Re: Compile programs on sarge (glibc 2.3.2) to run on woody (glibc 2.2.5)?
On Sat, 05 Feb 2005 21:30:19 +0100, Roberto Sanchez wrote:
>
>Reid Priedhorsky wrote:
>>
>> I would like to compile programs on my sarge box and have them run
>> correctly on a woody box. Sometimes (most of the time?), this works fine.
>> However, sometimes it breaks; when running on the woody box I get:
>>
>> /lib/libc.so.6: version `GLIBC_2.3' not found (required by ./dlc)
>>
>> An objdump analysis on the program reveals dependency on two glibc 2.3
>> symbols:
>>
>> 00000000 F *UND* 00000071 __ctype_toupper_loc@@GLIBC_2.3
>> 00000000 F *UND* 00000071 __ctype_b_loc@@GLIBC_2.3
>>
>> When compiled on woody, the program works fine on both systems.
>> The two __ctype* symbols appear in the woody-compiled program as:
>>
>> 080adf48 g O .bss 00000004 __ctype_toupper@@GLIBC_2.0
>> 080adf4c g O .bss 00000004 __ctype_b@@GLIBC_2.0
>>
>> So -- is there anything I can do to request that glibc 2.3-specific
>> symbols not be used when I compile programs (using current sarge gcc and
>> toolchain)?
>>
>
> Out of curiosity, why not build a Woody chroot with the development
> tools only installed inside of it? That way you can develop on your
> Sarge box and then compile and fix any issues if the compile fails
> in your Woody chroot.
Well, that's more effort and disk space than I'd like to put into it.
Right now I can just copy the source over to the woody box and compile
there.
To clarify, this happens on several different programs, some of which I
wrote and some of which I didn't. "The program" in my original post is one
that didn't work that I decided to investigate recently. The other time it
happened, there was something similar but it might have been a different
symbol that was the culprit. I wouldn't know how to "fix any issues"
because I don't know what is bringing in the glibc-2.3 symbols -- I
certainly didn't call the offending functions in the program I wrote.
Thanks,
Reid
Reply to: