Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 3887ac7e authored by Marco Nelissen's avatar Marco Nelissen
Browse files

Skip the endianness marker when converting UCS-2 to UTF-8, since it 

ends up in the UTF-8 string otherwise.
parent 655306f8

media/libstagefright/id3/ID3.cpp

+21 −1
Original line number Diff line number Diff line
@@ -24,6 +24,7 @@ 
#include <media/stagefright/MediaDebug.h>

#include <media/stagefright/Utils.h>

#include <utils/String8.h>

#include <sys/endian.h>



namespace android {
@@ -336,7 +337,26 @@ void ID3::Iterator::getString(String8 *id) const { 
    } else {

        // UCS-2

        // API wants number of characters, not number of bytes...

        id->setTo((const char16_t *)(mFrameData + 1), n / 2);

        int len = n / 2;

        const char16_t *framedata = (const char16_t *) (mFrameData + 1);

        char16_t *framedatacopy = NULL;

        if (*framedata == 0xfffe) {

            // endianness marker doesn't match host endianness, convert

            framedatacopy = new char16_t[len];

            for (int i = 0; i < len; i++) {

                framedatacopy[i] = swap16(framedata[i]);

            }

            framedata = framedatacopy;

        }

        // If the string starts with an endianness marker, skip it

        if (*framedata == 0xfeff) {

            framedata++;

            len--;

        }

        id->setTo(framedata, len);

        if (framedatacopy != NULL) {

            delete[] framedatacopy;

        }

    }

}