scripts/build/arch/arm.sh
author "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
Sun Jan 25 17:53:14 2009 +0000 (2009-01-25)
changeset 1150 218f7a0e8d71
parent 903 9fb0f81b4416
child 1591 11460fc587e6
permissions -rw-r--r--
On 20090120.2203+0100, Andy Johnson <ajohnson@aecno.com> wrote:
The glibc.sh script doesn't handle the glibc versions with
an underscore very well (bash expected integer error). I
have attached a small patch for that. Instead of looking
for "not period" I changed the sense to look for numbers.
I initially tried to make it look for either a period or
an underscore, but that didn't work like I wanted (probably
because I did something wrong).

Original patch modified to be more robust.

/trunk/scripts/build/libc/glibc.sh | 8 4 4 0 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
     1 # Compute ARM-specific values
     2 
     3 CT_DoArchTupleValues() {
     4     # The architecture part of the tuple:
     5     CT_TARGET_ARCH="${CT_ARCH}${target_endian_eb}"
     6 
     7     # The system part of the tuple:
     8     case "${CT_LIBC},${CT_ARCH_ARM_EABI}" in
     9         *glibc,y)   CT_TARGET_SYS=gnueabi;;
    10         uClibc,y)   CT_TARGET_SYS=uclibcgnueabi;;
    11         none,y)     CT_TARGET_SYS=eabi;;
    12     esac
    13 
    14     # In case we're EABI, do *not* specify any ABI!
    15     # which means, either we do not have an ABI specified, or we're not EABI.
    16     CT_TestOrAbort "Internal error: CT_ARCH_ABI should not be set for EABI build." -z "${CT_ARCH_ABI}" -o -z "${CT_ARCH_ARM_EABI}"
    17 }