On 30/05/20 4:52 AM, connor.r.no...@gmail.com wrote:
In an effort to clean up my python logging practices when creating libraries, I have
begun reading into "Advanced Logging" and converting my logging practices into
logging configuration `.ini` files:
[link](https://docs.python.org/3.4/howto/logging.html#configuring-logging)
My question is: When defining a FileHandler in a `.ini` file, all of the
examples that I've seen hardcode the name and location of the log file. In the
code snippet below, `python.log` is hardcoded into the `.ini` file:
```
[handler_hand02]
class=FileHandler
level=DEBUG
formatter=form02
args=('python.log', 'w')
```
[code reference
link](https://docs.python.org/3.4/library/logging.config.html#logging-config-fileformat)
Desired Behavior:
On each run of the program, a new logfile is created in the `logs/` directory named
"<timestamp>_program.log".
Current Behavior:
The format in the example above overwrites a single file called "python.log" on
each run, which is not the desired behavior.
Question: Is there a standard procedure for using .ini files to create a new
logfile prepended with the current Unix timestamp on each run of the program?
There is nothing in the PSL to do this (to my knowledge) -
correction/education welcome...
However, it is a fairly standard pattern to take a base-name from a
config file (.ini in MSFT wording) and then to post-process to suit the
application's purposes.
In this case, adding a timestamp makes sense - although perhaps a more
human-readable option unless file-name length is a concern.
The pattern is to:
Compute the (log) file-name.
Check to see if it name already exists (in this directory).
Re-compute if necessary - rinse-and-repeat.
Something like: Make unique file name
https://code.activestate.com/recipes/577200-make-unique-file-name/
--
Regards =dn
--
https://mail.python.org/mailman/listinfo/python-list
