EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  <20042005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  <20042005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: RE: calc and calcout records in EPICS 3.14.6
From: "Thompson, David H." <[email protected]>
To: EPICS Tech-Talk <[email protected]>
Date: Wed, 17 Nov 2004 08:40:16 -0500
Compiler bugs are one of the scariest aspects of software development
for me, I guess because I understand the least about compilers and how
to test them.

So:

Is there anyone out there that has an old version of Linux that has a
version of gcc between 2.96 and now? I would like to get reports on
which versions work and which versions exhibit the error.  That way we
can know when the fix was applied in the gcc source leading to a patch
for our 2.96.

It would also be useful to know if the X86 versions have this bug as
well.
Be sure and include -O2 in the flags when compiling the test.

double stack[]={0.0,1.0};
double * pStack=&stack[0];

int i=(*pStack==*(pStack+1));

If i is 1 you have the bug.


Ernest and I are putting together a web page on our server,
contributions are welcome at this address. 

-----Original Message-----
From: Mark Rivers [mailto:[email protected]] 
Sent: Wednesday, November 17, 2004 7:52 AM
To: Ralph Lange; Andrew Johnson
Cc: EPICS Tech-Talk
Subject: RE: calc and calcout records in EPICS 3.14.6

Andrew and Ralph,

We've been using the SNS versions of PPC and 68K compilers without being
aware of the bugs that Andrew mentioned.  Are these documented somewhere
on the Web?

Thanks,
Mark

>   > There are also problems with a 68K C++ compiler built from these
> sources;
>   > in that case the issue is related to passing bool variables as
> arguments
>   > to functions or member functions - again I haven't chased the
detail.
> 
> Plus a problem with NULL pointers being treated wrong in expressions
> (after setting a pointer to NULL, comparing it to NULL in an
expression
> would yield a FALSE).
> 
> Afaik this problem is appeared only in the 68k version of the original
> Tornado 2.0.2 compiler and was fixed with a later patch. Anyway:
> compilers built from the sources available at SNS are not showing this
> bug. (That's why we rebuilt even our HP-UX based compiler from the SNS
> sources.)





Navigate by Date:
Prev: RE: calc and calcout records in EPICS 3.14.6 Mark Rivers
Next: EPICS meeting at KEK Dalesio, Leo
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  <20042005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: calc and calcout records in EPICS 3.14.6 Andrew Johnson
Next: RE: calc and calcout records in EPICS 3.14.6 Thompson, David H.
Index: 1994  1995  1996  1997  1998  1999  2000  2001  2002  2003  <20042005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 10 Aug 2010 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·