[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-wget] Revisiting AI_ADDRCONFIG
From: |
Maciej W. Rozycki |
Subject: |
Re: [Bug-wget] Revisiting AI_ADDRCONFIG |
Date: |
Wed, 29 Sep 2010 18:39:07 +0100 (BST) |
User-agent: |
Alpine 2.00 (LFD 1167 2008-08-23) |
On Wed, 29 Sep 2010, Hrvoje Niksic wrote:
> >> Note that you can disable IPv6 in Wget itself with --inet4-only or the
> >> equivalent .wgetrc command. That will cause Wget to request AF_INET
> >> family from getaddrinfo and (presumably) avoid the spurious lookup that
> >> causes problems.
> > I could also build a patched wget with that flag set.
>
> This part of my response was about how to avoid the waiting behavior in
> this particular case, even without rebuilding Wget. My understanding of
> the cited text is that it claims this to be impossible without using
> AI_ADDRCONFIG, i.e. patching and rebuilding Wget.
>
> The relevant question for adding AI_ADDRCONFIG would be: other than
> working around this resolver bug, what else is gained by using
> AI_ADDRCONFIG in the current Wget?
FWIW, the configure script may be taught about the known-broken hosts WRT
AI_ADDRCONFIG (it's the configure's role after all) and let the good ones
benefit from it. From own experience you also need to #define
AI_ADDRCONFIG to 0 on hosts that are old/twisted enough not to support it
at all. Last time I checked Windows/MinGW used to fall into this category
for example.
If you are still uneasy about the breakage out there, then how about
making it an AC_ARG_ENABLE feature with the default of "disable"? The
actual (quite straightforward) implementation would be left as an exercise
for the requester.
Maciej