On track for Python 2.6.4 final this Sunday?
Are we on track to release 2.6.4 final this Sunday or do we need
another rc?
Yesterday, Tarek committed another setuptools related fix and said
that he was going to run a bunch of build tests locally. Tarek, how
did that go?
Please note that issue 7064 is still open as a release blocker. Can
we close that (again) now?
http://bugs.python.org/issue7064
I am still being very conservative about what goes in 2.6.4. Only
regressions introduced in 2.6.3 are being accepted. There's plenty of
time after that for patches to go in for 2.6.5.
-Barry
Barry Warsaw wrote:
Are we on track to release 2.6.4 final this Sunday or do we need another rc?
Yesterday, Tarek committed another setuptools related fix and said that he was going to run a bunch of build tests locally. Tarek, how did that go?
Please note that issue 7064 is still open as a release blocker. Can we close that (again) now?
http://bugs.python.org/issue7064
I am still being very conservative about what goes in 2.6.4. Only regressions introduced in 2.6.3 are being accepted. There's plenty of time after that for patches to go in for 2.6.5.
It would be nice to get this issue resolved out for 2.6.4:
http://bugs.python.org/issue4120
The problem is that extensions built with 2.6.x will not work when used with a User-only installation of Python on machines that don't already have the MS VC90 CRT DLLs installed system-wide.
There's a patch available that appears to work, but someone with more manifest-Windows-DLL-fu than me will need to have a look at it.
One possible approach would be to use the patch in 2.6.4 and then continue digging deeper until 2.6.5 is released.
-- Marc-Andre Lemburg eGenix.com
Professional Python Services directly from the Source (#1, Oct 13 2009)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
::: Try our new mxODBC.Connect Python Database Interface for free ! ::::
eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/
It would be nice to get this issue resolved out for 2.6.4:
http://bugs.python.org/issue4120
The problem is that extensions built with 2.6.x will not work when used with a User-only installation of Python on machines that don't already have the MS VC90 CRT DLLs installed system-wide.
As this bug already exists in 2.6.2, I don't think the change is eligible for 2.6.4.
In addition, I want to review it, which I won't be able to until Sunday.
One possible approach would be to use the patch in 2.6.4 and then continue digging deeper until 2.6.5 is released.
This also speaks against the patch. Anything being changed in this area ideally should be the final state of affairs for the rest of 2.6.x.
Regards, Martin
"Martin v. Löwis" wrote:
It would be nice to get this issue resolved out for 2.6.4:
http://bugs.python.org/issue4120
The problem is that extensions built with 2.6.x will not work when used with a User-only installation of Python on machines that don't already have the MS VC90 CRT DLLs installed system-wide.
As this bug already exists in 2.6.2, I don't think the change is eligible for 2.6.4.
In addition, I want to review it, which I won't be able to until Sunday.
Then I'd suggest to wait another week with 2.6.4 to give you a chance to look at the patch.
One possible approach would be to use the patch in 2.6.4 and then continue digging deeper until 2.6.5 is released.
This also speaks against the patch. Anything being changed in this area ideally should be the final state of affairs for the rest of 2.6.x.
Of course, but we don't live in an ideal world, otherwise we would have noticed before releasing 2.6 :-)
I guess the problem is that developer typically already have the MS VC90 CRT DLLs installed system-wide, so they simply don't notice.
-- Marc-Andre Lemburg eGenix.com
Professional Python Services directly from the Source (#1, Oct 13 2009)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
::: Try our new mxODBC.Connect Python Database Interface for free ! ::::
eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/
As this bug already exists in 2.6.2, I don't think the change is eligible for 2.6.4.
In addition, I want to review it, which I won't be able to until Sunday.
Then I'd suggest to wait another week with 2.6.4 to give you a chance to look at the patch.
That won't make the change more eligible.
Of course, but we don't live in an ideal world, otherwise we would have noticed before releasing 2.6 :-)
Oh, I did notice. I had been cautioning for years that switching to a newer VS version (2005 or later) would cause severe problems. We are still trying to recover from the switch to VS 2008. That said, staying with VS 2003 really hadn't been an option, either. It's just said that Microsoft has created a new DLL hell in their attempt to fix the old one, and that they fail to admit it (assuming that everybody should be using .NET, anyway).
Regards, Martin
"Martin v. Löwis" wrote:
As this bug already exists in 2.6.2, I don't think the change is eligible for 2.6.4.
In addition, I want to review it, which I won't be able to until Sunday.
Then I'd suggest to wait another week with 2.6.4 to give you a chance to look at the patch.
That won't make the change more eligible.
