At Thu, 27 Apr 2000 13:40:56 -0400 (EDT),
Andrew Park <apark@xxxxxxxxxxxxxxx> wrote:
> Man ... I've been trying to reply for last half an hour or so, but
> I keep getting more replies ... :)
> I appreciate the following people's feed back.
> Kip Macy
> Steve Lord
> Andi Kleen
> who graciously commented on my question. Now it's my turn. :)
> I'm not asking for backward compatibility I'm asking whether XFS will
> be current(ward) compatible (is that even a word?) :) when it comes
> down to it. If XFS is in ready-state and official linux kernel compiler
> is gcc 2.7.2.x still, would XFS be adjusted to compile under it then?
> or as some companies might exhibit "Too Bad! You do it our way or no way!"
> is the type of statement I'll be hearing from this mailing list?
> (just wondering ... no offense intended)
I don't know for sure; again it isn't something we have investigated..
Unless there is some techincal issue that really really can not be
solved by means other than upgrading gcc, XFS will comply with
whatever version of gcc is recommended.
> My apologies for misunderstanding the status of XFS, but from the Linux
> University sponsored by SGI that I attended in Toronto it sounded quite
> like as if it is ready for experiment in semi-production
That's up to the individual to determine their own risk factor.
I would just use the term experimental right now... semi-production...
hmm not quite yet, but it's getting closer every day.
> Knowing the XFS status now I understand (not quite agreeing though) why
> compiling under gcc2.7.2x is not much of issue.
> [cont. below]
> On 28 Apr 2000 suzukis@xxxxxxxxxxxxxxxxxxxxxx wrote:
> > >I (IMHO) think it is rather stupid if XFS wouldn't compile
> > >with the official linux kernel compiler.
> > Of course, I agree enabling linux-xfs to be compiled
> > with gcc-18.104.22.168 is expected - but I don't think yet
> > it's not time for the linux-xfs core developpers to
> > check the back compatibilities. Today linux-xfs is
> > in bleeding-edge status, Possibly still there might
> > be bugs due to the implementation itself (SORRY!),
> > realized in any versions of gcc. For the core developpers
> > of SGI, now it's time to fix such bugs.
> > So, please don't ask the gcc-version-independency yet.
> > If you wish NOW - you (and me :-)) should do it.
> It may take me a while, but I'll see what I can do.
> > But, keeping (at least) 1 environment similar to the
> > core developpers is not bad to find a bug, I think.
> > Don't you think so?
> Agreed, and thanks again for the reply.
> Andrew Park
> CDFlab Systems Administrator www.cdf.utoronto.ca |
> Team BlueShirt Developer www.blueshirt.org |