[Mono-list] Re: XmlNamespaceManager.LookupPrefix
Hello, I suggest we move ahead with the fix. OK, just committed. I know that time is money :) I wish DotGNU also decide to move. Regards, -- Atsushi Eno ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
[Mono-list] Re: XmlNamespaceManager.LookupPrefix
Hello, For now, lets assume that the bug is in the documentation, so lets move closer to the .NET behavior, and submit the bug to ECMA/Microsoft so they can fix their docs. About this topic, I made contact with DotGNU developers ML. Gopal V, one of the developer, said that it would be better to wait until the next revision of ECMA is released. I think it is better choice. Miguel, how do you think? Related to this, I found another ECMA or MS bug in XmlTextWriter.LookupPrefix. If prefix was not found, then - Mono == null - MS (both document and implementation) == null - ECMA spec. == String.Empty. - pnetlib == String.Empty - OCL == null Moreover, XmlNode.GetPrefixOfNamespace(uri) returns String.Empty. (MS document has no hints) ... I'm totally confused ;-( Now I have completed (I hope so) those modification. I attached. It was a bit complicated and might have another bugs. (At least, no failures are found with my environment.) I also checked with Sebastien's System.Security testcases. I assume many failure are cleared. Thanks, -- Atsushi Eno XML.diff Description: Binary data
Re: [Mono-list] Re: XmlNamespaceManager.LookupPrefix
Hello, For now, lets assume that the bug is in the documentation, so lets move closer to the .NET behavior, and submit the bug to ECMA/Microsoft so they can fix their docs. About this topic, I made contact with DotGNU developers ML. Gopal V, one of the developer, said that it would be better to wait until the next revision of ECMA is released. I think it is better choice. Miguel, how do you think? The ECMA standard is a useful starting point for many things in Mono, but it is not an ultimate reference. After all we are implementing the .NET superset instead of the ECMA spec. On the other hand, we know for a fact that the documentation might be out of sync with the implementation. This happens very often with large projects. This is a bug in the specification, so lets just move ahead an do the right thing. There is no point in waiting for the updated documentation, we already know its wrong. I suggest we move ahead with the fix. Miguel. ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
[Mono-list] Re: XmlNamespaceManager.LookupPrefix
Hello. Nick, James, thank you. System.Xml is part of the ECMA standard, yes? In this case shouldn't a documentation discrepancy indicate Microsoft needing to fix things, as otherwise they'd be breaking the standard? :) I looked into ECMA spec and pnetlib (of DotGNU) source, and found that both of them are based on MS's documentation. I instinctually want to agree with James. But I also think different. We should carefully consider which is totally good. I think, for this time, it is ECMA specification that is worthless. Intel's OCL 0.5 source points out the same problem. It comments out the code which returns String.Empty, and is written to return null. The only thing I worry is about compatibility of DotGNU. As far as I saw, pnetlib's XmlTextWriter.LookupPrefix method is inconsistent with MS documentation. I assume their attitude is not clear with respect to ECMA specifications. I'll talk to them about this problem. Thanks, -- Atsushi Eno ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Re: XmlNamespaceManager.LookupPrefix
Hello, I looked into ECMA spec and pnetlib (of DotGNU) source, and found that both of them are based on MS's documentation. Remember one thing: ECMA documentation is based on the Microsoft documentation, and sometimes the Microsoft documentation does contain errors on it or does not match the actual implementation. Sometimes the bug is on the documentation, and the right thing to do is to fix the documentation instead of the code. For now, lets assume that the bug is in the documentation, so lets move closer to the .NET behavior, and submit the bug to ECMA/Microsoft so they can fix their docs. We have a file where we log these things in Mono mcs/class/doc/API-notes Best wishes, Miguel ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list