Re: [9] Review request : JDK-6933879: URISyntaxException when non-alphanumeric characters are present in scope_id
On 15/12/14 12:01, Alan Bateman wrote: On 15/12/2014 11:25, Chris Hegarty wrote: Konstantin, I did reply to this RFR [1], with a question, that is still unanswered. -Chris. [1] http://mail.openjdk.java.net/pipermail/net-dev/2014-December/008782.html I can think of configurations where the scenario could arise. It make me wonder if this will require a spec clarification as scope_id is defined in Inet6Address to allow any String and so raises the question as to the characters permitted in the host component. I think that we can add '_' and '.' . ':' I am less sure about. I don't see any reason to update the spec here, given that the set of allowable character is not clearly defined in the relevant RFC's. The change, as proposed in the webrev, is reasonable ( with ':' removed ), but you should also include an update to jdk/test/java/net/URI/Test.java, to cover this. -Chris.
Re: [9] Review request : JDK-6933879: URISyntaxException when non-alphanumeric characters are present in scope_id
On 16/01/2015 10:49, Chris Hegarty wrote: : I don't see any reason to update the spec here, given that the set of allowable character is not clearly defined in the relevant RFC's. I think we need to create a bug to look into this more. Inet6Address allows the scope of be any String but URI puts restrictions on what is legal. The proposed change updates the list of allowed characters but I assume it will need to re-visited again as systems with more exotic interface names are encountered. -Alan
Re: [9] Review request : JDK-6933879: URISyntaxException when non-alphanumeric characters are present in scope_id
On 16/01/15 11:29, Alan Bateman wrote: On 16/01/2015 10:49, Chris Hegarty wrote: : I don't see any reason to update the spec here, given that the set of allowable character is not clearly defined in the relevant RFC's. I think we need to create a bug to look into this more. Inet6Address allows the scope of be any String but URI puts restrictions on what is legal. The proposed change updates the list of allowed characters but I assume it will need to re-visited again as systems with more exotic interface names are encountered. Yes that is possible, but to date this is the first time I've seen this come up. Given the lack of clear specification in the relevant RFC's (relating to scope ids ), I don't think we should be making too many specification changes in this area. But I agree with your comment, URI has a restriction that it not clearly specified in the Java spec. -Chris.
Re: [9] Review request : JDK-6933879: URISyntaxException when non-alphanumeric characters are present in scope_id
Hi Chris, Alan, thank you for reviewing. I have made a new webrev http://cr.openjdk.java.net/~kshefov/6933879/webrev.01 I have removed ":" and added a test case. -Konstantin 16.01.2015 14:42, Chris Hegarty пишет: On 16/01/15 11:29, Alan Bateman wrote: On 16/01/2015 10:49, Chris Hegarty wrote: : I don't see any reason to update the spec here, given that the set of allowable character is not clearly defined in the relevant RFC's. I think we need to create a bug to look into this more. Inet6Address allows the scope of be any String but URI puts restrictions on what is legal. The proposed change updates the list of allowed characters but I assume it will need to re-visited again as systems with more exotic interface names are encountered. Yes that is possible, but to date this is the first time I've seen this come up. Given the lack of clear specification in the relevant RFC's (relating to scope ids ), I don't think we should be making too many specification changes in this area. But I agree with your comment, URI has a restriction that it not clearly specified in the Java spec. -Chris.
Re: [9] Review request : JDK-6933879: URISyntaxException when non-alphanumeric characters are present in scope_id
On 16/01/15 12:50, Konstantin Shefov wrote: Hi Chris, Alan, thank you for reviewing. I have made a new webrev http://cr.openjdk.java.net/~kshefov/6933879/webrev.01 I have removed ":" and added a test case. This looks ok to me. -Chris. -Konstantin 16.01.2015 14:42, Chris Hegarty пишет: On 16/01/15 11:29, Alan Bateman wrote: On 16/01/2015 10:49, Chris Hegarty wrote: : I don't see any reason to update the spec here, given that the set of allowable character is not clearly defined in the relevant RFC's. I think we need to create a bug to look into this more. Inet6Address allows the scope of be any String but URI puts restrictions on what is legal. The proposed change updates the list of allowed characters but I assume it will need to re-visited again as systems with more exotic interface names are encountered. Yes that is possible, but to date this is the first time I've seen this come up. Given the lack of clear specification in the relevant RFC's (relating to scope ids ), I don't think we should be making too many specification changes in this area. But I agree with your comment, URI has a restriction that it not clearly specified in the Java spec. -Chris.
