From d3331fba6eaab44646010b556106da38925d43e0 Mon Sep 17 00:00:00 2001
From: jsmall-nvidia <jsmall@nvidia.com>
Date: Thu, 6 Feb 2020 14:31:09 -0500
Subject: Literal handling improvements (#1202)

* WIP: 64 literal diagnostic and truncation.

* Improve how integer truncation is handled/supported.
Added literal-int64.slang test.
Set a suffix on all literals.
Fixed problem on C++ based targets where l suffix was not the same as int() cast. So on C++ derived emitters, int() is used instead of l suffix to have same behavior across targets.

* Add literal diagnostic testing.

* Allow lexer to lex - in front of literals.

* Fix lexing and converting int literal with -.

* Too large small values of floats become inf.
Handling writing inf types out on different targets.
Add function to deterimine if a float literals kind.

* Roll back the support of lexer lexing negative literals.

* Fixed tests broken because of diagnostics numbers.
Improved _isFinite

* Fix compilation on linux.

* Fix problem with abs on linux - use Math::Abs.

* Fix typo.

* * Improve warnings for float literals zeroed
* Improved 64 bit type documentation
* Handle half
* Improved comments
* Fixed tests broken
* Use capital letters for suffixes.

* Make default behavior on outputting a int literal that is an 'int32_t' is cast (not suffix) to avoid platform inconsistencies.
Improve documentation for 64 bit types.
Make tests cover material in docs.

* Fixed tests.

* Rename FloatKind::Normal -> Finite

* Fix half zero check.
---
 tests/diagnostics/float-literal.slang          | 23 ++++++++++++++++++
 tests/diagnostics/float-literal.slang.expected |  9 ++++++++
 tests/diagnostics/int-literal.slang            | 32 ++++++++++++++++++++++++++
 tests/diagnostics/int-literal.slang.expected   |  7 ++++++
 4 files changed, 71 insertions(+)
 create mode 100644 tests/diagnostics/float-literal.slang
 create mode 100644 tests/diagnostics/float-literal.slang.expected
 create mode 100644 tests/diagnostics/int-literal.slang
 create mode 100644 tests/diagnostics/int-literal.slang.expected

(limited to 'tests/diagnostics')

diff --git a/tests/diagnostics/float-literal.slang b/tests/diagnostics/float-literal.slang
new file mode 100644
index 000000000..46ce048b9
--- /dev/null
+++ b/tests/diagnostics/float-literal.slang
@@ -0,0 +1,23 @@
+//DIAGNOSTIC_TEST:SIMPLE:
+
+
+float doSomething(float a)
+{
+    // Too large with become +inf
+    a += 5e+40;
+    // Will be narrowed to 0
+    a += 9e-50;
+    
+    double b = 0.0f;
+    
+    // These shouldn't produce warning as they can fit 
+    b += 5e+40l;
+    b += 9e-50l;
+    
+    // Cos these don't have l suffix they should also produce warnings
+    // and produce -inf and 0
+    b += -5e+40;
+    b += -9e-50;
+    
+    return a + float(b);
+}
diff --git a/tests/diagnostics/float-literal.slang.expected b/tests/diagnostics/float-literal.slang.expected
new file mode 100644
index 000000000..347317666
--- /dev/null
+++ b/tests/diagnostics/float-literal.slang.expected
@@ -0,0 +1,9 @@
+result code = 0
+standard error = {
+tests/diagnostics/float-literal.slang(7): warning 39999: float literal '5e+40' unrepresentable, converted to 'inf'
+tests/diagnostics/float-literal.slang(9): warning 39999: '9e-50' is smaller than the smallest representable value for type float, converted to '0'
+tests/diagnostics/float-literal.slang(19): warning 39999: float literal '5e+40' unrepresentable, converted to 'inf'
+tests/diagnostics/float-literal.slang(20): warning 39999: '9e-50' is smaller than the smallest representable value for type float, converted to '0'
+}
+standard output = {
+}
diff --git a/tests/diagnostics/int-literal.slang b/tests/diagnostics/int-literal.slang
new file mode 100644
index 000000000..3724d0e14
--- /dev/null
+++ b/tests/diagnostics/int-literal.slang
@@ -0,0 +1,32 @@
+//DIAGNOSTIC_TEST:SIMPLE:
+
+int doSomething(int a)
+{
+    // Warning can't fit
+    int c0 = 0x800000000;
+    
+    // No warning as top bits are just ignored
+    int c1 = -1ll;
+    
+    int c2 = int(-1u);
+    
+    // Should sign extend 
+    int c3 = 0x80000000;
+    
+    // Should give a warning (ideally including the preceeding -)
+    // Currently we don't have the -, because the lexer lexes - independently
+    int c4 = -0xfffffffff;
+    
+    // 
+    a += c0 + c1 + c2;
+    
+    int64_t b = 0;
+
+    // Ok
+    b += 0x800000000ll;
+    
+    uint64_t c5 = -2ull;
+    
+    return a + int(b);
+}
+
diff --git a/tests/diagnostics/int-literal.slang.expected b/tests/diagnostics/int-literal.slang.expected
new file mode 100644
index 000000000..ffc5ff6d2
--- /dev/null
+++ b/tests/diagnostics/int-literal.slang.expected
@@ -0,0 +1,7 @@
+result code = 0
+standard error = {
+tests/diagnostics/int-literal.slang(6): warning 39999: integer literal '0x800000000' too large for type 'int' truncated to '0'
+tests/diagnostics/int-literal.slang(18): warning 39999: integer literal '0xfffffffff' too large for type 'int' truncated to '-1'
+}
+standard output = {
+}
-- 
cgit v1.2.3