[Scons-users] Question about binfo

William Blevins wblevins001 at gmail.com
Thu Feb 25 12:47:01 EST 2016


I'm not sure, but thanks for bringing this to attention because I have
looked into binfo related cache issues in the past and was not aware of
this behavior. I wonder if it is related.

Thanks,
William

On Thu, Feb 25, 2016 at 3:55 PM, Tom Tanner (BLOOMBERG/ LONDON) <
ttanner2 at bloomberg.net> wrote:

> In the Node/__init__.py, in the clear() method, it says this:
>
> # The del_binfo() call here isn't necessary for normal execution,
> # but is for interactive mode, where we might rebuild the same
> # target and need to start from scratch.
> self.del_binfo()
>
> This is causing me some interesting problems. I'm trying to stash the
> *ENTIRE* command line used to build an object (partly for rerunning
> commands, partly for getting better details of actual command line changes
> rather than the rather brief summary produced currently).
>
> Actually calculating the command line seems to be somewhat expensive (you
> would not believe the length of one of ours) so I wanted to only calculate
> it on an as needed basis and then stick it in the nodes binfo so it gets
> saved the the .scons.dblite file.
>
> But that line above stops this working. And I'm not sure I see the point
> of it. If 'interactive' should repeatedly build an object even if nothing
> has changed, wouldn't it be better to delete the object instead?
>
> For reference - expensive is - if I make this happen as needed, it saves
> about 2 minutes in a 12 minute build.
>
> Incidentally, I was going to put the stashing of the command line in
> prepare, but it seems that 'prepare' is being called for every node in the
> tree, which means I need to put the generator into 'build' or 'built' and
> 'explain', probably the latter as I really can't tell offhand how thread
> safe the code is.
>
>
>
> _______________________________________________
> 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/20160225/706f5c9b/attachment.html>


More information about the Scons-users mailing list