Hi Nik, On So 25 Nov 2012 15:43:01 CET Dominik George wrote:
in the ocurse of the BSP Essen I took a close look at this report and the proposed patch. First of all, I am downgrading the severity of the bug to important to remove it from the RC list. The background is that the changes are most likely not trivial enough to get a freeze exception for wheezy and the issue is not critical enough to justify it as RC.
ACK.
Looking at the code, I have a few questions:
addrs = None
try:
_ttl, addrs = resolve_ipv4(hostname)
except:
_ttl, addrs = resolve_ipv6(hostname)
Is preferring IPv4 over IPv6 the expected behaviour? It diverts from the
normal behaviour of other code.
Nope, normally preferring IPv6 over IPv4 is the default. However, the patch introduces IPv6 functionality, but it with the current patch:
o resolve_ipv4 gets called o waits for DNS timout if the address is IPv6-only o resolve_ipv6 gets calledShifting the order around (calling resolve_ipv6, then resolve_ipv4) will result in a IPv6 DNS timeout whenever a hostname is to be resolved.
if ipv4_res is not None:
for res in ipv4_res:
(et al), this looks the same as what was there before, but it is a bug.
The ipv4_res could be False, which is also not iterable.
if ipv4_res:
would be better, proper exception handling would be best.
Probably yes, however, if that patch will not make it into wheezy anyway, then the package maintainer should get upstream's libev rewrite of gevent into Debian unstable (1.0~b1), so it is available for other distros that pull packages in from Debian unstable.
Greets, Mike -- DAS-NETZWERKTEAM mike gabriel, rothenstein 5, 24214 neudorf-bornstein fon: +49 (1520) 1976 148 GnuPG Key ID 0x25771B31 mail: mike.gabr...@das-netzwerkteam.de, http://das-netzwerkteam.de freeBusy: https://mail.das-netzwerkteam.de/freebusy/m.gabriel%40das-netzwerkteam.de.xfb
pgpvEYr8oHDfK.pgp
Description: Digitale PGP-Unterschrift
