Re: Strange error resolving non-existent hostname

[Andreas Gustafsson <gson%gson.org@localhost>]

Thomas Klausner wrote:
> Recently, a mercurial test has started failing for me (7.99.21/amd64):
> 
> @@ -80,7 +80,7 @@
>    $ echo 'http://does.not.exist/bundle.hg' > server/.hg/clonebundles.manifest
>    $ hg clone http://localhost:$HGPORT 404-url
>    applying clone bundle from http://does.not.exist/bundle.hg
> -  error fetching bundle: * not known (glob)
> +  error fetching bundle: No address associated with hostname
>    abort: error applying bundle
>    (if this error persists, consider contacting the server operator or disable clone bundles via "--config experimental.clonebundles=false")
>    [255]
> 
> ERROR: test-clonebundles.t output changed
> 
> 
> In other words, the test tries resolving does.not.exist and expects
> any error message ending with 'not known' but gets the one for
> EAI_NODATA, in contrast to all other operating systems (according to
> the mercurial author).
> 
> Can someone else please try this, to make sure it's not my DNS setup?
> 
> cd /usr/pkgsrc/devel/py-mercurial && make test

I see the same error on 6.1.4.

Here's a simpler test:

   $ telnet nosuch
   nosuch: No address associated with hostname

I get the above (incorrect) output on both 6.1.4 and -current.

> Any ideas what could be the reason for this?

Looks like a bug in NetBSD's getaddrinfo() to me.
-- 
Andreas Gustafsson, gson%gson.org@localhost