[Scons-users] Previous build information lost with ProtoBuilder/Gch interaction

[William Blevins]

We would probably need more information about your scripts.

If any sources are being dynamically generated (emitter changes between
1st-2nd compile) or an emitter outputs the wrongs files, then SCons may
rebuild more than once and could possibly never be up-to-date in the latter
case.  If either of these issues are the culprit, then "--tree=<options>"
will be more useful than explain because the tree will show either files
that don't exist, be missing files that exist, or the tree will change from
build to build.

V/R,
William

On Mon, Aug 3, 2015 at 10:36 PM, Carl Cerecke <carl.cerecke at compacsort.com>
wrote:

> I have two Sconscript files, both building a single static library,
> and both called from the same SConstruct (they are the only two
> Sconscript files called).
>
> One StaticLibrary uses ProtocBuilder
> (https://bitbucket.org/scons/scons/wiki/ProtocBuilder) as part of its
> build. Let's call this library libProto.
> The other StaticLibrary uses g++ precompiled headers
> (https://bitbucket.org/scons/scons/wiki/GchBuilder) as part of its
> build. Let's call this libGch.
>
> Building from scratch works fine.
>
> Problem: Rebuilding (after no changes) builds everything again and
> --debug=explain gives the message "scons: Cannot explain why
> `/path/to/build/Filename.o' is being rebuilt: No previous build
> information found"
>
> If I stop libGch from using precompiled headers, the rebuild (for both
> libraries) works as expected.
>
> If I stop LibProto from depending on the .cc file that was generated
> by the ProtocBuilder, then the rebuild (for both libraries) works as
> expected.
>
> What is utterly confusing is that, whether or not libGch finds
> previous build information or not, depends on whether an unrelated
> dependency exists in a *separate* Sconscript file that should be
> completely independent!
>
> Both Sconscript files use a clone of the environment they imported
> from the parent SConstruct, so they should be independent.
>
> I've tried env.Decider('MD5') and env.Decider('MD5-timestamp').
> I've tried looking at the differences in .sconsign.dblite between
> libGch working correctly, and libGch failing to find previous build
> information - nothing seemed obviously amiss.
> Likewise I've looked at any difference in environment, but again, it
> all seems OK.
>
> I checked the code in Node.__init__.py and noticed that the 'old'
> variable in the explain() method doesn't have any of the attributes
> {'bsources','bdepends','bimplicit'} which causes the exception. It
> does have the -sigs version of those attributes, but they are simply
> empty lists. I don't know what that means, though.
>
> I'm using scons 2.3.1
>
> Any ideas on what might be going wrong?
>
> I might extract a small example set of files which exhibit the problem
> if that would help
> --
> Carl Cerecke
> _______________________________________________
> 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/20150804/529c708f/attachment.html>