fix: linux build

[ShortDevelopment]

palinternal.h defines a fallback for the min and max macros in case the compiler does not:

ChakraCore/pal/src/include/pal/palinternal.h

Lines 689 to 696 in 36becec

	// do not change the location of the definitions below.
	// these defs should be always at the end
	#ifndef max
	#define max(a,b) (((a) > (b)) ? (a) : (b))
	#endif
	#ifndef min
	#define min(a,b) (((a) < (b)) ? (a) : (b))
	#endif

finite.cpp imports palinternal.h before <math> so we end up with two conflicting definitions of min and max:

ChakraCore/pal/src/cruntime/finite.cpp

Lines 22 to 24 in 36becec

	#include "pal/palinternal.h"
	#include "pal/dbgmsg.h"
	#include <math.h>

---

@rhuanjl This is the first in a stack of PRs that will hopefully fix MacOS ci.
Until then, MacOS builds will fail...

[ShortDevelopment]

As expected

• the MacOS build fails due to an out-of-date minimum CMake version
  See fix: MacOS build #7040
• Linux builds pass now

Strangely, one win-x86 test fails.
Maybe due to "Out of Memory"?

ChakraCore/test/typedarray/allocation.js

Lines 6 to 12 in 36becec

	for (i = 0; i < 1000; i++) {
	var a = new ArrayBuffer(10000000);
	for (j = 0; j < 20; j++) {
	var b = new Object();
	}
	}
	WScript.Echo("pass");

[rhuanjl]

As expected

• the MacOS build fails due to an out-of-date minimum CMake version

  See fix: MacOS build #7040

• Linux builds pass now

Strangely, one win-x86 test fails.

Maybe due to "Out of Memory"?

ChakraCore/test/typedarray/allocation.js

Lines 6 to 12 in 36becec

	for (i = 0; i < 1000; i++) {
	var a = new ArrayBuffer(10000000);
	for (j = 0; j < 20; j++) {
	var b = new Object();
	}
	}
	WScript.Echo("pass");

That win x86 test has been flakey for a long time - sometimes passes, sometimes hits out of memory.

[ShortDevelopment]

Hey, thank you for the quick merge!
Next in the stack is #7040 but I still have some strange issues with byte-code regen on linux...