... I think you misunderstand how 'set origin' is supposed to work. It sets the origin of the graph, in paper coordinates, which means that to move the graph...
3361
Petr Mikulik
mikulik@...
Mar 7, 2000 11:17 am
... Positive offsets work as expected. I expect that negative offsets should work as well and move the graph to the left -- i.e., the behaviour is not ...
3362
Hans-Bernhard Broeker
broeker@...
Mar 7, 2000 11:28 am
... Well, thinking about that again, I suspect they actually don't. I'd expect to crash gnuplot on some of the bitmap image terminals ('gif', 'png' or 'pbm')...
3363
Johannes Zellner
johannes@...
Mar 7, 2000 11:39 am
... I guess Petr wants to move the graph to the *left*. And that does not work! I'd also call this asymmetry (allow shifting the graph off the right but not...
3364
Hans-Bernhard Broeker
broeker@...
Mar 7, 2000 12:11 pm
... Both behaviours are a bug, actually. Gnuplot terminal drivers are not required to behave well if they receive commands to draw outside their reported paper...
3365
Jeff Spirko
spirko@...
Mar 7, 2000 1:49 pm
... +++ graphics.c Tue Mar 7 08:37:26 2000 @@ -4132,6 +4132,7 @@ int sign = 1; double l10; int power; + double mtmp; /* Tmp storage for mantissa, see...
3366
Hans-Bernhard Broeker
broeker@...
Mar 7, 2000 2:51 pm
... I doubt it does... The real problem is that the behaviour on output depends on the number of digits and the precision the user asked for in the '%s'...
3367
Jeff Spirko
spirko@...
Mar 7, 2000 3:03 pm
... You're right, my bad. The test case was a log scale, where the mantissa actually does come out to 10.0. Also, I didn't realize that %t would *ever* round...
3368
Hans-Bernhard Broeker
broeker@...
Mar 7, 2000 3:25 pm
... That's the good point about your patch: it gets this part of the task done in the correct place. mant_exp() is the place to do that. But it'll need another...
3369
Hans-Bernhard Broeker
broeker@...
Mar 7, 2000 3:26 pm
... Ooops, seems I overlooked that %T output then depends on what the format spec for %t was.... not even to mention what would happen if some crazy user did ...
3370
Jeff Spirko
spirko@...
Mar 7, 2000 4:03 pm
... Or, a more reasonable " %T\n%4.1t * 10" (maybe on a text terminal), or even printed the number twice with different precisions. ... I was just...
3371
Ilya Zakharevich
ilya@...
Mar 7, 2000 10:22 pm
... This should be specified by a flag in the terminal structure. Why? How would the caller of the driver know that a left-anchored proportional-width-font...
3372
Chris Atenasio
chris@...
Mar 8, 2000 12:35 am
sin(x)**2 + cos(x)**2 Doesn't plot as a straight line... Anyone know why? Apologies if this is the wrong place to post. - Chris ... Chris Atenasio...
3373
Lars Hecking
lhecking@...
Mar 8, 2000 1:28 am
... "Straight line" is relative. Check the actual y range of the plot, then switch autoscale off for the y axis and replot. The sampling rate may have a say in...
... A single flag alone wouldn't help a bit. The actual problem of clipping is way too complicated to be describable by a single flag. The important part would...
3376
Jens Schneider
Jens.Schneider@...
Mar 8, 2000 1:24 pm
Hi folks, I just want to report some problems with pm3d and postscript: it seems to produce illegal postscript-code, at least my ghostscript doesn't like it. ...
3377
Petr Mikulik
mikulik@...
Mar 8, 2000 2:34 pm
... Really "splot sin(x)=y**2"? If you mean set term post color; set out "test.ps" set pm3d set pm3d solid splot sin(x)*y**2 then it works. ... Petr Mikulik,...
3378
Johannes Zellner
johannes@...
Mar 8, 2000 2:42 pm
Jens, and: maybe you can try if it produces `valid' ps w/o pm3d ? gnuplot> unset pm3d gnuplot> set term post color gnuplot> set out 'test.ps' gnuplot> plot x*x...
3379
Johannes Zellner
johannes@...
Mar 8, 2000 2:43 pm
... It works perfectly here with: gv 3.5.8 gs Aladdin Ghostscript 5.50 (1998-9-16) I just installed the very same gs version you have (guess: you're on a...
3380
Jens Schneider
Jens.Schneider@...
Mar 8, 2000 3:22 pm
Hm. I tried the example from help pm3d set pm3d at s hidden3d 100 set style line 100 lt 5 lw 0.5 unset hidden3d unset surf splot x*x+y*y and then set post...
3381
Ilya Zakharevich
ilya@...
Mar 8, 2000 4:35 pm
... Nope. You did not notice: *this is (except the flag) already implemented for PARI*. All you need is to clip 3 primitives: points, lines, and text. All...
3382
Jens Schneider
Jens.Schneider@...
Mar 8, 2000 4:44 pm
Still, the following script generates a wrong ps-file. I attach it to this mail. set pm3d at s hidden3d 100 set style line 100 lt 5 lw 0.5 unset hidden3d unset...
3383
Johannes Zellner
johannes@...
Mar 8, 2000 5:34 pm
... [...] yes. I can reproduce it. You don't have to send examples any more :-) We'll try to fix that. -- Johannes [[[[ unsubscribe from info-gnuplot-beta via...
3384
Petr Mikulik
mikulik@...
Mar 8, 2000 7:06 pm
... I've just submitted the bugfix to Johannes' CVS server. Technical details: now, for "set pm3d at s hidden3d 100": color.c resets "PS_relative_ok" after...
3385
Alexander Mai
st002279@...
Mar 8, 2000 8:55 pm
... gs 6.00 fails, too. My HP printer doesn't like it either. BTW, I think there was a request to send this personally. In general idea attachments >>10KB to a...
3386
Johannes Zellner
johannes@...
Mar 8, 2000 10:17 pm
On Wed, 8 Mar 2000, Alexander Mai wrote: [...] ... you should probably buy another printer then :-) Petr Mikulik has already found the problem, thanks Petr....
3387
Alexander Mai
st002279@...
Mar 9, 2000 5:05 pm
Just tried to build the pm3d-version. Out of the box fails on DU/alpha for various reaons: - someone cluttered gcc stuff in the src/Makefile :-(( - C++...
3388
Alexander Mai
st002279@...
Mar 9, 2000 5:06 pm
... [...] TERM_PUBLIC int X11_waitforinput() { char c; #if defined(USE_MOUSE) && !defined(OS2) fd_set fds; static struct gp_event_t ge; static int l=0; int n; ...
