It ended up a bit more complicated; that patch leaks memory when a socket isn't closed manually. There's also an issue with how the close flag on _fileobject works -- rather than simply decrementing the refcount the way that socket.close does, it should close the socket regardless of the refcount.
On Tue, Jul 15, 2008 at 9:24 AM, Michael Foord <[EMAIL PROTECTED]> wrote: > Curt Hagenlocher wrote: > >> Those changes will be in 2.0b4. >> >> > Is this your patch from March, added to the following work item? > > http://www.codeplex.com/IronPython/WorkItem/View.aspx?WorkItemId=10825 > > Michael > >> Some of CPython's socket tests are problematic in the face of >> non-deterministic finalization. They assume, for instance, that a socket >> will implicitly be closed when the variable goes out of scope. Because of >> this, and because users haven't had a chance to bang on them yet, I wasn't >> confident enough in the changes I made to advocate porting them back to >> 1.1.2. <http://1.1.2 <http://1.1.0.2/>.> >> >> On Tue, Jul 15, 2008 at 8:38 AM, Michael Foord <[EMAIL PROTECTED]<mailto: >> [EMAIL PROTECTED]>> wrote: >> >> Hello guys, >> >> The only socket related issue I see in that list is: >> >> 16356 1.1.2: socket.getnameinfo(...) broken under Vista >> >> We were hoping that 1.1.2 would finally put to rest IronPython 1 >> issues with urllib / urllib2. Has that been done? >> >> All the best, >> >> Michael >> >> Srivatsn Narayanan wrote: >> >> >> Hello IronPython Community, >> >> I'm pleased to announce the release of IronPython 1.1.2 RC1. >> In this release twenty-one CodePlex work Items that were >> identified as high priority on the IronPython mailing list >> were fixed. It's worth noting that this included the >> implementation of the "_winreg" CPython module and a few >> performance improvements in targeted areas. Please also note >> that there are two potentially breaking changes in this release: >> >> · nt.unlink will now throw an exception if the file doesn't >> exist, as it does in CPython. >> >> · The signature of >> IronPython.Runtime.Operations.Ops.Id<http://ironpython.runtime.operations.ops.id/> >> <http://ironpython.runtime.operations.ops.id/>() has changed >> >> and it now returns an object instead of long. This maps to the >> id() function in python. Note that there is no change for >> python caused by this and only code which is directly calling >> Ops.Id from a statically typed language like C# or VB will be >> affected. >> >> Assuming no major issues are discovered in this build, we hope >> to rerelease it as our last planned IronPython 1.1 release, >> IronPython 1.1.2, in several weeks. Please let us know if you >> encounter any issues as soon as possible. >> >> The following CodePlex Work Items were closed: >> >> · 16368 1.1.2: nt.access is missing >> >> · 16402 1.1.2: implement _winreg module >> >> · 15105 1.1.2: endpos is zero in IronPython 1.1.1 >> >> · 16335 1.1.2: Event handlers can cause circular references >> and leak memory >> >> · 16337 1.1.2: Trivial: Implement float.__lt__(float) >> >> · 16338 1.1.2: Using lambda in class definition will add >> <lambda$..> into the >> >> · 16342 1.1.2: calling base class __call__ invokes constructor >> instead >> >> · 16343 1.1.2: problem with __slots__ and __init__ in >> new-style classes >> >> · 16347 1.1.2: Trivial: popen shouldn't open new window >> >> · 16348 1.1.2: Removes the inexistent file did not throw >> OSError in IP >> >> · 16350 1.1.2: int() doesn't convert representable longs to int >> >> · 16351 1.1.2: dict.update doesn't take keyword arguments - >> differs from CPython >> >> · 16353 1.1.2: Trivial: int('0x20', 16) fails to parse, long too >> >> · 16355 1.1.2: unpacking single element tuples in >> for-statement, listcomp and generator >> >> · 16356 1.1.2: socket.getnameinfo(...) broken under Vista >> >> · 16360 1.1.2: Class with slots and getattr not compatible >> >> · 16363 1.1.2: Can't call method w/ nullable as 1st argument >> w/ greater than 5 arguments >> >> · 16364 1.1.2: Backport fix for compiled regular expressions >> >> · 16365 1.1.2: Tuple hashing improvements >> >> · 16366 1.1.2: PyCF_DONT_IMPLY_DEDENT support in compile >> >> · 16749 1.1.2 (Trivial): Modifier of >> PythonEngine.DefaultCompilerContext(..) >> >> We'd like to thank everyone in the community who contributed >> to these bugs: Ronnie Maor, jackeyoo, sanxiyn, Michael Foord, >> Kamil Dworakowski, lthompson, romank, Jeff Brown, rridge, >> David Fraser, pobrien, ebaklund. >> >> You can download IronPython 1.1.2 RC1 at: >> >> http://www.codeplex.com/IronPython/Release/ProjectReleases.aspx?ReleaseId=15198 >> >> >> The IronPython Team >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Users mailing list >> [email protected] <mailto:[email protected]> >> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com >> >> >> >> -- http://www.ironpythoninaction.com/ >> http://www.voidspace.org.uk/ >> http://www.trypython.org/ >> http://www.ironpython.info/ >> http://www.theotherdelia.co.uk/ >> http://www.resolverhacks.net/ >> >> _______________________________________________ >> Users mailing list >> [email protected] <mailto:[email protected]> >> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Users mailing list >> [email protected] >> http://lists.ironpython.com/listinfo.cgi/users-ironpython.com >> >> > > > -- > http://www.ironpythoninaction.com/ > http://www.voidspace.org.uk/ > http://www.trypython.org/ > http://www.ironpython.info/ > http://www.theotherdelia.co.uk/ > http://www.resolverhacks.net/ > > _______________________________________________ > Users mailing list > [email protected] > http://lists.ironpython.com/listinfo.cgi/users-ironpython.com >
_______________________________________________ Users mailing list [email protected] http://lists.ironpython.com/listinfo.cgi/users-ironpython.com
