[Stackless] merging changes

Anselm Kruis a.kruis at science-computing.de
Wed Nov 6 20:24:17 CET 2013


Hi,

Am 01.11.2013 15:24, schrieb Kristján Valur Jónsson:
> Okay, I have merged all of those in.
>
> One thing I noticed is that Anselm has been very tidy in keeping the stackless/changelog.txt up to date.
> This is something I have not been doing :)
>
> I have made a bunch of changes to 2.7 and need to update the changelog accordingly.

Good idea. We trying to sell a product based on stackless to very large 
and sometimes bureaucratic IT organisations. They like regular releases 
and nice change logs.:-)

>
> The changelog then needs to be migrated to 3.2 and 3.3.
>
> However, it seems a bit awkward to be keeping the same minor release cadence as cpython.
> Surely we can adopt a micro version number, or something,
>
> e.g. 2.7.5.1  is cpython 2.7.5 with stackless version 1
>
> Does this make sense?

I also thought about it and I'm worried about compatibility. We have 
formal version numbers in sys.hexversion and sys.version_info as well as 
informal information in sys.version and sys.subversion. On windows we 
also have the uuid in the msi installer.

If we add a stackless specific version information, we must adhere to 
the definitions for sys.hexversion and sys.version_info. The 
documentation of sys.hexversion states, that the serial-bits of a final 
version are always 0. The "releaselevel" bits of a final release are 
always 0xf.

Now if we think of stackless as an enhanced version of C-Python, we 
could use the "serial"-component of the version information. If we 
define the serial of a final stackless release to be "1" and use our own 
uuid, we could even install the stackless msi and the C-Python msi 
without conflicts.

> Maybe that is all moot since I don't know if we are making proper regular stackless releases anymore :(
Why not?

