Ignore:
Timestamp:
Jul 24, 2006, 6:02:14 PM (18 years ago)
Author:
Ken Moffat <zarniwhoop@…>
Children:
e970bd81
Parents:
5e56006
Message:

Fix most of the tests in glibc-32 on multilib, thanks to Go Moko. Anybody with access to mips64 should consider whether this needs to be applied to N32, please.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • final-system/multilib/glibc.xml

    r5e56006 r8b3ca42  
    8888    xpointer="xpointer(//*[@os='l'])"/>
    8989
     90    <para os="l1">In multilib, we tend to think that compiling for
     91    <literal>${CLFS_TARGET32}</literal> is <emphasis>not</emphasis>
     92    cross-compiling. Glibc takes the traditional view that if you are building
     93    for a different host then you are cross-compiling, so you won&quot;t be running
     94    the tests and therefore you don&quot;t need the locale files. When we run
     95    the tests, many will fail if the locale files are missing. The following
     96    sed allows these tests to succeed:</para>
     97
     98<screen os="l2"><userinput>sed -i '/cross-compiling/s@ifeq@ifneq@g' ../glibc-2.4/localedata/Makefile</userinput></screen>
     99
    90100    <xi:include xmlns:xi="http://www.w3.org/2003/XInclude"
    91101    href="../common/glibc.xml"
     
    103113    href="../common/glibc.xml"
    104114    xpointer="xpointer(//*[@os='p'])"/>
     115
     116<!-- now that we can see what is really failing, we can address arch-specific failures here -->
    105117
    106118    <xi:include xmlns:xi="http://www.w3.org/2003/XInclude"
Note: See TracChangeset for help on using the changeset viewer.