World: r4wp

I didn't, no.

Please don't repurpose 1921 :)

Let 1921 stay dismissed, OK.

One ticket for missing BROWSE on OSX, another one for the browse 
none crash.

https://github.com/rebol/r3/blob/master/src/os/win32/host-lib.c#L640

Will do. And I'm still revamping the REWORD ticket example code.

Well, I'll briefly write up the crash ticket.

If you could write both I'd appreciate it. I have limited R3 time 
today, I have an R2 task I have to work on.

Ok.

https://github.com/0branch/r3/commit/95e3481ffba473f9191b918dd2eb890e764ff32a

If that looks good, I can submit pull requests for both issues once 
Andreas has written up the issues?

I'm not sure if the second is the desired change :)

MarcS: one open questions is if we should use the __APPLE__ define 
(is that defined for iOS as well?) or rather use the R3-specific 
platform defines.

Oh, did I misunderstand?

Win32 handles a none URL gracefully, just opening an "empty" browser.

Oh, okay. But this meets Brian's suggestion?

Yes.

A function should only return none if returning another value is 
a valid option. If it never returns a value, it shouyld return unset.

Incidentally, there's no simple way to browse to none using the current 
technique (as far as I can see).

BrianH: oh okay, so this is a unit type / void.

Yeah.

e.g., on MacOSX, open uses the format of its argument to determine 
the appropriate handler. I'd imagine xdg-open does the same.

And i doubt browsing to file:/// is the desired default.

(Replace file:/// with anything else, I think the point stands.)

Or desired to allow at all, right.

What does `open http://` do on OSX?

Incidentally, what's R3 bazaar? I was a little confused as to whether 
that was a repo for experimental developments or something else.

Andreas: oh hey, about:blank

(Same with xdg-open.)

It's MaxV's thing.

But xdg-open takes me to http:///

(Invalid adress.)

Brian: but pull requests should still go to the main repo?

Yes.

address*

(Sorry for the 20 questions.)

Glad if we can help :)

browse none doesn't work right on Win32 unless IE is your default 
browser. It behaves differently wrong for every other default browser.

Aha, interesting.

Add that to the browse none ticket, Andreas, we can say "All Platforms".

So we probably better rather remove this feature altogether.

> MarcS: one open question re your original pull is if we should 
use the __APPLE__ define

> (is that defined for iOS as well?) or rather use the R3-specific 
platform defines.

It would be worth taking none out of the argument spec and marking 
the ticket as a problem until we figure out how to do this correctly.

Stylistically I'd imagine Rebol's defines are preferred?

Yes please, if we have an appropriate one. And if we don't, we should.

But are they reliable? I can't see reference to either __APPLE__ 
or __MACH__ in the codebase

They are build-driven.

You explicitly build for OSX, then the platform defines are set.

TO_MAC, right?

That isn't set correctly