[Scons-dev] SCons 2.4.0 Released

Tim Jenness tjenness at lsst.org
Wed Sep 23 14:36:15 EDT 2015


> On Sep 23, 2015, at 11:32 , Dirk Bächle <tshortik at gmx.de> wrote:
> 
> Hi all,
> 
> On 23.09.2015 12:28, Russel Winder wrote:
>> On Tue, 2015-09-22 at 09:13 +0200, Dirk Bächle wrote:
>>> 
>> […]
>>> 
>>> I volunteer to weed out some of the old Python 2.x stuff (where x < 7
>>> ;) ) on a separate bug branch. I have all the scons_testsuite
>>> stuff ready (just like for the slots change), in order to show that
>>> there are no large speed penalties coming with it.
>>> 
>>> 
>> Does this mean we can revert all the 2.6-isms that were introduced
>> instead of using nice 2.7, 3.3+-isms?
>> 
> 
> 
> I'd suggest to continue as follows:
> 
>  - Remove special Python constructs for Python 2.5 and smaller on a bug branch, then merge it
>    to tip of "default" as usual.
>  - I'd then like to add another issue on top of this where I do some changes to the core architecture
>    (pushing classes around to entangle the dependencies of Node and Taskmaster a little bit), again
>    merged to "default".
>  - Then, merge this new tip of "default" to the "python3" branch.
>  - Continue with bugfixing on "default", while on "python3" we can then start to remove 2.6 constructs
>    and work on getting the code to be 2.7/3.x compatible at the same time.
> 

Has anyone tried to run futurize with the “-1” option? It’s specifically designed to modify 2.x code to the 2.7 compatibility standard without requiring any external code (does not need to “import future” but does add “from __future__…”). It specifically does not try for python 3 support in that mode.

— 
Tim Jenness




More information about the Scons-dev mailing list