Regards
   Anselm



>
> K
>
>> -----Original Message-----
>> From: stackless-bounces at stackless.com [mailto:stackless-
>> bounces at stackless.com] On Behalf Of Anselm Kruis
>> Sent: 28. október 2013 20:30
>> To: The Stackless Python Mailing List
>> Subject: Re: [Stackless] merging changes
>>
>> Hi,
>>
>> Am 26.10.2013 12:51, schrieb Kristján Valur Jónsson:
>>> As long as the features have unittests (stackless/unittests/...) then there is
>> no special testing infrastructure needed.
>> That's perfectly valid as long as everything works as expected. But
>> sometimes things break in a subtle way and you need to closely investigate
>> the problem. For instance, a valid 2.7 test case could be invalid for 3.x.
>> Unfortunately I didn't use python 3.x yet and therefore I would probably
>> need a lot of time which I don't have.
>>
>>> Everything we add should be unit-testable if possible.  I have tried
>>> to make it a habit to add these tests for all that I do
>> and my 3.x testing when porting features consists of running the test suite.
>> +1
>>
>> As promised, here is a list of changes to be ported to 3.x.
>>
>> http://www.stackless.com/ticket/24
>> http://hg.python.org/stackless/rev/2612a56e0082
>>
>> http://www.stackless.com/ticket/22
>> http://www.stackless.com/changeset/00459789dc70
>> Perhaps we should refactor and simplify the f_execute logic at the end of
>> PyEval_EvalFrame_value. I didn't look into any 3.x version, but in
>> 2.7 it is really hard to read.
>>
>> http://www.stackless.com/ticket/20
>> http://www.stackless.com/changeset/91ff56a0cd3c
>> At least the test case.
>>
>> http://www.stackless.com/ticket/18
>> http://www.stackless.com/changeset/7d4f4b6fdf06
>> http://www.stackless.com/changeset/7513dc7259b3
>>
>> http://www.stackless.com/ticket/15
>> http://www.stackless.com/changeset/0f46dcc0abf0
>>
>> http://www.stackless.com/ticket/14
>> http://www.stackless.com/changeset/53f0e5446729
>>
>> http://www.stackless.com/ticket/12
>> http://www.stackless.com/changeset/56982efdeb04
>>
>> Regards
>>     Anselm
>>
>>
>>> K
>>>
>>> -----Original Message-----
>>> From: Anselm Kruis [mailto:a.kruis at science-computing.de]
>>> Sent: 25. október 2013 19:09
>>> To: Kristján Valur Jónsson
>>> Cc: stackless at stackless.com
>>> Subject: Re: [Stackless] merging changes
>>>
>>> Hi Kristján,
>>>
>>> here at s+c we still use Python 2.7 almost exclusively. Therefore I have no
>> real Python3 test infrastructure. I would appreciate it very much, if you could
>> merge my changes into 3.*. I already added comments about this to the
>> stackless issues, but I'll summarise it for you later this weekend.
>>>
>>> best regards
>>>      Anselm
>>>
>>>
>>> Am 25.10.2013 16:59, schrieb Kristján Valur Jónsson:
>>>> Hello there.
>>>> I notice that there have been a number of changes to 2.7 in the last
>> months by other people than me.
>>>> Anselm and Christian mostly.
>>>>
>>>> We should be careful to keep merging these changes to the 3.2 and 3.2
>> branches.
>>>> Merging from 2.7 to 3.2 is best done with the "graft" extension, I always
>> do this for individual changes.
>>>> Then, a branch merge from 3.2 to 3.3. can be performed, bringing the
>> changes all the way.
>>>>
>>>> Anselm and Christian, do you want to do this?  Alternatively, if you
>> identify the revisions that need merging, I can do it for you, n.p.
>>>>
>>>>
>>>> K
>>>>
>>>>
>>>>
>>>> Hello there.
>>>>
>>>> I notice that there have been a number of changes to 2.7 in the last
>>>> months by other people than me.
>>>>
>>>> Anselm and Christian mostly.
>>>>
>>>> We should be careful to keep merging these changes to the 3.2 and 3.2
>>>> branches.
>>>>
>>>> Merging from 2.7 to 3.2 is best done with the "graft" extension, I
>>>> always do this for individual changes.
>>>>
>>>> Then, a branch merge from 3.2 to 3.3. can be performed, bringing the
>>>> changes all the way.
>>>>
>>>> Anselm and Christian, do you want to do this?  Alternatively, if you
>>>> identify the revisions that need merging, I can do it for you, n.p.
>>>>
>>>> K
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Stackless mailing list
>>>> Stackless at stackless.com
>>>> http://www.stackless.com/mailman/listinfo/stackless
>>>>
>>>
>>
>> --
>>    Dipl. Phys. Anselm Kruis                       science + computing ag
>>    Senior Solution Architect                      Ingolstädter Str. 22
>>    email A.Kruis at science-computing.de             80807 München, Germany
>>    phone +49 89 356386 874  fax 737               www.science-computing.de
>> --
>> Vorstandsvorsitzender/Chairman of the board of management:
>> Gerd-Lothar Leonhart
>> Vorstand/Board of Management:
>> Dr. Bernd Finkbeiner, Michael Heinrichs, Dr. Arno Steitz, Dr. Ingrid Zech
>> Vorsitzender des Aufsichtsrats/ Chairman of the Supervisory Board:
>> Philippe Miltin
>> Sitz/Registered Office: Tuebingen
>> Registergericht/Registration Court: Stuttgart Registernummer/Commercial
>> Register No.: HRB 382196
>>
>>
>> _______________________________________________
>> Stackless mailing list
>> Stackless at stackless.com
>> http://www.stackless.com/mailman/listinfo/stackless
>
>
>
> _______________________________________________
> Stackless mailing list
> Stackless at stackless.com
> http://www.stackless.com/mailman/listinfo/stackless
>

-- 
  Dipl. Phys. Anselm Kruis                       science + computing ag
  Senior Solution Architect                      Ingolstädter Str. 22
  email A.Kruis at science-computing.de             80807 München, Germany
  phone +49 89 356386 874  fax 737               www.science-computing.de
-- 
Vorstandsvorsitzender/Chairman of the board of management:
Gerd-Lothar Leonhart
Vorstand/Board of Management:
Dr. Bernd Finkbeiner, Michael Heinrichs, 
Dr. Arno Steitz, Dr. Ingrid Zech
Vorsitzender des Aufsichtsrats/
Chairman of the Supervisory Board:
Philippe Miltin
Sitz/Registered Office: Tuebingen
Registergericht/Registration Court: Stuttgart
Registernummer/Commercial Register No.: HRB 382196




More information about the Stackless mailing list