No, but chances increase :-)
Of course, but we don't live in an ideal world, otherwise we would have noticed before releasing 2.6 :-)
Oh, I did notice. I had been cautioning for years that switching to a newer VS version (2005 or later) would cause severe problems. We are still trying to recover from the switch to VS 2008. That said, staying with VS 2003 really hadn't been an option, either. It's just said that Microsoft has created a new DLL hell in their attempt to fix the old one, and that they fail to admit it (assuming that everybody should be using .NET, anyway).
Well, hopefully we can work out those problems and then stay with VS 2008 a bit.
MS is already working on VS 2010:
http://www.microsoft.com/visualstudio/en-us/products/2010/default.mspx
List of C/C++ changes:
http://msdn.microsoft.com/en-us/library/dd465215(VS.100).aspx
Not sure what this means:
""" The libraries deployment model is no longer based on the fusion mechanism, and manifests are no longer used. Instead, the name of each dynamic link library contains its version number. """
Perhaps they finally realized that adding version number to DLLs is better than trying to match manifests.
-- Marc-Andre Lemburg eGenix.com
Professional Python Services directly from the Source (#1, Oct 13 2009)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
::: Try our new mxODBC.Connect Python Database Interface for free ! ::::
eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/
On Oct 13, 2009, at 11:01 AM, M.-A. Lemburg wrote:
Then I'd suggest to wait another week with 2.6.4 to give you a chance to look at the patch.
That's not a good option, IMO. We have a known broken 2.6.3 out there
and we owe it to our users to correct our mistake and given them the
release they should have gotten in the first place.
We don't need to wait too long for 2.6.5 though. A few months would
be appropriate.
-Barry
Barry Warsaw wrote:
On Oct 13, 2009, at 11:01 AM, M.-A. Lemburg wrote:
Then I'd suggest to wait another week with 2.6.4 to give you a chance to look at the patch.
That's not a good option, IMO. We have a known broken 2.6.3 out there and we owe it to our users to correct our mistake and given them the release they should have gotten in the first place.
We don't need to wait too long for 2.6.5 though. A few months would be appropriate.
Would it be reasonable to shorten that period, if the fix for the mentioned problem gets ready for prime time earlier ?
-- Marc-Andre Lemburg eGenix.com
Professional Python Services directly from the Source (#1, Oct 13 2009)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
::: Try our new mxODBC.Connect Python Database Interface for free ! ::::
eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/
On Oct 13, 2009, at 1:07 PM, M.-A. Lemburg wrote:
Would it be reasonable to shorten that period, if the fix for the mentioned problem gets ready for prime time earlier ?
I think there are many 2.6.x bugs queued up for after 2.6.4 is
released. I'm not at all opposed to setting a date approximately 1
month from now for the first 2.6.5 rc. That should give people plenty
of time to get their top fixes in.
i-like-timed-releases-ly y'rs, -Barry
Barry Warsaw wrote:
On Oct 13, 2009, at 1:07 PM, M.-A. Lemburg wrote:
Would it be reasonable to shorten that period, if the fix for the mentioned problem gets ready for prime time earlier ?
I think there are many 2.6.x bugs queued up for after 2.6.4 is released. I'm not at all opposed to setting a date approximately 1 month from now for the first 2.6.5 rc. That should give people plenty of time to get their top fixes in.
i-like-timed-releases-ly y'rs,
Agreed. Adding some pressure that way helps :-)
-- Marc-Andre Lemburg eGenix.com
Professional Python Services directly from the Source (#1, Oct 13 2009)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
::: Try our new mxODBC.Connect Python Database Interface for free ! ::::
eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611 http://www.egenix.com/company/contact/
>> We don't need to wait too long for 2.6.5 though. A few months would be
>> appropriate.
MAL> Would it be reasonable to shorten that period, if the fix for the
MAL> mentioned problem gets ready for prime time earlier ?I think it would be worthwhile to prioritize all outstanding bugs which have been mentioned in the context of 2.6.[345] and run a bug day with the top priority being to fix those bugs. If that task is completed, then move onto other stuff. Once those primary bugs are tackled schedule a 2.6.5 release.
Skip
At 08:27 AM 10/13/2009 -0400, Barry Warsaw wrote:
Are we on track to release 2.6.4 final this Sunday or do we need another rc?
Yesterday, Tarek committed another setuptools related fix and said that he was going to run a bunch of build tests locally. Tarek, how did that go?
FWIW, that change won't work to fix the problem on Windows if a package's setup.py is written with cross-platform paths (i.e., the usual case).
