[coreboot-gerrit] New patch to review for coreboot: 2da339f libpayload: UTF-16LE to ASCII conversion

[Patrick Georgi (pgeorgi@google.com)]

Patrick Georgi (pgeorgi at google.com) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/8733

-gerrit

commit 2da339fb48fcaea86cc2544818d8426538443ec0
Author: Dan Ehrenberg <dehrenberg at chromium.org>
Date:   Fri Nov 21 15:50:27 2014 -0800

    libpayload: UTF-16LE to ASCII conversion
    
    This patch adds a simple function to convert a string in UTF-16LE
    to ASCII.
    
    TEST=Ran against a string found in a GPT with the intended outcome
    BRANCH=none
    BUG=none
    
    Change-Id: I94ec0a32f5712259d3d0caec2233c992330228e3
    Signed-off-by: Patrick Georgi <pgeorgi at chromium.org>
    Original-Commit-Id: 1104db8328a197c7ccf6959a238277f416a2113a
    Original-Signed-off-by: Dan Ehrenberg <dehrenberg at chromium.org>
    Original-Change-Id: I50ca5bfdfbef9e084321b2beb1b8d4194ca5af9c
    Original-Reviewed-on: https://chromium-review.googlesource.com/231456
    Original-Reviewed-by: Julius Werner <jwerner at chromium.org>
---
 payloads/libpayload/include/string.h |  7 +++++++
 payloads/libpayload/libc/string.c    | 19 +++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/payloads/libpayload/include/string.h b/payloads/libpayload/include/string.h
index 952261f..ca5d513 100644
--- a/payloads/libpayload/include/string.h
+++ b/payloads/libpayload/include/string.h
@@ -69,6 +69,13 @@ char *strerror(int errnum);
 /** @} */
 
 /**
+ * @defgroup string Unicode functions
+ * @{
+ */
+char *utf16le_to_ascii(uint16_t *utf16_string, int maxlen);
+/** @} */
+
+/**
  * @defgroup string OpenBSD based safe string functions
  * @{
  */
diff --git a/payloads/libpayload/libc/string.c b/payloads/libpayload/libc/string.c
index 71dd1a6..549fe0c 100644
--- a/payloads/libpayload/libc/string.c
+++ b/payloads/libpayload/libc/string.c
@@ -653,3 +653,22 @@ char *strerror(int errnum)
 	snprintf(errstr, sizeof(errstr), "Unknown error %d", errnum);
 	return errstr;
 }
+
+/*
+ * Simple routine to convert UTF-16 to ASCII, giving up with ? if too high.
+ * A single code point may convert to ?? if in not in the BMP.
+ * @param utf16_string A string encoded in UTF-16LE
+ * @param maxlen Maximum possible length of the string in code points
+ * @return Newly allocated ASCII string
+ */
+char *utf16le_to_ascii(uint16_t *utf16_string, int maxlen)
+{
+	char *ascii_string = xmalloc(maxlen + 1);  /* +1 for trailing \0 */
+	ascii_string[maxlen] = '\0';
+	int i;
+	for (i = 0; i < maxlen; i++) {
+		uint16_t wchar = utf16_string[i];
+		ascii_string[i] = wchar > 0x7f ? '?' : (char)wchar;
+	}
+	return ascii_string;
+}