Celestia Forums

Chris,

I have finished coding and testing some major improvements. They are
ready for checking in...

Please, acknowledge my parallel email, as soon as you got it. I am
also short in free time as you know, and I never know whether my
emails reach you without a response...(It seems to be a problem associated with the DNS's we are using here). On the other hand, I have a daily email load of 10-25 without the slightest problems, except ns.shatters.net.
Everything is handled via a professionally configured mail server over here...

Here are the "highlights":

1) Gained a speed factor <= 10 (!) for small FOV's due to culling of /rendered
galaxies/, if they are outside the window (analogous to the culling in the
standard renderList). Notably, for the popular new, much enlarged list of
galaxies, galaxy rendering was a /major/, awkward slow-down at small fov up to
now.

2) Fixed the "click & display" of starnames (also in Windows) notably for small
fields of view. The bug was that 'pickTolerance' was assumed proportional to
the FOV, which gave vanishing tolerance at the smallest fields! I now put
pickTolerance = 3.0f/height, which reproduces the old sensitivity at FOV=45
deg. It works perfectly now even at the minimal field of 36".

3) Significantly tuned and incorporated the Automag scheme into the code, such
that switching between "fixedMag" and AutoMag by means of the GUI dialog
(Linux) or CTRL Y is consistent everywhere. The AutoMag scheme now is very
nice and /efficient/ for the newest 2M star data like 2.0C! Things like AutoMag
adaptation of faintestMag, while jumping with the middle mouse button between
45deg and the last fov, are also incorporated...

4) Some tuning for gtkmain dialogs under Linux (showCometTails etc).

Bye Fridger

PS: AutoMag on|off still needs to be included into the Windows GUI which I
cannot do...

Chris,

just got this back again....

============================================
**********************************************
** THIS IS A WARNING MESSAGE ONLY **
** YOU DO NOT NEED TO RESEND YOUR MESSAGE **
**********************************************

The original message was received at Sat, 21 Sep 2002 22:03:48 +0200 (MET DST)
from rem-schrempp.desy.de [131.169.241.64]

----- The following addresses had transient non-fatal errors -----
<claurel@ns.shatters.net>

----- Transcript of session follows -----
<claurel@ns.shatters.net>... Deferred: Name server: ns.shatters.net.: host name lookup failure
Warning: message still undelivered after 4 hours
Will keep trying until message is 5 days old

I think the problem is that I don't have reverse DNS working, and some mail servers are blocking shatters.net because of it . . .

--Chris

Chris,

I was about to check in my code this morning, when I discovered, you had checked in code 12 min's earlier, which needed a little reshuffling of mine;-).

Actually, your code did not compile (sprintf), since #include <cstdio> was missing in celestiacore.cpp. I have added it.

This is just to notify you, in case you again do not get my email via the sourgeforge-developer mailing list...

Bye Fridger

I did get your mail from the developers list . . .

Sorry about the sprintf problem . . . I was checking in code right before heading to bed, and was too lazy to compile on Linux first. Oops.

What do you think of the notifications when you change limiting magnitude, time rate, etc? Useful?

--Chris

chris wrote:I did get your mail from the developers list . . .

Sorry about the sprintf problem . . . I was checking in code right before heading to bed, and was too lazy to compile on Linux first. Oops.

What do you think of the notifications when you change limiting magnitude, time rate,
etc? Useful?

--Chris

I have a little bit split feelings, but basically I like the flashes
(colour and fontsize might still be optimized, perhaps).

I think, the flash messages are certainly useful, until we have toolbars that will permanently and clearly indicate the various toggle states of the important options.

On the negative side, I feel that (unexperienced) users might easily
get confused about the cumulative toggle states after pushing a number of keys triggering these few second flashes...(It happened to me
too;-)). What I liked very much, though, is the indication of the limiting Mags upon operation of the [,] key adjustments, after AutoMag has been switched off.

Some of the messages are also visible in other corners of the display (like "Follow", "Pause", "Time factor",...) and thus are redoubling to some extent.

Bye Fridger

Chris,

after my previous mail, it occured to me that your flash message would be /ideal/ for indicating in case of /AutoMag/ in a transient fashion the actual limiting Mag as function of the fov. The needed changes are absolutely trivial (deleting an 'if' statement;-))


Bye Fridger

Chris,

There do seem to be nameserver problems not related to the lack of a reverse lookup. My system is unable to resove either http://www.shatters.net or ennui.shatters.net right now. As a result, I've had to resort to using the numeric address in my Web browser.

For what it's worth traceroute shows substantial delays getting through your ISP, which probably are causing nameserver timeouts:


25 99 ms 99 ms 99 ms 10.gig0-0-0.sttl-agw1.sttl.uswest.net [206.81.192.248]
26 104 ms 117 ms 129 ms 216.160.70.11
27 304 ms 416 ms 642 ms 63.224.48.70
28 434 ms 526 ms 595 ms 63.224.48.65

selden wrote:Chris,

There do seem to be nameserver problems not related to the lack of a reverse lookup. My system is unable to resove either http://www.shatters.net or ennui.shatters.net right now. As a result, I've had to resort to using the numeric address in my Web browser.

For what it's worth traceroute shows substantial delays getting through your ISP, which probably are causing nameserver timeouts:


25 99 ms 99 ms 99 ms 10.gig0-0-0.sttl-agw1.sttl.uswest.net [206.81.192.248]
26 104 ms 117 ms 129 ms 216.160.70.11
27 304 ms 416 ms 642 ms 63.224.48.70
28 434 ms 526 ms 595 ms 63.224.48.65

Let me add that ennui.shatters.net is virtually /never/ resolved by our nameservers and http://www.shatters.net only works since the number is kept in the local nameserver cache...

Bye Fridger