Also, for the record, it's not really a setuptools-related fix; it's just attempting to fix collateral damage caused by the attempt to fix the problems caused by the original API change... which affected all build_ext customizations relying on the docstring of the changed method, not just the one in setuptools.
On Tue, Oct 13, 2009 at 4:52 PM, P.J. Eby <pje@telecommunity.com> wrote:
At 08:27 AM 10/13/2009 -0400, Barry Warsaw wrote:
Are we on track to release 2.6.4 final this Sunday or do we need another rc?
Yesterday, Tarek committed another setuptools related fix and said that he was going to run a bunch of build tests locally. Tarek, how did that go?
I still need to do some more tests, I didn't have time to try the various projects under win32. It's planned to night.
The tests are consisting of compiling and insatling a dozain of projects on linux and win32 (and bith when possible)
FWIW, that change won't work to fix the problem on Windows if a package's setup.py is written with cross-platform paths (i.e., the usual case).
Also, for the record, it's not really a setuptools-related fix; it's just attempting to fix collateral damage caused by the attempt to fix the problems caused by the original API change... which affected all build_ext customizations relying on the docstring of the changed method, not just the one in setuptools.
Yes the doctest was pretty fuzzy about what is an extension name. It's will be improved in 2.7.
The current code is doing os.path.join()'s to join the Extension name with the build path, leading to the collateral damage you are mentioning. To fix the problem, the API has to be as permissive as it was before.
I have already three or four tests to fix that problem in test_build_ext. If you can take a look at them and think of any other test that would be missing, that would be appreciated. Another pair of eye is never enough.
Tarek
On Oct 13, 2009, at 11:16 AM, Tarek Ziadé wrote:
I still need to do some more tests, I didn't have time to try the various projects under win32. It's planned to night.
The tests are consisting of compiling and insatling a dozain of projects on linux and win32 (and bith when possible)
Great thanks, let us know how it goes. If there's any question at all
about the fixes that have gone in since rc1, we should spin an rc2.
-Barry
On Oct 13, 2009, at 12:57 PM, Barry Warsaw wrote:
On Oct 13, 2009, at 11:16 AM, Tarek Ziadé wrote:
I still need to do some more tests, I didn't have time to try the various projects under win32. It's planned to night.
The tests are consisting of compiling and insatling a dozain of projects on linux and win32 (and bith when possible)
Great thanks, let us know how it goes. If there's any question at
all about the fixes that have gone in since rc1, we should spin an
rc2.
Just curious, how big a server does the linux portion of the test
require memory-wise? Would, e.g. an Ubuntu 9.0x with 256MB do the job?
I ask because those types of servers are available for short-run jobs
for pennines per hour and I would be willing to work on the code to
fire-up, load them with the test suite, do the run, and shut them
down. I happen to have a bunch of code lying around for doing such
things...
S
On Tue, 13 Oct 2009 at 12:57, Barry Warsaw wrote:
On Oct 13, 2009, at 11:16 AM, Tarek Ziad� wrote:
I still need to do some more tests, I didn't have time to try the various projects under win32. It's planned to night.
The tests are consisting of compiling and insatling a dozain of projects on linux and win32 (and bith when possible)
Great thanks, let us know how it goes. If there's any question at all about the fixes that have gone in since rc1, we should spin an rc2.
I always thought that the idea of a release candidate was that if there were any fixes _at all_ that there would be a new rc. Only when no bugs needing fixed are found does the rc turn into the actual release. But I understand that this is an idealized rc/release policy :)
--David (RDM)
On Oct 13, 2009, at 1:41 PM, R. David Murray wrote:
I always thought that the idea of a release candidate was that if
there were any fixes _at all_ that there would be a new rc. Only when no bugs needing fixed are found does the rc turn into the actual release. But I understand that this is an idealized rc/release policy :)
I'll invoke the Umbrella Rule now. If we don't do one we'll wish we
had. If we do, we won't have needed it.
-Barry
On Tue, Oct 13, 2009 at 10:45 AM, Barry Warsaw <barry@python.org> wrote:
On Oct 13, 2009, at 1:41 PM, R. David Murray wrote:
I always thought that the idea of a release candidate was that if there were any fixes _at all_ that there would be a new rc. Only when no bugs needing fixed are found does the rc turn into the actual release. But I understand that this is an idealized rc/release policy :)
I'll invoke the Umbrella Rule now. If we don't do one we'll wish we had. If we do, we won't have needed it.
