Re: Closing out 2.7 beta 4
Jim Baker <[email protected]>
| Newsgroups | gmane.comp.lang.jython.devel |
|---|---|
| Message-ID | <CAOhO=aPiNM+POtWKmgkmM7HS3B4hzvAZXZeuOT8AOmJPG=zFDQ@mail.gmail.com> |
On Mon, Oct 13, 2014 at 4:06 PM, Jeff Allen <[email protected]> wrote: > Jim: > > Thanks for all the work that went into this succinct position report. I've > several other things going on at the moment that leave me not much time for > Jython, during your beta 4 window. (Better from mid Nov.) But I'll pick off > something from your high priority set and see where it goes. > Your fantastic work is something we all appreciate! Glad to see one such bug already get squashed! > ... > > Failing tests > > In general, the most important problems to be fixed are those that > directly tested by the regrtest. Here’s is the current list of failures, > from regrtest itself: > > [exec] 16 fails unexpected: > [exec] test_asynchat test_asyncore test_dict_jy test_email > [exec] test_email_renamed test_gzip test_import test_inspect > [exec] test_isinstance test_json test_socket test_ssl test_strtod > [exec] test_tarfile test_univnewlines test_xmlrpc > > This is useful on its own! I often wonder whether what I see, mostly on > Windows, is typical. I'm still getting a fair amount of "cannot delete" > failure from unclosed files on Windows. > Here's what I think is happening: those os.unlink (= os.remove) errors are probably happening on test code that believes it is running on Posix (because it's looking at os.name, sees "java" and assumes *not Windows*). But there's a different behavior between Posix and Windows, per os.remove's docs (https://docs.python.org/2/library/os.html#os.remove): *On Windows, attempting to remove a file that is in use causes an exception to be raised; on Unix, the directory entry is removed but the storage allocated to the file is not made available until the original file is no longer in use.* This would not be the first time we have seen this in the stdlib, including the test suite. > > For a while someone was hosting a build-bot for us, but it became sporadic > and has gone now. Such a thing would be useful but more so if we decided > not to have failing tests but count skips removed as our measure of > progress. (Skips ought not to be counted for CPython and GC specifics.) Is > a build-bot possible? > I asked Darjus Loktevic privately about this, since he worked on the issue of CI support at the EuroPython sprints. Darjus has done some initial work on Travis CI integration with our github mirror, and it's possible we could also have this done for any PRs against that mirror, which would be a huge gain in reviewer efficiency. But I will let him respond in more detail on this thread! > ... > > We plan to kickoff Jython 3.5 development at the PyCon 2015 sprints in > Montreal; as with previous versioning, > > ... > > Jython 3.5 should target a minimum of Java 8; and we should upgrade to > using Antlr 4 for our parser support. > > Ambitious! But interesting. It will be a stretch to be developing on two > fronts: you don't have to read Python-dev for very long to be reminded of > that. > Agreed. But getting to language support, vs standard library, should be something that ca be done very quickly. Frank Wierzbicki has a prototype Antlr 3 grammar in https://bitbucket.org/fwierzbicki/jython-grammar, which should allow us to quickly add support like function annotations. Presumably this is what we should do first, ideally before the PyCon 2015 sprint. Another thing to know is that Python 3's stdlib is supposed to use Python with optional C accelerators (this is combined during module import). This helps PyPy, so it should help us too. Interestingly, Antlr 4 ships with a sample Python 3 grammar: https://github.com/antlr/grammars-v4/blob/master/python3/Python3.g4 Now to get that working with all of the support we have done with Python AST, along with code compilation effort and incremental parse support for the console, will not be minor. But at least we have something we can look at when we move to Antlr 4. There are other interesting aspects, such as Argument Clinic, and how that would relate to our internal expose infrastructure. I'm sure we will find many details, hopefully none too frustrating! - Jim ------------------------------------------------------------------------------ Comprehensive Server Monitoring with Site24x7. Monitor 10 servers for $9/Month. Get alerted through email, SMS, voice calls or mobile push notifications. Take corrective actions from your mobile device. http://p.sf.net/sfu/Zoho _______________________________________________ Jython-dev mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jython-dev