[Bf-committers] Keep SCons relevant to CMake

[Campbell Barton]

On Mon, Dec 22, 2014 at 9:57 AM, Sergey Sharybin <sergey.vfx at gmail.com> wrote:
> Hey,
>
> This rant is caused by the following change:
> https://developer.blender.org/rBe67fd7a (Cambo, sorry, this commit just
> happened to be the most recent which demonstrates the issue we're having).
>
> Well, it was fair enough change to split appdir routines into dedicated
> file keeping path utils small and nice. And was fair enough change to fix
> CMake compilation. But hey, where's the change to SCons?!
>
> I would ask everyone to pay more attention on making sure changes to either
> of SCons or CMake build systems are also ported to another system. I
> wouldn't expect it to be huge overhead to do such a tweaks for both systems.
>
> And yes, all the rants about "hey, SCons is stupid and CMake is just much
> nicer" could be send directly to /dev/null. I do not care about such a
> statements for as long SCons continues to be officially supported and used
> for release builds process.
>
> End of rant, now let's get back to 2.73 release work.
>
> --
> With best regards, Sergey Sharybin

I don't think this is really a special case, (re: SCons/CMake),
Its stupid mistake which are hard to entirely avoid.
In this case I was a bit distracted from regular Blender development,
but of course its very idiot not to fix both build-systems.
In this particular case Blender runs just fine without BINRELOC for
many users. Not an excuse - but why it went unnoticed.

What I regret is not being able to fix own errors,
normally I keep an eye on the tracker and try to make to investigate
any reports related to my own work.

However fixing other peoples bugs is pretty common too, in the last
week I found 2 really stupid mistakes in other developers new code
which I fixed without making a fuss (not to place blame, just to point
out its common).

-- 
- Campbell