[regexp] Include unicode/uvernum.h in parser

This patch explicitly includes unicode/uvernum.h in the regular
expression parser.

It should be removed once we no longer need to check
`U_ICU_VERSION_MAJOR_NUM` during preprocessing, i.e. once Node.js
updates their ICU. This is an ongoing effort:
https://github.com/nodejs/node/pull/16876

BUG=v8:4743

Change-Id: I3cd9447b481249a9035d9fb00745057da8809c58
Reviewed-on: https://chromium-review.googlesource.com/758407Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Cr-Commit-Position: refs/heads/master@{#49253}

src/regexp/regexp-parser.cc

@@ -14,6 +14,10 @@
 #include "src/regexp/jsregexp.h"
 #include "src/utils.h"
 
+// TODO(mathias): Remove this when we no longer need to check
+// `U_ICU_VERSION_MAJOR_NUM`.
+#include "unicode/uvernum.h"
+
 #ifdef V8_INTL_SUPPORT
 #include "unicode/uniset.h"
 #endif  // V8_INTL_SUPPORT