From: Alex Dowad <alexinbeijing@gmail.com>
Date: Sun, 6 Sep 2020 12:42:55 +0000 (+0200)
Subject: Add identify filter for UTF-32{,BE,LE}
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=7047e5d2c409c073799139490b92ffde2e6b1afb;p=php

Add identify filter for UTF-32{,BE,LE}
---

diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf32.c b/ext/mbstring/libmbfl/filters/mbfilter_utf32.c
index 2566ff5476..ddd942d6e9 100644
--- a/ext/mbstring/libmbfl/filters/mbfilter_utf32.c
+++ b/ext/mbstring/libmbfl/filters/mbfilter_utf32.c
@@ -30,6 +30,10 @@
 #include "mbfilter.h"
 #include "mbfilter_utf32.h"
 
+static int mbfl_filt_ident_utf32(int c, mbfl_identify_filter *filter);
+static int mbfl_filt_ident_utf32le(int c, mbfl_identify_filter *filter);
+static int mbfl_filt_ident_utf32be(int c, mbfl_identify_filter *filter);
+
 static const char *mbfl_encoding_utf32_aliases[] = {"utf32", NULL};
 
 const mbfl_encoding mbfl_encoding_utf32 = {
@@ -65,6 +69,24 @@ const mbfl_encoding mbfl_encoding_utf32le = {
 	&vtbl_wchar_utf32le
 };
 
+const struct mbfl_identify_vtbl vtbl_identify_utf32 = {
+	mbfl_no_encoding_utf32,
+	mbfl_filt_ident_common_ctor,
+	mbfl_filt_ident_utf32
+};
+
+const struct mbfl_identify_vtbl vtbl_identify_utf32be = {
+	mbfl_no_encoding_utf32be,
+	mbfl_filt_ident_common_ctor,
+	mbfl_filt_ident_utf32be
+};
+
+const struct mbfl_identify_vtbl vtbl_identify_utf32le = {
+	mbfl_no_encoding_utf32le,
+	mbfl_filt_ident_common_ctor,
+	mbfl_filt_ident_utf32le
+};
+
 const struct mbfl_convert_vtbl vtbl_utf32_wchar = {
 	mbfl_no_encoding_utf32,
 	mbfl_no_encoding_wchar,
@@ -289,3 +311,128 @@ int mbfl_filt_conv_wchar_utf32le(int c, mbfl_convert_filter *filter)
 
