gnunet-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] gnunet-arm behaviour


From: Schanzenbach, Martin
Subject: Re: [GNUnet-developers] gnunet-arm behaviour
Date: Tue, 30 Apr 2019 20:26:24 +0200


> On 30. Apr 2019, at 20:00, Christian Grothoff <address@hidden> wrote:
> 
> Signed PGP part
> On 4/30/19 2:08 PM, Schanzenbach, Martin wrote:
>> Hi,
>> 
>> everytime I show somebody how to start gnunet, the behaviour of gnunet-arm 
>> seems to be a major pain point because it exhibits two behaviours which, 
>> combined, are quite odd.
>> Those two are:
>> 
>> 1. gnunet-arm -s does not hang but return the user to the terminal
>> 2. Logging by default is in that same terminal
>> 
>> (2) basically defeats all advantages (1) would give the user.
>> Can we either change (1) or (2) in order to make it more consistent with 
>> other tools?
> 
> I'd be happy to see (2) change, but somehow I thought that in the
> systemd age logging to stderr was the 'consistent' thing to do (having
> systemd then manage logging).
> 
> Anyway, maybe we should log to stderr as any gnunet-service (and expect
> systemd to launch gnunet via gnunet-service-arm instead of gnunet-arm!)
> and have the command-line tool gnunet-arm redirect logging to a file.
> I think that would indeed make the most sense (but I'm not sure if that
> is "consistent with other tools", given that you left "other" rather
> wide open).
> 
>> So either have gnunet-arm -s hang and on Ctrl-C it also implicity behaves 
>> like gnunet-arm -e
>> OR
> 
> Definitively not this one. If you want that, run gnunet-service-arm
> directly.
> 
>> By default, log into a file, maybe even have a gnunet-log or gnuent-arm 
>> --logs in the future.
> 
> Eh, you are aware of the "-l FILE" option of gnunet-arm? I wouldn't mind
> having a configuration setting that by default logs to some log file,
> see arm.conf:18 for an example. Maybe we should have a separate option
> just for gnunet-arm, so it works _only_ for gnunet-arm and without
> requiring the user to pass -l explicitly every time?

Doesn't that only affect gnunet-arm's logs? But my issue was more with sane 
defaults. I already through that there might be a switch for this, but the 
default should be silent I think. At least for gnunet-arm.

> 
> 
> 

Attachment: signature.asc
Description: Message signed with OpenPGP


reply via email to

[Prev in Thread] Current Thread [Next in Thread]