Re: [PING global maintainer/testsuite expert] Re: [patch bfd]: Win32 coff-i386 and coff-x86_64 broken in_reloc_p

[Kai Tietz <ktietz70 at googlemail dot com>]

2010/3/31 Dave Korn <dave.korn.cygwin@googlemail.com>:
> On 31/03/2010 20:16, Kai Tietz wrote:
>> 2010/3/31 Ralf Wildenhues <Ralf.Wildenhues@gmx.de>:
>>> * Kai Tietz wrote on Wed, Mar 31, 2010 at 09:05:28PM CEST:
>>>> 2010/3/31 Ralf Wildenhues:
>>>>> wc -c is POSIX and portable, but you cannot be sure about the amount of
>>>>> whitespace in the output: for example, IRIX, AIX, Tru64, FreeBSD,
>>>>> OpenBSD wc produce varying amounts of leading whitespace, GNU doesn't.
>>>>> Does tcl strip that? ?Otherwise, if all you're trying to find out is
>>>>> whether a file is nonempty or not, testing the exit status of shell
>>>>> 'test -s' works.
>>>> There are no whitespaces. It is an binary file-format and no text.
>>> I was talking about the output of
>>> $ wc -c somefile
>>>
>>> which can be either of
>>> ? ?1204 somefile
>>> ? ? ?1204 somefile
>>> ? ? ? ? ?1204 somefile
>>> 1204 somefile
>>>
>>> not the contents of "somefile". ?I haven't checked TABs vs. spaces.
>>
>> Ah, ok. Sorry, I got this wrong. I am not sure if 'string match' trims
>> for white-spaces.
>
> ?Probably not. ?How about:
>
>> - ? ?if ![string match "0 tmpdir/$testname.base" $err] then {
>> + ? ?if ![string match "0 tmpdir/$testname.base" [string trim $err]] then {
>
> ?That ought to do it. ?It's certainly a safe precautionary change, so Kai,
> it's pre-approved if you'd like to just test it and make sure it doesn't break.
>
> ? ?cheers,
> ? ? ?DaveK
>
>

Yes, works nicely. I committed patch with ChangeLog

        * ld-pe//pe-compile.exp (run_basefile_test): Trim result of wc
        before string compare.

Cheers,
Kai
-- 
|  (\_/) This is Bunny. Copy and paste
| (='.'=) Bunny into your signature to help
| (")_(") him gain world domination