Tatsuhiro Tsujikawa <tatsuhiro.t_at_gmail.com> wrote:

> I ran tests several times, but I didn't see much difference between
> patched and unpatched sources. Both completed tests around 6 minutes.

In my case a full run of the test suite, not counting compilation
time, takes nearly 45 minutes. Yes I know, I need more processing
power ;-) So execution time is something I usually notice. I will do
some investigation into this next week when time permits, to verify if
there's actually a time increase or not.

>> For the benefit of scripts parsing curl's stderr It would be
>> interesting if all metalink related messages sent to stderr started
>> with a common literal such as 'Metalink:'.
>>
>
> Agreed. Patch attached.

And pushed a couple of minutes ago.

I've added six more Metalink tests with 'funny' command line options.
And these reveal the following:

Test case 2010 pushed disabled given that it seems to trigger
unexpected behavior. Option '--include' (-i) used along with
--metalink should most probably not include the HTTP headers in the
downloaded file.

And there's something else to ponder. Should option '--dump-header'
(-D) work as it is doing right now along with '--metalink' option, or
should we make sure that no HTTP headers are dumped nor file created
when both options are used?

If --metalink along with --dump-header is 'fixed', metalink test case
definitions will also need fixing.

-- 
-=[Yang]=-
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-users
FAQ:        http://curl.haxx.se/docs/faq.html
Etiquette:  http://curl.haxx.se/mail/etiquette.html