Index: src/blob.c
==================================================================
--- src/blob.c
+++ src/blob.c
@@ -1095,35 +1095,30 @@
** done. If useMbcs is false and there is no BOM, the input string is assumed
** to be UTF-8 already, so no conversion is done.
*/
void blob_to_utf8_no_bom(Blob *pBlob, int useMbcs){
char *zUtf8;
- int bomSize = 0;
- if( starts_with_utf8_bom(pBlob, &bomSize) ){
+ int bomSize = starts_with_bom(pBlob);
+ if( bomSize == 3 ){
struct Blob temp;
zUtf8 = blob_str(pBlob) + bomSize;
blob_zero(&temp);
blob_append(&temp, zUtf8, -1);
blob_swap(pBlob, &temp);
blob_reset(&temp);
#ifdef _WIN32
- }else if( starts_with_utf16le_bom(pBlob, &bomSize) ){
- /* Make sure the blob contains two terminating 0-bytes */
- blob_append(pBlob, "", 1);
- zUtf8 = blob_str(pBlob) + bomSize;
- zUtf8 = fossil_unicode_to_utf8(zUtf8);
- blob_zero(pBlob);
- blob_append(pBlob, zUtf8, -1);
- fossil_unicode_free(zUtf8);
- }else if( starts_with_utf16be_bom(pBlob, &bomSize) ){
- unsigned int i = blob_size(pBlob);
+ }else if( bomSize == 2 ){
zUtf8 = blob_buffer(pBlob);
- while( i > 0 ){
- /* swap bytes of unicode representation */
- char zTemp = zUtf8[--i];
- zUtf8[i] = zUtf8[i-1];
- zUtf8[--i] = zTemp;
+ if (*((unsigned short *)zUtf8) == 0xfffe) {
+ /* Found BOM, but with reversed bytes */
+ unsigned int i = blob_size(pBlob);
+ while( i > 0 ){
+ /* swap bytes of unicode representation */
+ char zTemp = zUtf8[--i];
+ zUtf8[i] = zUtf8[i-1];
+ zUtf8[--i] = zTemp;
+ }
}
/* Make sure the blob contains two terminating 0-bytes */
blob_append(pBlob, "", 1);
zUtf8 = blob_str(pBlob) + bomSize;
zUtf8 = fossil_unicode_to_utf8(zUtf8);
Index: src/checkin.c
==================================================================
--- src/checkin.c
+++ src/checkin.c
@@ -899,17 +899,17 @@
int binOk, /* Non-zero if binary warnings should be disabled. */
int encodingOk, /* Non-zero if encoding warnings should be disabled. */
const char *zFilename /* The full name of the file being committed. */
){
int eType; /* return value of looks_like_utf8/utf16() */
- int fUnicode; /* return value of starts_with_utf16_bom() */
+ int fUnicode; /* 1 if blob starts with UTF-16 BOM */
char *zMsg; /* Warning message */
Blob fname; /* Relative pathname of the file */
static int allOk = 0; /* Set to true to disable this routine */
if( allOk ) return 0;
- fUnicode = starts_with_utf16_bom(p, 0);
+ fUnicode = (starts_with_bom(p) == 2);
eType = fUnicode ? looks_like_utf16(p) : looks_like_utf8(p);
if( eType==0 || eType==-1 || fUnicode ){
const char *zWarning;
const char *zDisable;
const char *zConvert = "c=convert/";
Index: src/diff.c
==================================================================
--- src/diff.c
+++ src/diff.c
@@ -340,72 +340,31 @@
if( pnByte ) *pnByte = 3;
return bom;
}
/*
-** This function returns non-zero if the blob starts with a UTF-8
-** byte-order-mark (BOM).
+** This function returns detected BOM size if the blob starts with
+** a UTF-8, UTF-16le or UTF-16be byte-order-mark (BOM).
