[PD] Building Pd for MacOS (was MIDI timing FIFO overflowed receiving sysex)

[Paul Rankin]

On Tue, 27 Jun 2017, at 07:56 PM, Dan Wilcox wrote:
> 
> In the future, it would be helpful for us to know if the included scripts are not working so we can check & fix them :)
> 
> My entire goal of the autotools update was to make building Pd easier so user testing is important.

Cool. I will redo my build steps next week and report back.

Could the functionality of the mac/*.sh scripts be incorporated into the Makefile? This feels neater to me.

> > Ah yes, sorry I stand corrected. The above hack came from a PR I was going to submit about the script looking for the framework in /Library first, then falling back on /System, so the user can just type
> > 
> > $ ./osx-app.sh -s 8.6 VERSION
> > 
> > instead of 
> > 
> > $ ./osx-app.sh -w /Library/Frameworks/Tk.framework/Versions/Current/Resources/Wish.app VERSION
> > 
> > So, no hacking necessary, but I found the hacking easier than remembering where Wish was.
> 
> 
> I can add this behavior. I built the script with only the system installed Tcl/Tk to check. Good to know the location of the custom ActiveState install.

One caveat here is that I'm pretty sure the framework location changed from 8.4 or 8.5, or maybe on the macOS side... Not sure. But I do recall the ActiveTcl documentation having an outdated path listed. Maybe there is a more programatic way of finding the framework...

> > I agree the naming convention is useful. I don't think the build should fail without it though.
> 
> I was a bit lazy there as there isn't an easy way to grab the version string without parsing the first line in configure.ac. Your suggestion of just leaving it off without the version argument is even easier. I believe I *did* add the version string info already in a previous commit.

I was being even lazier... I was going to open a PR to make VERSION optional in the script, but then I didn't and just complained about it instead.


-- 
www.paulwrankin.com