RE: textutils version 2.1 bug

[Parimi, Venkateshwara Rao]

>I still suggest that you install the Sun-recommended patches in
><ftp://sunsolve.sun.com/pub/patches/2.6_Recommended.tar.Z>.
>It will save your time and ours.

Sorry for not informing earlier, In fact I have all these patches installed.

>> What I want to know is, what would be the next check done by the
configure
>> script that causing Segmentation fault.

>You should be able to find that out by running "ksh -x ./configure"
>instead of plain "./configure".

I have tried it as well but nothing is reported except few additional
details in the beginning of the script like SHELL etc.

>> 1) src/wheel-gen.pl has #!/usr/bin/perl hard-coded(I have perl in
>> /usr/local/bin).
>>      Can this be auto detected?

>It could, but does it fix things to replace "#!/usr/bin/perl -w" with
>"#!/usr/bin/env perl"?  I don't know what the -w does; can that be done
>inside the perl script instead?

Changing it to "#!/usr/bin/env perl -w" should be fine.
"-w" option of perl produces nice diagnostic information, should there be
any problem, and it is always recommended to use.

>> 2) Basic test of du fails because of the test file is written with less
>> bytes than required (>64 bytes).

>I vaguely recall that this is fixed in coreutils 5.0.1; can you try that?
>ftp://alpha.gnu.org/gnu/coreutils/coreutils-5.0.1.tar.gz

I shall give a try then, thanks.

>> b) fail-2eperm of rm fails because of the error string expected is
different
>> >From the actual error string returned
>> parimiv > diff fail-2eperm fail-2eperm.fixed
>> 46c46
>> < rm: cannot remove `a/b': Operation not permitted
>> ---
>> > rm: cannot remove `a/b': Not owner

>This test-case bug still remains in 5.0.1, I think.  But it's not a
>bug in rm, as far as I can see.

If different flavors of OS has different error strings, opting for checking
the return code instead of comparing the return string would be better I
think. 

Thanks
-Venkatesh