Re: v4.3.0.4 not working on SSE1 CPUs anymore

On Fri, Aug 8, 2014 at 3:25 PM, Wols Lists <antlists@youngman.org.uk> wrote:
> On 08/08/14 12:55, Norbert Thiebaud wrote:
> > What compelling reason is there to run the lastest version of a
> > software on such old architecture ? and even more to the point, what
> > compelling reason is there to slow down everybody else for that.
>
> Maybe there's still a reasonable number of such systems about? It was a
> pretty decent XP system. I only retired my K7 thunderbird about a year
> ago, because I was given a newer "ancient" system to replace it.

I'm not sure how many systems of this era are still in use -- anyone
know where we could get some data? Maybe some other FOSS project like
Mozilla has some data?

Maniaxx -- Did this problem start with LO 4.3? (I believe that we
switched to Visual Studio 2012 by LO 4.0 or 4.1)

> (Or is it a lot of effort to maintain a SSE build for old systems on the
> website? I'm not offering ...)

Building on Windows is a bit more painful than building on GNU/Linux,
however if someone really wanted a one-off build, I guess that they
could follow the instructions
https://wiki.documentfoundation.org/Development/Windows_Build_Dependencies

> I know we can't keep supporting old stuff for ever, but I've got 80s-era
> software I would still like to run,

In general, 80s-era software (or older) should either be updated to
run on modern machines or be run in a VM.

> and I know of people who can't
> afford to replace aging computers...

I'm quite sympathetic to those who can't afford to upgrade their
existing hardware. If someone can get some hard data that there's a
non-trivial percentage of our users (or potential users) who are not
able to use our software, and who *could* run our software if we just
changed a compile-time flag, then I think we should give the matter
further consideration.

Best,
--R