Re: [PATCH] silent-many-generic: be stricter in silent rules pattern

[Stefano Lattarini]

On 02/28/2012 08:52 PM, Dave Goodell wrote:
> Your patch does not work for me.
>
Yep, that's because of this embarassing mistake:  I used
 '*CC '*)
where I should have (obviously) used:
  *'CC '*)

> Attached is a fixed version of your patch that matches the offending $CXX
> value correctly, handles spaces in compiler wrapper arguments,
>
Ouch, you're right, I had forgotten to properly escape some '"' when creating
the wrapper script :-( *blush*

> and fixes a typo.
> 
> If you have a better alternative to the conservative "case" statement that
> I used, please use it.  Between the recommendations in the autoconf manual
> and the recent thread on autoconf@ about ${1+"$@"}, I'm a bit unsure about
> the best way to proceed besides using the case block.
>
No, you patch is good enough as it stands (and, even more important, is
*tested*); I've applied it in your name, with the following slightly-edited
commit message:

    tests: fix spurious failure with Portland Compilers

    * tests/silent-many-generic.test (configure.ac): Be more careful
    in the workaround to avoid spurious failures with C++ compilers
    containing the string "CC" in their names; this avoids spurious
    failures with at least the Sun Studio C++ compiler (when named
    "sunCC") and the Portland Group C++ Compiler ("pgCC").
    Also handle arguments with whitespace correctly in the 'am--cxx'
    temporary wrapper.
    * THANKS: Update.

    Co-authored-by: Stefano Lattarini <address@hidden>
    Copyright-paperwork-exempt: yes

Thanks,
  Stefano