gnutls-devel
[Top][All Lists]
Advanced

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

Re: gnutls-cli should handle network errors more gracefully


From: Nikos Mavrogiannopoulos
Subject: Re: gnutls-cli should handle network errors more gracefully
Date: Mon, 26 Nov 2012 22:26:44 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:10.0.10) Gecko/20121027 Icedove/10.0.10

On 11/26/2012 08:54 PM, Mantas Mikulėnas wrote:

> On Mon, Nov 26, 2012 at 8:44 PM, Nikos Mavrogiannopoulos
> <address@hidden> wrote:
>> On Sun, Nov 25, 2012 at 10:10 PM, Mantas Mikulėnas <address@hidden> wrote:
>>
>>>>> The expected result here is that `gnutls-cli` would keep trying other
>>>>> addresses returned from getaddrinfo(), until at least one succeeds.
>>>>
>>>>
>>>> I'm curious, how do other networking programs avoid this issue? Do they
>>>> keep trying or do they provide an option to force IPv4 or 6?
>>>
>>> The ones I just tested – `lftp`, `curl`, `nc-openbsd`, `mutt`, `ncat`,
>>> `wget`, inetutils `telnet` and `ssh` – keep trying addresses until one
>>> succeeds, or until the program runs out of addresses to try.
>>
>> Hello,
>>  Could you try this patch and verify that it works in your use case?
>> http://git.savannah.gnu.org/gitweb/?p=gnutls.git;a=commitdiff;h=5da62fb671474b1ffaad5dd3062895ca109c47bf
>>
>> regards,
>> Nikos
> 
> The patch doesn't change anything; `gnutls-cli` still exit(1)'s after
> the first error and does not attempt other addresses. (Also,
> `gnutls-cli` now only shows the hostname, while previously it used to
> show the actual IP address it tried.)


Well, since you have the test environment could you provide a patch for
that to work as expected? Is replacing the exit() with continue, enough?

regards,
Nikos



reply via email to

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