Detect icu4c error codes more precisely
Negative values are warnings, not errors; failing on them breals encoding when it would otherwise work. Unfortunately, these negative numbers are being converted into large positive ones through the cgo bridge, so this becomes a slightly more complicated check than necessary.
This commit is contained in:
18
c_bridge.c
18
c_bridge.c
@@ -15,7 +15,7 @@ const int detectCharset(void *detector,
|
||||
|
||||
// Put input bytes in the detector.
|
||||
ucsdet_setText((UCharsetDetector*)detector, (char*)input, input_len, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
@@ -25,7 +25,7 @@ const int detectCharset(void *detector,
|
||||
|
||||
// Perform analysis and return all guesses and their count.
|
||||
bestGuesses = ucsdet_detectAll((UCharsetDetector*)detector, &matchCount, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
@@ -42,19 +42,19 @@ const int detectCharset(void *detector,
|
||||
|
||||
// Fill guessed encoding
|
||||
bestGuessedCharset = ucsdet_getName(bestGuess, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
// Fill guessed language
|
||||
bestGuessedLanguage = ucsdet_getLanguage(bestGuess, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
// Fill its confidence rating
|
||||
int32_t conf = ucsdet_getConfidence(bestGuess, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
@@ -77,13 +77,13 @@ int convertToUtf16(const char *srcEncoding,
|
||||
UConverter *conv;
|
||||
|
||||
conv = ucnv_open(srcEncoding, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
/* Convert from original encoding to UTF-16 */
|
||||
int len = ucnv_toUChars(conv, dest, destCapacity, src, srcLength, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
@@ -102,13 +102,13 @@ int convertFromUtf16(const char *destEncoding,
|
||||
UConverter *conv;
|
||||
|
||||
conv = ucnv_open(destEncoding, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
/* Convert from UTF-16 to destination encoding */
|
||||
int len = ucnv_fromUChars(conv, dest, destCapacity, src, srcLength, status);
|
||||
if (*status != U_ZERO_ERROR) {
|
||||
if U_FAILURE(*status) {
|
||||
return 0;
|
||||
}
|
||||
|
||||
|
@@ -79,7 +79,7 @@ func (conv *CharsetConverter) ConvertToUtf8(input []byte, srcEncoding string) ([
|
||||
C.int32_t(len(input)),
|
||||
(*C.int)(unsafe.Pointer(&status)))
|
||||
|
||||
if status == U_ZERO_ERROR {
|
||||
if isSuccess(status) {
|
||||
nConvLen := C.convertFromUtf16(
|
||||
Utf8CString,
|
||||
(*C.char)(unsafe.Pointer(&conv.utf8Buffer[0])),
|
||||
@@ -88,7 +88,7 @@ func (conv *CharsetConverter) ConvertToUtf8(input []byte, srcEncoding string) ([
|
||||
C.int32_t(convLen),
|
||||
(*C.int)(unsafe.Pointer(&status)))
|
||||
|
||||
if status == U_ZERO_ERROR {
|
||||
if isSuccess(status) {
|
||||
resStr := conv.utf8Buffer[:nConvLen]
|
||||
return ([]byte)(resStr), nil
|
||||
}
|
||||
|
20
detect.go
20
detect.go
@@ -11,10 +11,22 @@ import (
|
||||
)
|
||||
|
||||
const (
|
||||
U_ZERO_ERROR = 0 // ICU common constant error code which means that no error occured
|
||||
MatchDataBufferSize = 25 // Size of the buffer for detection results (Max count of returned guesses per detect call)
|
||||
U_ZERO_ERROR = 0 // ICU common constant error code which means that no error occured
|
||||
U_ERROR_LIMIT = 0x7FFFFFFF // Dirty hack, negative error codes are are being turned into large positive ints
|
||||
MatchDataBufferSize = 25 // Size of the buffer for detection results (Max count of returned guesses per detect call)
|
||||
)
|
||||
|
||||
// Go implementation of the icu U_SUCCESS macro. Negative status codes are
|
||||
// warnings, 0 is a success without warnings, > 0 is an error
|
||||
func isSuccess(status int) bool {
|
||||
return status <= U_ZERO_ERROR || status >= U_ERROR_LIMIT
|
||||
}
|
||||
|
||||
// Go implementation of the icu U_FAILURE macro.
|
||||
func isFailure(status int) bool {
|
||||
return status > U_ZERO_ERROR && status < U_ERROR_LIMIT
|
||||
}
|
||||
|
||||
// CharsetDetector provides ICU charset detection functionality.
|
||||
type CharsetDetector struct {
|
||||
ptr *C.UCharsetDetector // ICU struct needed for detection
|
||||
@@ -39,7 +51,7 @@ func NewCharsetDetector() (*CharsetDetector, error) {
|
||||
|
||||
det.ptr = C.ucsdet_open((*C.UErrorCode)(statusPtr))
|
||||
|
||||
if status != U_ZERO_ERROR {
|
||||
if isFailure(status) {
|
||||
return nil, fmt.Errorf("ICU Error code returned: %d", status)
|
||||
}
|
||||
|
||||
@@ -70,7 +82,7 @@ func (det *CharsetDetector) GuessCharset(input []byte) (matches []Match, err err
|
||||
(*C.MatchData)(unsafe.Pointer(&det.resBuffer[0])),
|
||||
C.int(MatchDataBufferSize))
|
||||
|
||||
if status == U_ZERO_ERROR {
|
||||
if isSuccess(status) {
|
||||
// Convert the returned number of entries from result buffer to a slice
|
||||
// that will be returned
|
||||
count := int(guessCount)
|
||||
|
Reference in New Issue
Block a user