[Scons-users] SConscripts on the fly

[Bill Deegan]

Spencer,

You could read the config files directly in python and process them to do
the same as a generated SConscript would via a pseudobuilder..

https://docs.python.org/2/library/configparser.html

Assuming that your current package builds are described by a config file..

-Bill

On Tue, May 16, 2017 at 3:27 PM, Spencer Yost <syost at triad.rr.com> wrote:

> Thanks folks,
>
> Part of the reason is because our current build system uses a file in each
> package that specifies specific linkage instructions (i.e. libraries),
> final deliverable ownership and permissions,  installation directories,
> etc.  It also specifies different types of build behavior(compile, simple
> copy, authomated unit test, etc).
>
> With a simple pearl program using the config:: STD package to read the old
> file, I can generate an SConscript on the fly very easily that will mimic
> final deliverables of the old system.   Since it's so easy I figured just
> going that route would be the best way to showcase SCons.  I also like how
> we can use this mechanism to ease the transition from old system to new
> system.
>
> I may have mentioned it, I'm not sure. What I am trying to do is to
> convince a very large development department that the way we build stuff is
> antiquated, complicated and difficult to port and we need to move to SCons.
> So right now I do have to try to stay on the easiest, quickest path to an
> impressive show and tell without cutting too many corners insofar as best
> practices because I also don't want SCons looking like a bunch of python
> hacks embedded in package directories
>
>
>
> Spencer Yost
>
> On May 16, 2017, at 12:16 PM, Bill Deegan <bill at baddogconsulting.com>
> wrote:
>
> +1 on Eric's sugguestion.
> Unless you're planning on using variant_dir in your SConscript calls to
> build multiple variants.
>
> -Bill
>
> On Tue, May 16, 2017 at 8:13 AM, Eric Lunderberg <
> eric.lunderberg at gmail.com> wrote:
>
>> As a potential alternative, if the SConscripts can be automatically
>> generated from a top-level SConstruct, why not instead generate the build
>> actions directly?  That way, the intermediate SConscript would not need to
>> be generated at all.  If some directories require customized behavior, then
>> those directories could be the only ones that contain a SConscript, and the
>> top-level SConstruct could preferentially use such a SConscript if it
>> exists.
>>
>> On Tue, May 16, 2017 at 11:02 AM, William Blevins <wblevins001 at gmail.com>
>> wrote:
>>
>>> I've never tried doing something like that but maybe making a custom
>>> builder that produces sconscript files would be appropriate.
>>>
>>> On May 16, 2017 10:23 AM, "Spencer Yost" <syost at triad.rr.com> wrote:
>>>
>>>> Thanks again for y'all's help a few weeks ago, we have SCons behaving
>>>> pretty much as we expect. Still tweaking and playing to put together a
>>>> final show and tell.
>>>>
>>>> Quick history:  As you may recall, we have hundreds of packages in a
>>>> reverse domain tree that encompasses 10,000+  files.  We have decided that
>>>> we would prefer creating the SConscript file in each package dynamically on
>>>> the fly. There are several reasons for this that I will not bore you with.
>>>>
>>>> I'm trying to figure out the best way to do this. A PreAction() clearly
>>>> won't work. I could use brute force and simply execute something out of the
>>>> top level SConstruct where we find them and call SConscript(), but that
>>>> seems like a hack.  It just "seems" like there should be a better way.
>>>>
>>>> What would you folks consider to be best practice for creating
>>>> SConscripts on the fly?
>>>>
>>>> Thanks in advance!
>>>>
>>>> Spencer Yost
>>>>
>>>> _______________________________________________
>>>> Scons-users mailing list
>>>> Scons-users at scons.org
>>>> https://pairlist4.pair.net/mailman/listinfo/scons-users
>>>>
>>>>
>>> _______________________________________________
>>> Scons-users mailing list
>>> Scons-users at scons.org
>>> https://pairlist4.pair.net/mailman/listinfo/scons-users
>>>
>>>
>>
>> _______________________________________________
>> Scons-users mailing list
>> Scons-users at scons.org
>> https://pairlist4.pair.net/mailman/listinfo/scons-users
>>
>>
> _______________________________________________
> Scons-users mailing list
> Scons-users at scons.org
> https://pairlist4.pair.net/mailman/listinfo/scons-users
>
>
> _______________________________________________
> Scons-users mailing list
> Scons-users at scons.org
> https://pairlist4.pair.net/mailman/listinfo/scons-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist4.pair.net/pipermail/scons-users/attachments/20170516/d45ae4cd/attachment.html>