Delete IsAtLeastOS10_9() and IsAtMostOS10_9()

base/sys_info_mac.mm

Index: base/sys_info_mac.mm
diff --git a/base/sys_info_mac.mm b/base/sys_info_mac.mm
index aab1103d4c401d14efc6bf7eecb30c208b9a5d98..2e3d9fc32dceaee3af8c7f4f4a706a99c659caaf 100644
--- a/base/sys_info_mac.mm
+++ b/base/sys_info_mac.mm
@@ -51,7 +51,7 @@ void SysInfo::OperatingSystemVersionNumbers(int32_t* major_version,
     // cases in 10.9, rely on ::Gestalt(..). Since this code is only needed for
     // 10.9.0 and 10.9.1 and uses the recommended replacement thereafter,
     // suppress the warning for this fallback case.
-    DCHECK(base::mac::IsOS10_9());

[Nico, 2016/08/29 19:44:23]

Hm, the lhs seems better to me. When we dropped old macOS versions in the past,
we grepped for (say) "10.8" and that let us find a bunch of code to delete. This
is now harder. Is there any rush to remove Is10_9?

[Sidney San Martín, 2016/08/30 15:27:04]

FWIW, grepping doesn't get any harder. When we drop 10.9, we'd search for:

- IsAtLeastOS10_10(): now guaranteed to be true with or without `!`, time to
delete some code.
- IsOS10_10() -> !IsAtLeastOS10_11()
- IsAtMost10_10() -> !IsAtLeastOS10_11()

…and then we could delete the *10_10() functions.

On the other hand, if it's more of a semantic meaning/readability thing —
IsOS10_9() doesn't actually mean the same thing as !IsOS10_10(), even though it
does in practice — then how do you feel about reinstating IsOS10_9() and
reverting the more-confusing changes?

[Sidney San Martín, 2016/08/30 15:27:55]

See latest patch set.

+    DCHECK(!base::mac::IsAtLeastOS10_10());
 #pragma clang diagnostic push
 #pragma clang diagnostic ignored "-Wdeprecated-declarations"
     Gestalt(gestaltSystemVersionMajor,