I just posted the attached message to the libwin32 list, but I realize now that this is probably a core P5P issue, given that Win32.pm ships in the core.
I can't make up my mind if the proposed changes below are acceptable "bug fixes", or if they should be rejected because of backwards compatibility concerns. Windows 2008 R2 is still in beta, so changing the return value for it should be acceptable. Windows Home Server was released in 2007, but had serious file system problems that weren't fixed until July 2008. I doubt it is used much beyond home use, so I feel like changing the value will have very low impact too. At least one big Perl application (the Logitech SqueezeCenter) does need to know if it is running on WHS or not. Windows 2003 R2 is the hardest to justify, as it has been released for the longest time. But it would be weird to attach the "R2" to the 2008 name but not to 2003. Of course you can argue that GetOSName() is just a convenience function to return a display name, and people can always use GetOSVersion() to check for themselves, but I doubt people always do that, as GetOSName() hides all the ugly heurists from you. So, do you think any of these 3 return value changes are acceptable for the core? Cheers, -Jan On Thu, 18 Jun 2009, Jan Dubois wrote: > > Right now "Windows Home Server" it is being reported as "Win2003". > I'm tempted to say it should return "WinHomeServer" instead. > > What about "Windows 2003 R2" and "Windows 2008 R2"? Should > we append the "R2" to the OSName as well? The R2 versions > are quite a bit different from the original releases, so > there is some justification to report them differently > (just as we report 2008 different from Vista even though > internally they use the same Windows version numbers). > > Best places I could find to describe the differences in the > 2003 and 2008 R2 releases: > > http://technet.microsoft.com/en-ca/windowsserver/bb428898.aspx > http://www.microsoft.com/windowsserver2008/en/us/R2.aspx > > Thoughts?