*/
-int starts_with_utf8_bom(const Blob *pContent, int *pnByte){
+int starts_with_bom(const Blob *pContent){
const char *z = blob_buffer(pContent);
- int bomSize = 0;
+ int c1, bomSize = 0;
const unsigned char *bom = get_utf8_bom(&bomSize);
- if( pnByte ) *pnByte = bomSize;
- if( blob_size(pContent)<bomSize ) return 0;
- return memcmp(z, bom, bomSize)==0;
-}
-
-/*
-** This function returns non-zero if the blob starts with a UTF-16le or
-** UTF-16be byte-order-mark (BOM).
-*/
-int starts_with_utf16_bom(const Blob *pContent, int *pnByte){
- const char *z = blob_buffer(pContent);
- int c1, c2;
-
- if( pnByte ) *pnByte = 2;
- if( blob_size(pContent)<2 ) return 0;
- c1 = z[0]; c2 = z[1];
- if( (c1==(char)0xff) && (c2==(char)0xfe) ){
- return 1;
- }else if( (c1==(char)0xfe) && (c2==(char)0xff) ){
- return 1;
- }
- return 0;
-}
-
-/*
-** This function returns non-zero if the blob starts with a UTF-16le
-** byte-order-mark (BOM).
-*/
-int starts_with_utf16le_bom(const Blob *pContent, int *pnByte){
- const char *z = blob_buffer(pContent);
- int c1, c2;
-
- if( pnByte ) *pnByte = 2;
- if( blob_size(pContent)<2 ) return 0;
- c1 = z[0]; c2 = z[1];
- if( (c1==(char)0xff) && (c2==(char)0xfe) ){
- return 1;
- }
- return 0;
-}
-
-/*
-** This function returns non-zero if the blob starts with a UTF-16be
-** byte-order-mark (BOM).
-*/
-int starts_with_utf16be_bom(const Blob *pContent, int *pnByte){
- const char *z = blob_buffer(pContent);
- int c1, c2;
-
- if( pnByte ) *pnByte = 2;
- if( blob_size(pContent)<2 ) return 0;
- c1 = z[0]; c2 = z[1];
- if( (c1==(char)0xfe) && (c2==(char)0xff) ){
- return 1;
+ if( (blob_size(pContent)>=bomSize)
+ && (memcmp(z, bom, bomSize)==0) ){
+ return bomSize;
+ }
+ /* Only accept UTF-16 BOM if the blob has an even number of bytes */
+ if( (blob_size(pContent)<2) || (blob_size(pContent)&1) ) return 0;
+ c1 = *((unsigned short *)z);
+ if( (c1==0xfffe) || (c1==0xfeff) ){
+ if( blob_size(pContent)>=4 ){
+ /* For UTF-32 BOM, always return 0. */
+ if( ((unsigned short *)z)[1] == 0 ) return 0;
+ }
+ return 2;
}
return 0;
}
/*
@@ -2369,11 +2328,11 @@
if( zLimit==0 || zLimit[0]==0 ) zLimit = "-1";
iLimit = atoi(zLimit);
showLog = find_option("log",0,0)!=0;
fileVers = find_option("filevers",0,0)!=0;
db_must_be_within_tree();
- if (g.argc<3) {
+ if( g.argc<3 ){
usage("FILENAME");
}
file_tree_name(g.argv[2], &treename, 1);
zFilename = blob_str(&treename);
fnid = db_int(0, "SELECT fnid FROM filename WHERE name=%Q", zFilename);
@@ -2383,11 +2342,11 @@
fid = db_int(0, "SELECT rid FROM vfile WHERE pathname=%Q", zFilename);
if( fid==0 ){
fossil_fatal("not part of current checkout: %s", zFilename);
}
cid = db_lget_int("checkout", 0);
- if (cid == 0){
+ if( cid == 0 ){
fossil_fatal("Not in a checkout");
}
if( iLimit<=0 ) iLimit = 1000000000;
compute_direct_ancestors(cid, iLimit);
mid = db_int(0, "SELECT mlink.mid FROM mlink, ancestor "