RFR 4906983: java.net.URL constructors throw MalformedURLException in undocumented way

[Chris Hegarty]

Sebastian,

> On 10 Sep 2015, at 20:32, Sebastian Sickelmann <sebastian.sickelmann at gmx.de> wrote:
> 
> Hi,
> 
> first thanks to Chris and David for their helpful input . I looked through the existing 
> Testcases and found one that is already testing for negative-port numbers.
> So i extended the @bug line with "4906983" which I hope is the right solution to do it.

It is.

> I am with Chris, when he says normally you only have numbers between 1 and
> 65535 (because many protocols are using tcp). So i changed to documentation as 
> Chris suggested it.

Thanks.

> But ports above this "natural" barrier are valid too. It depends on the protocol what
> to do with the port information. So I also extended the testcase to check that their are
> valid port numbers also above 65535 and the special -1.
> 
> But i asked myself should
> 	new URL("http://server:-1/path");
> be realy a valid URL?

Probably not, but I think it is possibly too late to change this now.

> What do you think?
> Special thanks to David who hosted the new webrev:
> http://cr.openjdk.java.net/~dbuck/4906983.1/ <http://cr.openjdk.java.net/~dbuck/4906983.1/>

I think these changes are good as is ( pending the outcome of the mail exchange with Mark Sheppard ).

-Chris.

> -- Sebastian
> 
> 
> Am 10.09.2015 um 12:38 schrieb Chris Hegarty:
>> Another minor comment...
>> 
>> While what you have suggested is not incorrect, I’m afraid it is giving the wrong impression about the typical acceptable port ranges. A port of Integer.MAX_VALUE is not all that useful, since it typically maps to a TCP port number ( but not always ). Maybe just remove the valid values from @param port, and add something like the following to MalformedURLException: “.., or the port is a negative number other than -1” ?
>> 
>> -Chris.
>> 
>> On 10 Sep 2015, at 11:13, Chris Hegarty <chris.hegarty at oracle.com> <mailto:chris.hegarty at oracle.com> wrote:
>> 
>>> On 8 Sep 2015, at 21:01, Sebastian Sickelmann <sebastian.sickelmann at gmx.de> <mailto:sebastian.sickelmann at gmx.de> wrote:
>>> 
>>>> Hi,
>>>> 
>>>> Please find my small patch[1] to javadoc in java.net.URL that adresses
>>>> JDK-4906983(javadoc-fix)[2].
>>>> 
>>>> I signed the SCA/OCA some time ago. Feel free to check at the OCA
>>>> Signatures-List[3]
>>>> 
>>>> thanks to david buck for hosting this patch on cr.openjdk.java.net.
>>>> 
>>>> -- Sebastian Sickelmann
>>>> 
>>>> [1] http://cr.openjdk.java.net/~dbuck/4906983.0/ <http://cr.openjdk.java.net/~dbuck/4906983.0/>
>>> Just to confirm this is a spec only change, that documents long standing existing behaviour, right?
>>> 
>>> I think we should add a minimal testcase to cover this.
>>> 
>>> Thanks,
>>> -Chris.
>>> 
>>>> [2] https://bugs.openjdk.java.net/browse/JDK-4906983 <https://bugs.openjdk.java.net/browse/JDK-4906983>
>>>> 
>>>> [3] http://www.oracle.com/technetwork/community/oca-486395.html <http://www.oracle.com/technetwork/community/oca-486395.html>
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/net-dev/attachments/20150913/bd06dbca/attachment.html>