Traceback (most recent call last): File "umbrella_rules.py", line 6, in <module> UmbrellaRule() File "unbrella_rules.py", line 4, in UmbrellaRule raise DuplicateRuleName('http://whatplanetisthis.com/2008/02/the-umbrella-rule-and-sharing-rules-in-g...') DuplicateRuleName: http://whatplanetisthis.com/2008/02/the-umbrella-rule-and-sharing-rules-in-g...
-- --Guido van Rossum (home page: http://www.python.org/~guido/)
On Oct 13, 2009, at 2:00 PM, Guido van Rossum wrote:
Traceback (most recent call last): File "umbrella_rules.py", line 6, in <module> UmbrellaRule() File "unbrella_rules.py", line 4, in UmbrellaRule raise DuplicateRuleName('http://whatplanetisthis.com/2008/02/the-umbrella-rule-and-sharing-rules-in-g...') DuplicateRuleName: http://whatplanetisthis.com/2008/02/the-umbrella-rule-and-sharing-rules-in-g...
MisinformedMeteorologistRule (probably more PC than stupid weatherman
rule :)
-Barry
I always thought that the idea of a release candidate was that if there were any fixes _at all_ that there would be a new rc. Only when no bugs needing fixed are found does the rc turn into the actual release.
This was also my understanding; that's the point of calling it "candidate". Since the code base of 2.6.4rc1 was not released as-is, we would need to consider another candidate.
But then, Barry doesn't like release candidates in the first place.
Regards, Martin
On Oct 13, 2009, at 6:10 PM, Martin v. Löwis wrote:
I always thought that the idea of a release candidate was that if
there were any fixes _at all_ that there would be a new rc. Only when no bugs needing fixed are found does the rc turn into the actual
release.This was also my understanding; that's the point of calling it "candidate". Since the code base of 2.6.4rc1 was not released as-is, we would need to consider another candidate.
But then, Barry doesn't like release candidates in the first place.
No, but let's do one anyway!
So, we can either make Sunday's release rc2 and do the final release
one week later, or I can try to get an rc2 out in the next day or two,
with a final release mid-next week.
Thoughts? -Barry
So, we can either make Sunday's release rc2 and do the final release one week later, or I can try to get an rc2 out in the next day or two, with a final release mid-next week.
Thoughts?
I won't be able to produce Windows binaries until Saturday. Now sure how that would fit into the "within the next day or two" plan.
Regards, Martin
On Oct 13, 2009, at 10:15 PM, Martin v. Löwis wrote:
So, we can either make Sunday's release rc2 and do the final
release one week later, or I can try to get an rc2 out in the next day or two,
with a final release mid-next week.Thoughts?
I won't be able to produce Windows binaries until Saturday. Now sure
how that would fit into the "within the next day or two" plan.
That seems to argue for doing rc2 on Sunday the 18th. If I tag the
release some time Saturday, you could have the binaries by Sunday
right? Then we'll take one more week and release the final on the 25th.
-Barry
On Oct 14, 2009, at 8:06 AM, Martin v. Löwis wrote:
That seems to argue for doing rc2 on Sunday the 18th. If I tag the release some time Saturday, you could have the binaries by Sunday right?
Correct.
Cool. I've updated the Python Release Schedule calendar to reflect
the new dates.
-Barry
On Tue, Oct 13, 2009 at 5:16 PM, Tarek Ziadé <ziade.tarek@gmail.com> wrote:
On Tue, Oct 13, 2009 at 4:52 PM, P.J. Eby <pje@telecommunity.com> wrote:
At 08:27 AM 10/13/2009 -0400, Barry Warsaw wrote:
Are we on track to release 2.6.4 final this Sunday or do we need another rc?
Yesterday, Tarek committed another setuptools related fix and said that he was going to run a bunch of build tests locally. Tarek, how did that go?
I still need to do some more tests, I didn't have time to try the various projects under win32. It's planned to night.
The tests are consisting of compiling and insatling a dozain of projects on linux and win32 (and bith when possible)
Barry,
I'll continue my win32 smoke tests tomorrow, it's not over. So far, so good.
windows XP (vmware) - VC9 - running setup.py build+install
- zope.interface 3.5.2 + setuptools 0.6c9
- Twisted-8.2.0 + setuptools 0.6c9
- pythonwin32-314 +plain distutils: works but I had a failure on 'mc.exe' on one extension which is unrelated obviously. pywin32 builds fine otherwise. Still investigating. on this one
Tarek
participants (9)
-
"Martin v. Löwis" -
Barry Warsaw -
Guido van Rossum -
M.-A. Lemburg -
P.J. Eby -
R. David Murray -
skip＠pobox.com -
ssteinerX＠gmail.com -
Tarek Ziadé