21 Nov 2012 at 12:53

GoogleCodeExporter commented 8 years ago 
Fixed in socketmon and procmon to be 16 characters for messageID(timestamp). It 
is in mcroseconds.

r959

Original comment by toddd...@gmail.com on 21 Nov 2012 at 1:12

GoogleCodeExporter commented 8 years ago 
Fixed in filemon and filedescr monitors to be 16 characters for 
messageID(timestamp). It is in microseconds.

r963

Original comment by toddd...@gmail.com on 21 Nov 2012 at 4:07

GoogleCodeExporter commented 8 years ago 
The aeManager and aeProxy have been updated to have a micro second timestamp 
effectively adding three zeros to the millisecond number.  I do not see the 
benefit making the timestamp microseconds as it not a common cross language 
timestamp.

Original comment by tbrt....@gmail.com on 21 Nov 2012 at 4:32

GoogleCodeExporter commented 8 years ago 
Verified with unit testing

Original comment by tbrt....@gmail.com on 9 Dec 2012 at 9:24