[mach-o]: atomize zero-terminated literals correctly.

When looking through sections with zero-terminated string-literals (__cstring
or __ustring) we were constantly rechecking the first few bytes of the string
for '\0' rather than advancing along. This obviously failed unless all strings
within the section had the same length as that first one.

llvm-svn: 211682

diff --git a/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp b/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp
index 7de85e8..1a03ede 100644 (file)
--- a/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp
+++ b/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp
@@ -352,9 +352,9 @@ std::error_code processSection(DefinedAtom::ContentType atomType,
       case atomizeUTF8:
         // Break section up into zero terminated c-strings.
         size = 0;
-        for (unsigned int i=0; offset+i < e; ++i) {
+        for (unsigned int i = offset; i < e; ++i) {
           if (section.content[i] == 0) {
-            size = i+1;
+            size = i + 1 - offset;
             break;
           }
         }
@@ -362,9 +362,9 @@ std::error_code processSection(DefinedAtom::ContentType atomType,
       case atomizeUTF16:
         // Break section up into zero terminated UTF16 strings.
         size = 0;
-        for (unsigned int i=0; offset+i < e; i += 2) {
-          if ((section.content[i] == 0) && (section.content[i+1] == 0)) {
-            size = i+2;
+        for (unsigned int i = offset; i < e; i += 2) {
+          if ((section.content[i] == 0) && (section.content[i + 1] == 0)) {
+            size = i + 2 - offset;
             break;
           }
         }

diff --git a/lld/test/mach-o/parse-literals.yaml b/lld/test/mach-o/parse-literals.yaml
index 46ddb90..24568ea 100644 (file)
--- a/lld/test/mach-o/parse-literals.yaml
+++ b/lld/test/mach-o/parse-literals.yaml
@@ -18,7 +18,7 @@ sections:
     address:         0x0000000000000100
     content:         [ 0x68, 0x65, 0x6C, 0x6C, 0x6F, 0x00, 0x74, 0x68, 
                        0x65, 0x72, 0x65, 0x00, 0x77, 0x6F, 0x72, 0x6C, 
-                       0x64, 0x00 ]
+                       0x00 ]
   - segment:         __TEXT
     section:         __literal4
     type:            S_4BYTE_LITERALS
@@ -51,7 +51,7 @@ sections:
     address:         0x0000000000000100
     content:         [ 0x68, 0x00, 0x65, 0x00, 0x6C, 0x00, 0x6C, 0x00, 
                        0x6F, 0x00, 0x00, 0x00, 0x74, 0x00, 0x68, 0x00, 
-                       0x65, 0x00, 0x72, 0x00, 0x65, 0x00, 0x00, 0x00 ]
+                       0x65, 0x00, 0x72, 0x00, 0x00, 0x00 ]
 ...
 
 
@@ -64,13 +64,13 @@ sections:
 # CHECK:    content:         [ 74, 68, 65, 72, 65, 00 ]
 # CHECK:  - scope:           hidden
 # CHECK:    type:            c-string
-# CHECK:    content:         [ 77, 6F, 72, 6C, 64, 00 ]
+# CHECK:    content:         [ 77, 6F, 72, 6C, 00 ]
 # CHECK:  - scope:           hidden
 # CHECK:    type:            utf16-string
 # CHECK:    content:         [ 68, 00, 65, 00, 6C, 00, 6C, 00, 6F, 00, 00, 00 ]
 # CHECK:  - scope:           hidden
 # CHECK:    type:            utf16-string
-# CHECK:    content:         [ 74, 00, 68, 00, 65, 00, 72, 00, 65, 00, 00, 00 ]
+# CHECK:    content:         [ 74, 00, 68, 00, 65, 00, 72, 00, 00, 00 ]
 # CHECK:  - scope:           hidden
 # CHECK:    type:            const-4-byte
 # CHECK:    content:         [ 01, 02, 03, 04 ]