 	return c;
 }
+
+static int mbfl_filt_ident_utf32(int c, mbfl_identify_filter *filter)
+{
+	/* The largest valid codepoint is 0x10FFFF; we don't want values above that
+	 * Neither do we want to see surrogates
+	 * For UTF-32 (not LE or BE), we do also need to look for a byte-order mark */
+	switch (filter->status) {
+	case 0: /* 1st byte */
+		if (c == 0xff) {
+			filter->status = 1;
+			return c;
+		}
+		filter->filter_function = mbfl_filt_ident_utf32be;
+		break;
+
+	case 1: /* 2nd byte */
+		if (c == 0xfe) {
+			filter->status = 2;
+			return c;
+		}
+		filter->filter_function = mbfl_filt_ident_utf32be;
+		(filter->filter_function)(0xff, filter);
+		break;
+
+	case 2: /* 3rd byte */
+		if (c == 0) {
+			filter->status = 3;
+			return c;
+		}
+		filter->filter_function = mbfl_filt_ident_utf32be;
+		(filter->filter_function)(0xff, filter);
+		(filter->filter_function)(0xfe, filter);
+		break;
+
+	case 3: /* 4th byte */
+		if (c == 0) {
+			/* We found a little-endian byte-order mark! */
+			filter->status = 0;
+			filter->filter_function = mbfl_filt_ident_utf32le;
+			return c;
+		}
+		filter->filter_function = mbfl_filt_ident_utf32be;
+		(filter->filter_function)(0xff, filter);
+		(filter->filter_function)(0xfe, filter);
+		(filter->filter_function)(0, filter);
+		break;
+	}
+
+	return (filter->filter_function)(c, filter);
+}
+
+static int mbfl_filt_ident_utf32le(int c, mbfl_identify_filter *filter)
+{
+	switch (filter->status) {
+	case 0: /* 1st byte */
+		filter->status = 1;
+		break;
+
+	case 1: /* 2nd byte */
+		if (c >= 0xD8 && c <= 0xDF) {
+			filter->status = 4; /* might be surrogate if we are in BMP */
+		} else {
+			filter->status = 2;
+		}
+		break;
+
+	case 2: /* 3rd byte */
+		if (c > 0x10) {
+			filter->flag = 1; /* too big */
+		}
+		filter->status = 3;
+		break;
+
+	case 3: /* 4th byte */
+		if (c) {
+			filter->flag = 1; /* too big */
+		}
+		filter->status = 0;
+		break;
+
+	case 4: /* 3rd byte, previous byte looked like surrogate */
+		if (!c) {
+			filter->flag = 1; /* yep, it's a surrogate */
+		}
+		filter->status = 3;
+	}
+	return c;
+}
+
+static int mbfl_filt_ident_utf32be(int c, mbfl_identify_filter *filter)
+{
+	switch (filter->status) {
+	case 0: /* 1st byte */
+		if (c) {
+			filter->flag = 1; /* too big */
+		}
+		filter->status = 1;
+		break;
+
+	case 1: /* 2nd byte */
+		if (c > 0x10) {
+			filter->flag = 1; /* too big */
+		} if (c) {
+			filter->status = 4; /* not in the BMP */
+		} else {
+			filter->status = 2;
+		}
+		break;
+
+	case 2: /* 3rd byte */
+		if (c >= 0xD8 && c <= 0xDF) {
+			filter->flag = 1; /* reserved range for surrogates */
+		}
+		filter->status = 3;
+		break;
+
+	case 3: /* 4th byte */
+		filter->status = 0;
+		break;
+
+	case 4: /* 3rd byte, not in BMP */
+		filter->status = 3;
+	}
+	return c;
+}
diff --git a/ext/mbstring/libmbfl/filters/mbfilter_utf32.h b/ext/mbstring/libmbfl/filters/mbfilter_utf32.h
index 14b9fa4823..de1235e451 100644
--- a/ext/mbstring/libmbfl/filters/mbfilter_utf32.h
+++ b/ext/mbstring/libmbfl/filters/mbfilter_utf32.h
@@ -33,6 +33,9 @@
 extern const mbfl_encoding mbfl_encoding_utf32;
 extern const mbfl_encoding mbfl_encoding_utf32be;
 extern const mbfl_encoding mbfl_encoding_utf32le;
+extern const struct mbfl_identify_vtbl vtbl_identify_utf32;
+extern const struct mbfl_identify_vtbl vtbl_identify_utf32be;
+extern const struct mbfl_identify_vtbl vtbl_identify_utf32le;
 extern const struct mbfl_convert_vtbl vtbl_utf32_wchar;
 extern const struct mbfl_convert_vtbl vtbl_wchar_utf32;
 extern const struct mbfl_convert_vtbl vtbl_utf32be_wchar;
diff --git a/ext/mbstring/libmbfl/mbfl/mbfl_ident.c b/ext/mbstring/libmbfl/mbfl/mbfl_ident.c
index d2d991249f..a48762ee3d 100644
--- a/ext/mbstring/libmbfl/mbfl/mbfl_ident.c
+++ b/ext/mbstring/libmbfl/mbfl/mbfl_ident.c
@@ -169,6 +169,9 @@ static const struct mbfl_identify_vtbl *mbfl_identify_filter_list[] = {
 	&vtbl_identify_ucs2,
 	&vtbl_identify_ucs2be,
 	&vtbl_identify_ucs2le,
+	&vtbl_identify_utf32,
+	&vtbl_identify_utf32be,
+	&vtbl_identify_utf32le,
 	&vtbl_identify_false,
 	NULL
 };