Re: [gpsd-dev] scons quirks

[Hal Murray]

>> On the non-git world, revision.h gets updated each time you run scons.
>> In the git world, revision.h doesn't get updated by a second scons.

> Actually, it does.  But because scons checks out-of-dateness
> by MD5 hash of content rather than timestamp, the update
> doesn't trigger any rebuilds.

It's one step more complicated than that.  The write/create date doesn't get 
updated.


> There is a can of worms here, all proceeding from what we think the version
> stamp is for.  There are a couple of other possibilities for how to update
> revision.h when there's no .git directory. 

I don't have any good ideas, but at least I know what's going on now.  (Well, 
at least in that corner.)

I could push .git/ around and install git so everybody would use the git 
version.

One possibility would be to run something some code after the git-pull and 
before the rsync.  I can't see how to do that cleanly, but since it would be 
in a script, I'd be happy with a not-squeaky-clean solution that worked.  
(The sort of problem I'm trying to avoid is doing the pull (say by hand) but 
forgetting to run the magic step that updates revision.h.)

It might be reasonable to construct a revision string at run time rather than 
build time.  I think there is an easy way to get the link time-stamp, but I 
don't know how to do it.  The current SConstruct checks for a "dev" in 
gpsd_version which is a constant in SConstruct.

-------

There is still the problem of scons-install relinking things.


-- 
These are my opinions.  I hate spam.