This is probably more appropriate here.
mph
Begin forwarded message:
Date: Wed, 18 Jul 2007 18:40:49 -0700
From: Michael Hunter <michael.hun...@sun.com>
To: clearview-discuss at opensolaris.org
Cc: Renee.Danson at Sun.COM
Subject: [clearview-discuss] lantent string termination bug
In wpa_config_parse_ssid() ssid is set to the strdup(value) and
ssid_len is set to the strlen(value). ssid_len is checked to be <=
MAX_ESSID_LENGTH. I think that should probably be < so you can fit in
the termination or the constant should be renamed. I don't see an
obvious place where this is an issue around this code but I'm not
convinced that it is easy for the reader to figure out what your
conventions are (although there are clearly some +1s around that make
the above work). FWIW I tend to use max to mean the max of the value
and len to mean the length (amount) of storage including termination,
quoteing, etc. But in any case what convention you choose is up to you
although it should really be chosen to make the readers job easier.
In dladm_wlan_wpa_get_sr() you memcpy(we_bytes, ssid, ssid_len) which
just happens to work because the array is set to 0 by the caller
although I don't see that being stated explictly. You should state
that this is the assumption somewhere close by or terminate the strings.
No bugs here I think but maybe it would be worthwhile to play a little
slower with strings.
In wpa_supplicant.c:117 should this this be isprint() or is the ssid us
ascii centric? Should the replacement character depend on locale?
FWIW I was looking at this code 'cuz we had a bug where we were
formatting up some wireless information and passing it through popen()
and the shell was interpreting embedded quotes in a way quite
unhealthy. So that bug is our fault, but looking at your code it wasn't
immediately obvious to me that it was entirely copasetic.
mph
_________________________________
clearview-discuss mailing list
clearview-discuss at opensolaris.org
