Introduce a locale-independent version of isspace
The real isspace may give locale-dependent results, so use our own.
This also lets us simplify some of the silliness asn1_string_canon needs
to go through to never pass high bytes into isspace and islower. (I'm
otherwise leaving that function alone because I plan to, later, convert
the whole thing to CBS/CBB.)
Change-Id: Idd349095f3e98bf908bb628ea1089ba05c2c6797
Reviewed-on: https://boringssl-review.googlesource.com/c/boringssl/+/56486
Reviewed-by: Bob Beck <bbe@google.com>
Commit-Queue: David Benjamin <davidben@google.com>
diff --git a/crypto/conf/conf.c b/crypto/conf/conf.c
index f036498..8874896 100644
--- a/crypto/conf/conf.c
+++ b/crypto/conf/conf.c
@@ -780,7 +780,7 @@
lstart = list;
for (;;) {
if (remove_whitespace) {
- while (*lstart && isspace((unsigned char)*lstart)) {
+ while (*lstart && OPENSSL_isspace((unsigned char)*lstart)) {
lstart++;
}
}
@@ -794,7 +794,7 @@
tmpend = lstart + strlen(lstart) - 1;
}
if (remove_whitespace) {
- while (isspace((unsigned char)*tmpend)) {
+ while (OPENSSL_isspace((unsigned char)*tmpend)) {
tmpend--;
}
}
diff --git a/crypto/mem.c b/crypto/mem.c
index af20318..e9d3d72 100644
--- a/crypto/mem.c
+++ b/crypto/mem.c
@@ -315,6 +315,11 @@
return c;
}
+int OPENSSL_isspace(int c) {
+ return c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r' ||
+ c == ' ';
+}
+
int OPENSSL_strcasecmp(const char *a, const char *b) {
for (size_t i = 0;; i++) {
const int aa = OPENSSL_tolower(a[i]);
diff --git a/crypto/test/file_test.cc b/crypto/test/file_test.cc
index e1ab2aa..d7ef951 100644
--- a/crypto/test/file_test.cc
+++ b/crypto/test/file_test.cc
@@ -27,6 +27,7 @@
#include <string.h>
#include <openssl/err.h>
+#include <openssl/mem.h>
#include "../internal.h"
#include "./test_util.h"
@@ -57,11 +58,11 @@
// leading and trailing whitespace removed.
static std::string StripSpace(const char *str, size_t len) {
// Remove leading space.
- while (len > 0 && isspace(*str)) {
+ while (len > 0 && OPENSSL_isspace(*str)) {
str++;
len--;
}
- while (len > 0 && isspace(str[len - 1])) {
+ while (len > 0 && OPENSSL_isspace(str[len - 1])) {
len--;
}
return std::string(str, len);
diff --git a/crypto/x509/x_name.c b/crypto/x509/x_name.c
index a17d9d7..3e7cb94 100644
--- a/crypto/x509/x_name.c
+++ b/crypto/x509/x_name.c
@@ -443,12 +443,10 @@
len = out->length;
- // Convert string in place to canonical form. Ultimately we may need to
- // handle a wider range of characters but for now ignore anything with
- // MSB set and rely on the isspace() and tolower() functions.
+ // Convert string in place to canonical form.
// Ignore leading spaces
- while ((len > 0) && !(*from & 0x80) && isspace(*from)) {
+ while ((len > 0) && OPENSSL_isspace(*from)) {
from++;
len--;
}
@@ -456,7 +454,7 @@
to = from + len;
// Ignore trailing spaces
- while ((len > 0) && !(to[-1] & 0x80) && isspace(to[-1])) {
+ while ((len > 0) && OPENSSL_isspace(to[-1])) {
to--;
len--;
}
@@ -465,13 +463,8 @@
i = 0;
while (i < len) {
- // If MSB set just copy across
- if (*from & 0x80) {
- *to++ = *from++;
- i++;
- }
// Collapse multiple spaces
- else if (isspace(*from)) {
+ if (OPENSSL_isspace(*from)) {
// Copy one space across
*to++ = ' ';
// Ignore subsequent spaces. Note: don't need to check len here
@@ -480,7 +473,7 @@
do {
from++;
i++;
- } while (!(*from & 0x80) && isspace(*from));
+ } while (OPENSSL_isspace(*from));
} else {
*to++ = OPENSSL_tolower(*from);
from++;
diff --git a/crypto/x509v3/v3_conf.c b/crypto/x509v3/v3_conf.c
index 3261302..480bb3b 100644
--- a/crypto/x509v3/v3_conf.c
+++ b/crypto/x509v3/v3_conf.c
@@ -258,7 +258,7 @@
return 0;
}
p += 9;
- while (isspace((unsigned char)*p)) {
+ while (OPENSSL_isspace((unsigned char)*p)) {
p++;
}
*value = p;
@@ -279,7 +279,7 @@
return 0;
}
- while (isspace((unsigned char)*p)) {
+ while (OPENSSL_isspace((unsigned char)*p)) {
p++;
}
*value = p;
diff --git a/crypto/x509v3/v3_utl.c b/crypto/x509v3/v3_utl.c
index f6fdc4b..bb5db40 100644
--- a/crypto/x509v3/v3_utl.c
+++ b/crypto/x509v3/v3_utl.c
@@ -432,14 +432,14 @@
char *p, *q;
// Skip over leading spaces
p = name;
- while (*p && isspace((unsigned char)*p)) {
+ while (*p && OPENSSL_isspace((unsigned char)*p)) {
p++;
}
if (!*p) {
return NULL;
}
q = p + strlen(p) - 1;
- while ((q != p) && isspace((unsigned char)*q)) {
+ while ((q != p) && OPENSSL_isspace((unsigned char)*q)) {
q--;
}
if (p != q) {
diff --git a/include/openssl/mem.h b/include/openssl/mem.h
index f07698d..6216040 100644
--- a/include/openssl/mem.h
+++ b/include/openssl/mem.h
@@ -110,9 +110,14 @@
// OPENSSL_strnlen has the same behaviour as strnlen(3).
OPENSSL_EXPORT size_t OPENSSL_strnlen(const char *s, size_t len);
-// OPENSSL_tolower is a locale-independent version of tolower(3).
+// OPENSSL_tolower is a locale-independent version of tolower(3). It only
+// lowercases ASCII values. Other values are returned as-is.
OPENSSL_EXPORT int OPENSSL_tolower(int c);
+// OPENSSL_isspace is a locale-independent version of isspace(3). It only
+// recognizes '\t', '\n', '\v', '\f', '\r', and ' '.
+OPENSSL_EXPORT int OPENSSL_isspace(int c);
+
// OPENSSL_strcasecmp is a locale-independent version of strcasecmp(3).
OPENSSL_EXPORT int OPENSSL_strcasecmp(const char *a, const char *b);
