When will the next gcc come out

When will the next gcc come out

[Huang Bambo]

The current gcc( 4.3.4 ) has some bug with O3 option. it will
optimized out some needed code in some case.
I meet this bug when I compile ACE library.
If I use O2 option, everything goes fine.
So, maybe many people may waiting for the new version. :>
Thanks !

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: When will the next gcc come out

[Andy Koppe]

On 23 July 2010 09:24, Huang Bambo wrote:
> The current gcc( 4.3.4 ) has some bug with O3 option. it will
> optimized out some needed code in some case.
> I meet this bug when I compile ACE library.
> If I use O2 option, everything goes fine.

Have you got any actual evidence for this being a bug in gcc rather
than in your program? It's quite common that subtle program bugs get
exposed differently by different optimisations. This might also be
relevant:

http://en.wikipedia.org/wiki/Aliasing_(computing)#Conflicts_with_optimization

Andy

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: When will the next gcc come out

[Huang Bambo]

2010/7/23 Andy Koppe <andy.koppe@gmail.com>:
> On 23 July 2010 09:24, Huang Bambo wrote:
>> The current gcc( 4.3.4 ) has some bug with O3 option. it will
>> optimized out some needed code in some case.
>> I meet this bug when I compile ACE library.
>> If I use O2 option, everything goes fine.
>
> Have you got any actual evidence for this being a bug in gcc rather
> than in your program? It's quite common that subtle program bugs get
> exposed differently by different optimisations. This might also be
> relevant:

It's a very simple test program:

#include <ace/Log_Msg.h>

int ACE_TMAIN( int argc, ACE_TCHAR **argv )
{
        ACE_DEBUG(( LM_DEBUG, ACE_TEXT("%D\n") ));
        return 0;
}

Gcc optimized out some code like
*fp++ = *format_string ++;

When I gdb to that line and "print fp", gdb reported that fp is optized out...

If I compile ACE library with -O2, everything become OK.

I will check if this will happen in other platform when I have time.

> http://en.wikipedia.org/wiki/Aliasing_(computing)#Conflicts_with_optimization
>
> Andy
>
> --
> Problem reports:       http://cygwin.com/problems.html
> FAQ:                   http://cygwin.com/faq/
> Documentation:         http://cygwin.com/docs.html
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
>
>

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: When will the next gcc come out

[Dave Korn]

On 23/07/2010 09:24, Huang Bambo wrote:
> The current gcc( 4.3.4 ) has some bug with O3 option. it will
> optimized out some needed code in some case.
> I meet this bug when I compile ACE library.
> If I use O2 option, everything goes fine.
> So, maybe many people may waiting for the new version. :>

  4.5.0-1 has completed building and testing, I need to finish up the
packaging (setup.hints and README/release announcement) and I'll upload it
overnight tonight.

  It will be a test release to start with, owing to regressions in Ada.  After
it's done, I'll be starting on getting 4.5.1-1 ready, hopefully with fixed Ada.

    cheers,
      DaveK


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: When will the next gcc come out

[Dave Korn]

On 23/07/2010 10:02, Huang Bambo wrote:

> 
> Gcc optimized out some code like
> *fp++ = *format_string ++;
> 
> When I gdb to that line and "print fp", gdb reported that fp is optized out...

  You need to check in more detail than that.  GCC may have optimised away the
fp variable, but kept the functionality implied by the line.  You'd need to
study the assembly code to see if it really did what the C code said it should
do or not.

    cheers,
      DaveK

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple