On 5/7/07, Rana Dasgupta <[EMAIL PROTECTED]> wrote:
Hi Stephan,
   Thanks. Any idea why we choose not to do the same with x64?

Sometime ago we agreed that M1 should be aimed to x86. Also in [2] it
was proposed to publish x86 snapshots only. So milestone candidates
for Linux x86 were built with both libstd++ libraries.

However Linux x86_64 with libstd++.so.5 snapshot was included to M1
too. I don't know why (may be Tim found that it is worth to include it
- I don't have any objection for this).

-Stepan.

Rana

On 5/6/07, Stepan Mishura <[EMAIL PROTECTED]> wrote:
> On 5/4/07, Rana Dasgupta wrote:
> > Hi,
> >  We have published two different snapshots for 32 bit Linux builds
> > with libstdc++.so.5 and so.6, but only so.6 snapshots with 64 bit
> > builds. I understand that this was a problem for a couple of users
> > using slightly older versions of gcc. so.5 and so.6 is more a toolset
> > and not a platform issue, and unfortunately 64 bit does not seem to
> > imply new toolsets etc.
>
> Hi Rana,
>
> Linux snapshots are built with libstd++.so.5. In [1][2] threads
> building with libstd++.so.6 was discussed and for M1 we published x86
> snapshots with both dependencies.
> In [3] Tim suggested to do only the libstdc++v6 builds but we haven't
> discussed it yet.
>
> [1] http://mail-archives.apache.org/mod_mbox/harmony-dev/200704.mbox/[EMAIL 
PROTECTED]
>
> [2]http://mail-archives.apache.org/mod_mbox/harmony-dev/200704.mbox/[EMAIL 
PROTECTED]
>
> [3]http://mail-archives.apache.org/mod_mbox/harmony-dev/200704.mbox/[EMAIL 
PROTECTED]
>
> Thanks,
> Stepan.
>
> >  It's quite likely that we have discussed this already on other
> > threads, and I was not following along.
> >
> > Thanks,
> > Rana
> >
> >
> > On 4/29/07, Stepan Mishura <[EMAIL PROTECTED]> wrote:
> > > On 4/30/07, Stepan Mishura <[EMAIL PROTECTED]> wrote:
> > > > On 4/29/07, Tim Ellison wrote:
> > > > > Stepan Mishura wrote:
> > > > > > On 4/29/07, Tim Ellison wrote:
> > > > > >> It looks like a number of people have been testing the stable build
> > > > > >> candidate (thanks to everyone!) and so far it seems that there are 
no
> > > > > >> show stoppers for declaring this a good stable build.  It's our 
best so
> > > > > >> far, albeit with known problems etc.
> > > > > >>
> > > > > >> Unless anything serious come to light by tomorrow, I suggest we 
declare
> > > > > >> this our M1, and reopen the code for ongoing development.
> > > > > >>
> > > > > >
> > > > > > Does it make sense to rebuilding M1 with debug info in class files
> > > > > > included?
> > > > >
> > > > > Yes, if you could upload a new set of builds at the current repository
> > > > > state that would be great.  That would include the debug info now, and
> > > > > no additional functional change.  We can look at refining the JRE vs 
JDK
> > > > > info after M1.
> > > > >
> > > >
> > > > OK. I'm going to create snapshots for r533500.
> > > >
> > >
> > > Done
> > >
> > > > -Stepan.

Reply via email to