[Twisted-Python] Re: [Twisted-commits] r13090 - Use ThreadedResolver instead of BlockingResolver by default - the thread import issue is moot, PyOpenSSL imports thread unconditionally.
On Mon, 2005-03-07 at 22:31 -0700, Jp Calderone wrote:
Use ThreadedResolver instead of BlockingResolver by default - the thread import issue is moot, PyOpenSSL imports thread unconditionally.
Not true. Some BSDs don't do threading in Python, and we need to support them. Remember, the pyOpenSSL support is optional. The fact pyOpenSSL does the import unconditionally is a bug.
On Tue, 08 Mar 2005 10:45:21 -0500, Itamar Shtull-Trauring itamar@itamarst.org wrote:
On Mon, 2005-03-07 at 22:31 -0700, Jp Calderone wrote:
Use ThreadedResolver instead of BlockingResolver by default - the thread import issue is moot, PyOpenSSL imports thread unconditionally.
Not true. Some BSDs don't do threading in Python, and we need to support them. Remember, the pyOpenSSL support is optional. The fact pyOpenSSL does the import unconditionally is a bug.
I don't have access to a platform without threading support. There isn't even a buildslave running on one. If someone sets that up, I'll fix failures it reports.
Jp
On Tue, Mar 08, 2005 at 07:18:35PM +0000, Jp Calderone wrote:
On Tue, 08 Mar 2005 10:45:21 -0500, Itamar Shtull-Trauring itamar@itamarst.org wrote:
Not true. Some BSDs don't do threading in Python, and we need to support them. Remember, the pyOpenSSL support is optional. The fact pyOpenSSL does the import unconditionally is a bug.
I don't have access to a platform without threading support. There isn't even a buildslave running on one. If someone sets that up, I'll fix failures it reports.
NetBSD earlier than (and including) 1.6.2 lacked threading support (except for the pth userspace cooperative threading)
Unfortunately all of my NetBSD boxes have are now running 2.0/2.0.1/-current, so I don't have one you can use to build.
-w
On Tue, 2005-03-08 at 19:18 +0000, Jp Calderone wrote:
I don't have access to a platform without threading support. There isn't even a buildslave running on one.
That doesn't mean you ought to *break* it. Can't you put the installation of ThreadedResolver in as a try/except over import thread? Or can we have twisted.python.runtime.threadingSupported()? or something. But we oughtn't break support for non-threaded platforms when it's easy enough to test for it (sys.module["thread"] = None and then run code as usual).
On Tue, 08 Mar 2005 15:06:49 -0500, Itamar Shtull-Trauring itamar@itamarst.org wrote:
On Tue, 2005-03-08 at 19:18 +0000, Jp Calderone wrote:
I don't have access to a platform without threading support. There isn't even a buildslave running on one.
That doesn't mean you ought to *break* it. Can't you put the installation of ThreadedResolver in as a try/except over import thread? Or can we have twisted.python.runtime.threadingSupported()? or something. But we oughtn't break support for non-threaded platforms when it's easy enough to test for it (sys.module["thread"] = None and then run code as usual).
Yes. There are lots of ways it could work. How can I implement any of them, though, if I can't run a unit test and have it pass?
I could probably figure out a way to get it working without a buildslave, but this goes to a larger issue. I'm not going to waste my time trying to support a feature that we think we would like to support. Either we want to support it, in which case we need to test it, or we don't.
Anyway, I've had an idea while reading the posts in this thread. Python can be compiled without thread support intentionally.
Anyone want to volunteer a reliable host on which to run such a slave?
Jp
Twisted-Python mailing list Twisted-Python@twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python
Jp Calderone wrote:
Anyone want to volunteer a reliable host on which to run such a slave?
I actually have a few hosts hanging around here that I could put online (in particular, I have a mac that isn't doing anything, and the OS-X buildslave is offline) but figuring out how to set up a buildbot got me stuck last time.
Is there any updated documentation of this procedure?
participants (4)
-
Glyph Lefkowitz -
Itamar Shtull-Trauring -
Jp Calderone -
William Waites