Merge pull request #2682 from aG0aep6G/stride

std.utf.stride: reject sequence lengths greater than 4
dlang · Nov 12, 2014 · dc7eb9d · dc7eb9d
2 parents 858296f + 4ce7db5
commit dc7eb9d
Show file tree

Hide file tree

Showing 4 changed files with 21 additions and 6 deletions.
diff --git a/posix.mak b/posix.mak
@@ -88,8 +88,8 @@ WEBSITE_DIR = ../web
 DOC_OUTPUT_DIR = $(WEBSITE_DIR)/phobos-prerelease
 BIGDOC_OUTPUT_DIR = /tmp
 SRC_DOCUMENTABLES = index.d $(addsuffix .d,$(STD_MODULES) $(STD_NET_MODULES) $(STD_DIGEST_MODULES) $(STD_CONTAINER_MODULES) $(EXTRA_DOCUMENTABLES))
-STDDOC = $(DOCSRC)/std.ddoc
-BIGSTDDOC = $(DOCSRC)/std_consolidated.ddoc
+STDDOC = $(DOCSRC)/std.ddoc $(DOCSRC)/macros.ddoc
+BIGSTDDOC = $(DOCSRC)/std_consolidated.ddoc $(DOCSRC)/macros.ddoc
 # Set DDOC, the documentation generator
 DDOC=$(DMD) -m$(MODEL) -w -c -o- -version=StdDdoc \
 	-I$(DRUNTIME_PATH)/import $(DMDEXTRAFLAGS)

diff --git a/std/utf.d b/std/utf.d
@@ -138,7 +138,8 @@ unittest
     $(D index) defaults to $(D 0) if none is passed.
 
     Returns:
-        The number of bytes in the UTF-8 sequence.
+        The number of bytes in the UTF-8 sequence, a value between 1 and 4
+        (as per $(WEB tools.ietf.org/html/rfc3629#section-3, RFC 3629$(COMMA) section 3)).
 
     Throws:
         May throw a $(D UTFException) if $(D str[index]) is not the start of a
@@ -186,7 +187,7 @@ body
 {
     import core.bitop : bsr;
     immutable msbs = 7 - bsr(~c);
-    if (msbs < 2 || msbs > 6)
+    if (msbs < 2 || msbs > 4)
         throw new UTFException("Invalid UTF-8 sequence", index);
     return msbs;
 }
@@ -258,6 +259,20 @@ unittest
     });
 }
 
+unittest // invalid start bytes
+{
+    import std.exception: assertThrown;
+    immutable char[] invalidStartBytes = [
+        0b1111_1000, // indicating a sequence length of 5
+        0b1111_1100, // 6
+        0b1111_1110, // 7
+        0b1111_1111, // 8
+        0b1000_0000, // continuation byte
+    ];
+    foreach(c; invalidStartBytes)
+        assertThrown!UTFException(stride([c]));
+}
+
 
 /++
     $(D strideBack) returns the length of the UTF-8 sequence ending one code

diff --git a/win32.mak b/win32.mak
@@ -61,7 +61,7 @@ DMD=dmd
 ## Location of where to write the html documentation files
 
 DOCSRC = ../dlang.org
-STDDOC = $(DOCSRC)/std.ddoc
+STDDOC = $(DOCSRC)/std.ddoc $(DOCSRC)/macros.ddoc
 
 DOC=..\..\html\d\phobos
 #DOC=..\doc\phobos

diff --git a/win64.mak b/win64.mak
@@ -61,7 +61,7 @@ DMD=dmd
 ## Location of where to write the html documentation files
 
 DOCSRC = ../dlang.org
-STDDOC = $(DOCSRC)/std.ddoc
+STDDOC = $(DOCSRC)/std.ddoc $(DOCSRC)/macros.ddoc
 
 DOC=..\..\html\d\phobos
 #DOC=..\doc\phobos