On Mon, 15 Mar 2004, Norman Gray wrote:
> Greetings, all
>
> [This is a response to a message that Tim sent me directly, which I
> responded to as if it had been sent to the list. I should pay
> attention: that way lies potential great embarrassment.]
>
> On Sun, 14 Mar 2004, Tim Jenness wrote:
>
> > [fiddling with things yesterday, I'd managed to tread on Tim's toes,
> though not badly as it turned out]
>
> Ah, did I manage to stomp on something? I thought I'd be safe enough,
> as long as I sent a message out promptly enough after the commit.
> What happened -- should I have sent a `hands off' message to stardev,
> or is there anything else that would be good practice here?
>
> > clear your next task after primdat is GKS :-) I've left my notes on the
>
> I thought that, after primdat, and now possibly GKS (the smiley worries
> me strangely), I should spend a little time checking that things do build
> on other platforms, such as Solaris, and sorting out alink to remove
> the duct-tape currently disfiguring it. Also that I could tinker with
> things like nbs, to see if there's a more idiomatic solution than the
> one you've produced. That, while the more straightforward libraries
> are being added by other folk.
>
> > It feels very lonely at the moment. I note that Mark is the only other
>
> Given the success that you and Mark have had, I think that it's
> reasonable to say that the build system is now conditionally usable,
> and that if I concentrate on the odder libraries for the moment,
> they'll build up into a set of working examples. So I think that we
> could, and indeed should, sustain a couple of folk working on the autoconfing.
>
> However Mark mentioned that he, also, had things to do this week that
> would prevent him working on the autoconfing much, if at all. Do we
> have another two volunteers? Step right up.
To clarify: I have done just GRP and ARD, selected because I thought
they would be quite easy. It took me a little while to get into it,
but following the instructions in SSN/78 I'd agree that it is
reasonably straightforward to add new packages once you get the idea,
at least ones that aren't too weird in the first place.
I'm not working further on it for now, because I want to get a new
visible feature (table joins) into TOPCAT before NAM; this is probably
just about doable in the two weeks between now and then. I judge this
to be a higher short-term priority than the autoconf stuff, though if
commanded otherwise I can comply.
Mark
--
Mark Taylor Starlink Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/
|