same, Do you think it's the hardware? I wonder if a register is getting overwritten with normal processing but gdb is somehow saving it.
---------- Original Message ----------- From: Grant Grundler <[EMAIL PROTECTED]> To: Jim Buttafuoco <[EMAIL PROTECTED]> Cc: [email protected] Sent: Tue, 15 Mar 2005 09:36:26 -0700 Subject: Re: Postgresql 8.0.1 > On Tue, Mar 15, 2005 at 11:15:54AM -0500, Jim Buttafuoco wrote: > > This is really strange. If I run postgresql under gdb and break in > > the float4div function then I get the correct answer. If I remove > > the break point then I get the wrong answer. > > I am thinking I need to do the kernel upgrade you suggested. > > Yeah - the kernel you had was *really* old. > I just want to rule out kernel FP-assist bugs. > > > My only question is which one. whould I pick the 2.6.11-pa3 32 bit on. > > For the 710, that's sounds like the right one. > > > What does the pa0 - pa3 mean? > > We increment the -paN for every CVS commit and reset N to zero after > merging the latest "release" tree from linus. That way it's easy to > track which changes someone picked up from the parisc-linux CVS tree. > > grant > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] ------- End of Original Message ------- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
