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

Bug#761923: netcfg: please enable wpa_supplicant syslog and debug flags



Control: tag -1 patch pending

Cyril Brulebois <kibi@debian.org> (2014-09-16):
> Source: netcfg
> Version: 1.120
> Severity: normal
> Control: block -1 by 761922
> 
> Hi,
> 
> see discussion in [1] for background, and #761922 for the requested
> change in the wpa package. To enable this syslog mode, netcfg needs
> to pass the -s flag in its execlp call (wpa.c) but it MUST NOT do so
> before wpa has been uploaded (possibly migrated to testing) since
> wpa_supplicant would exit due to an unknown flag otherwise.
> 
>  1. https://lists.debian.org/debian-boot/2014/09/msg00517.html
> 
> It would also probably make sense to enable at least the first debug
> level (-d), and maybe the second one (-dd), even if I'm not sure.
> There might be a way to make it configurable, so that we can ask
> users to perform a single change when they report WPA-related
> issues?
> 
> I'd probably go for -s and -d in the next upload; and see how often
> we need -dd after that.

Testing with wpa 2.2-1 (as staged in wpa's svn as of r1889), that means
200+ wpa_supplicant lines on success, and 1200+ lines if I fail to enter
the right passphrase.

That might be a tad too much to be useful. Maybe having the possibility
to set -d or -dd “manually” (in case of errors) would be better. I'm
tempted to actually drop the -d flag for the next upload, and get back
to my initial idea: another bug report to track adding a “moar debug”
option.

Anyway, I've stacked the patch I've just tested in the pu/wpa branch for
now, and I'm tagging this bug report with patch + pending so that nobody
wastes time on it.

Mraw,
KiBi.

Attachment: signature.asc
Description: Digital signature


Reply to: