Le 8 août 2011 à 21:08, Anders Peter Fugmann a écrit :

> Hi Xavier,
> 
> On 2011-08-04 15:03, [email protected] wrote:
>> Could you elaborate on what was missing, or buggy in Bolt?
>> As long as it fits with its design, I may try to fix it to suit your needs.
> 
> When we tried bolt, the main issue was log rotation. We wanted to make sure 
> that logs were rotated so they would not fill up all disk space.
> 
> Bolt does indeed do log rotation, but the files created files by bold 
> includes either timestamp or pid, which changes between program invocation. 
> This makes to harder to see which file is the current, and leaves a lot of 
> log files during development.

The ability to put either the timestamp or the pid is just a possibility;
you can as well just use a "constant" file name. In that latter case, the
file will be replaced at each rotation.


> We therefore decided to send all log to stdout, and capture it though a 
> simple program that would reopen the file when it received SIGHUP.

The feature of rotating not on time information but on the reception
of a signal seems quite easy to add.


> (Logging to stdout is also helpful when developing).

My bad, the documentation is lacking the following information:
if you use "<stdout>" (without the quotes) as the filename, then
the data will be written on standard output.


> As we no longer used Bolt for log rotation, we decided to leave bolt, keeping 
> external dependencies down.
> 
> If bolt had the ability to reopen the log file on SIGHUB (or a function to 
> reopen the log file), we would probably still be using it. That combined with 
> the ability to log to stdout when developing would make it very useful.

As I have other incentives to publish an updated version of Bolt,
I will try to release a version featuring that signal thing "soon".


Regards,

Xavier Clerc



-- 
Caml-list mailing list.  Subscription management and archives:
https://sympa-roc.inria.fr/wws/info/caml-list
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

Reply via email to