On 8/21/2014 10:34 AM, dev wrote:
> I have been running nightly autobuilds for a few months and I do
> glance at the results from time to time. Which is to say I have
> neglected to follow up :-) and really fix or try to fix or single step
> or do something dammit on the only test failure I see on Solaris 10 :
>
>
> test 1513...[return failure immediately from progress callback]
>
> lib1513 returned 7, when expecting 42
> goto problem 2
> exit FAILED
> == Contents of files in the log/ dir after test 1513
> === Start of file stderr1513
> URL: http://localhost/1513
> === End of file stderr1513

Someone reported that last month as well [1]. I question the validity of
the test. CURLOPT_TIMEOUT has a documented return value of CURLE_OK [2].
If the timeout can occur at any point, how can any return code be
guaranteed; or why should any return code be guaranteed? (Real question
not rhetorical). The case of 1513 is to test the progress callback from
what I can see here but if there's no time for the callback to even be
called then personally I wouldn't expect CURLE_ABORTED_BY_CALLBACK (42).
If there is anything I can do to help with this let me know.

[1]: http://curl.haxx.se/mail/lib-2014-07/0121.html
[2]: http://curl.haxx.se/libcurl/c/CURLOPT_TIMEOUT.html

-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2014-08-21