[Libreoffice-commits] core.git: i18npool/source

[Tor Lillqvist (via logerrit)]

 i18npool/source/collator/gencoll_rule.cxx |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

New commits:
commit d7edac51367fe92c2659afdfd6e4d7c771dc5229
Author:     Tor Lillqvist <tml at collabora.com>
AuthorDate: Tue May 5 18:58:32 2020 +0300
Commit:     Andras Timar <andras.timar at collabora.com>
CommitDate: Thu May 7 07:08:44 2020 +0200

    ICU requires that a collator charset data array is four-byte-aligned
    
    See workdir/UnpackedTarball/icu/source/common/utrie2.cpp, the
    U_POINTER_MASK_LSB() check:
    
        if( length<=0 || (U_POINTER_MASK_LSB(data, 3)!=0) ||
            valueBits<0 || UTRIE2_COUNT_VALUE_BITS<=valueBits
        ) {
            *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
            return 0;
        }
    
    Apparently the data pointer there is always a multiple of four bytes
    from the start of the data array that the gencoll_rule code generates.
    
    Change-Id: I9b98b01b49b5800e1db8b077a4221b82d59510bf
    Reviewed-on: https://gerrit.libreoffice.org/c/core/+/93507
    Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice at gmail.com>
    Reviewed-by: Tor Lillqvist <tml at collabora.com>
    Reviewed-on: https://gerrit.libreoffice.org/c/core/+/93603
    Tested-by: Jenkins
    Reviewed-by: Andras Timar <andras.timar at collabora.com>

diff --git a/i18npool/source/collator/gencoll_rule.cxx b/i18npool/source/collator/gencoll_rule.cxx
index 091ff6b2fa90..2d55e0c8917d 100644
--- a/i18npool/source/collator/gencoll_rule.cxx
+++ b/i18npool/source/collator/gencoll_rule.cxx
@@ -50,7 +50,7 @@ static void data_write(char* file, char* name, sal_uInt8 *data, sal_Int32 len)
     fprintf(fp, "\nextern \"C\" {\n");
 
     // generate main dict. data array
-    fprintf(fp, "\nstatic const sal_uInt8 %s[] = {", name);
+    fprintf(fp, "\nalignas(4) static const sal_uInt8 %s[] = {", name);
 
     sal_Int32 count = 0;
     for (sal_Int32 i = 0; i < len; i++) {