On Monday, November 12, 2012 23:48:02 Daniel Stenberg wrote:
> On Fri, 9 Nov 2012, Kamil Dudka wrote:
> > On Thursday, November 08, 2012 15:41:19 Anthony Bryan wrote:
> >> we've had a few testers & this seems to work.
> >
> > Thanks for review! I have pushed them.
>
> I believe this patch series caused test 2005 to start failing when built
> with OpenSSL... When I revert the three commits cf75a64651e dca8ae5f02a84
> 49c37e6c1c11, it starts working again.
>
> $ ./runtests.pl -n 2005
>
> ...
>
> test 2005...[Metalink local XML file, HTTP resource]
>
> output (log/stderr2005) FAILED:
> --- log/check-expected 2012-11-12 23:47:12.000000000 +0100
> +++ log/check-generated 2012-11-12 23:47:12.000000000 +0100
> @@ -3,4 +3,4 @@
> Metalink: fetching (log/download2005) from (http://127.0.0.1:8990/2005)...
> Metalink: fetching (log/download2005) from (http://127.0.0.1:8990/2005) OK
> Metalink: validating (log/download2005)...
> -Metalink: validating (log/download2005) [sha-256] OK
> +Metalink: validating (log/download2005) [sha-256] FAILED (failed to
> initialize hash algorithm)
>
> - abort tests
> TESTDONE: 0 tests out of 1 reported OK: 0%
> TESTFAIL: These test cases failed: 2005
> TESTDONE: 1 tests were considered during 2 seconds.

Thanks for heads up! I did properly understand the logic behind the macros.
Now I realized that the OpenSSL functions are used directly without any
wrappers and they return 1 for success (in opposite to 0 as expected by all
the other wrappers). I have a patch for this and will submit it later today
when I have some additional testing done.

Kamil
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2012-11-13