gnuplot-beta — Developers and beta testers of new gnuplot code

Re: Should we keep this list subscriber-only?

[Petr M. <mi...@ph...>]

> I'd rather keep the list subscriber-only, and if possible to
> auto-generate a reply to non-subscribed messages that
> says something like:

I like the way to send a message to a group without sending two emails
before to subscribe there, and then again to unsubscribe from there.
It may bother users so much that they won't send any message at all.
(My own experience :-()

--
PM

Re: Windows EMF

[BBands <bb...@ya...>]

> 
> I have not heard back from anyone whether Word2000
> running
> under Windows does or does not have the same
> problem.
> 
> In the hope that someone reading this has a "real"
> W2K system
> to test this on, I attach the output from gnuplot's
> "test" command
> using both the old and the new (patched) emf driver.
>  On my
> system the new version places all the text
> characters on top
> of each other.  Do you see the same thing?

On a fully patched win 2k system--no emf hotfix--I can
see the new, but not the old. The pattern-fill bars
are empty, the line-width displays are all the same
width and the character-width display fills half the
box.

    --jab

=====
John Bollinger, CFA, CMT
www.BollingerBands.com

If you advance far enough, you arrive at the beginning.


	
		
__________________________________
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/ 

Re: loose plan (was Re: K&R C)

[Ethan M. <merritt@u.washington.edu>]

On Thursday 03 June 2004 10:26 am, Daniel J Sebald wrote:
> >   fillcurves	   use fill style for filledcurves (cgm gd pdf post svg)
>
> In the same category of a few of the above patches, someone (don't
> remember who) had done some work to integrate gray scale filled
> curve-like plots into PostScript.  That looked pretty nice, but I wonder
> if PostScript might be the only driver to easily support such a feature.

If I understand your description correctly, that is supported by the
above patch:

plot <whatever> using filledcurve linetype -1 fillstyle solid <greylevel>

But maybe you are thinking of the request for shaded 
areas in the background?  That should be possible to implement
for any driver that supports filled rectangles, but I don't recall
any suggestions at all as to what sort of syntax would be used
to request such a plot.


-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: loose plan (was Re: K&R C)

[Daniel J S. <dan...@ie...>]

Ethan A Merritt wrote:

>I have five patchsets that I consider stable and ready to go into
>CVS whenever continued development is split off from 4.0 bugfixing.
>
>   fillbounds	   adds "above" and "below" bound-clipping to filledcurves
>   datastrings	   'plot with labels', tic and plot labels from datafile, etc 
>   histograms	   plot with histograms
>   fillcurves	   use fill style for filledcurves (cgm gd pdf post svg x11)
>   XEvent  	   reorganizes X event loop to speed up refresh
>
>The 'set mouse key' patch works, but I'm still cleaning it up and it
>sure would be nice to have string variables so that I could return
>the keypress as a character rather than as a key code.
>  
>

In the same category of a few of the above patches, someone (don't 
remember who) had done some work to integrate gray scale filled 
curve-like plots into PostScript.  That looked pretty nice, but I wonder 
if PostScript might be the only driver to easily support such a feature.

Dan

Re: Windows EMF

[Ethan M. <merritt@u.washington.edu>]

On Thursday 03 June 2004 08:42 am, Hans-Bernhard Broeker wrote:
> On Thu, 3 Jun 2004, BBands wrote:
> > A Windows security patch to fix an EMF buffer overflow
> > vulnerability, KB 835732, prevents many EMFs from displaying.
>
> We're aware of that.  A patch for gnuplot is on our patch tracker, but
> according to Ethan, it still doesn't work.

To be precise, it doesn't work for me using Word2000 running
under linux.  Since this is far from the most common thing to do,
I don't know whether it really matters.  But the old (un-patched)
driver worked fine for me.

I have not heard back from anyone whether Word2000 running
under Windows does or does not have the same problem.

In the hope that someone reading this has a "real" W2K system
to test this on, I attach the output from gnuplot's "test" command
using both the old and the new (patched) emf driver.  On my
system the new version places all the text characters on top
of each other.  Do you see the same thing?

-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: timefmt does not have provision for milliseconds

[Daniel J S. <dan...@ie...>]

Vijayaraj AK wrote:

>Well, the data need not be generated every
>millisecond.  And yeah as you point out, it is not
>logged for hours. I want to plot lets say, 20-30min
>duration log of data sampled at 20-30ms interval. 
>  
>

Oh, well that is not so much ms resolution.  If you are displaying 20-30 
min at 20-30 ms sampling rate, that's 60,000 data points, quite a lot of 
data for one plot.  In other words, it doesn't match the resolution of 
the plotting device.  That is, if there is some event on the scale of 
20-100 ms, it may be the case that gnuplot will detect it because a 
min/max strategy is used.  But in terms of a display of 20-30 min, you 
wouldn't be able to tell if that was a 100 ms event or a 10 s event. 
 Some form of preprocessing and downsampling might be in order, although 
perhaps not essential.

My interpretation of a generalized time format might be 
year:month:day:hour:min:second:millisecond with an option for indicating 
what units the data is in, i.e., year, day, hour, etc.  And then the 
scale would adjust to display the appropriate units.  However, that's a 
lot of work.

Dan

Re: Should we keep this list subscriber-only?

[Hans-Bernhard B. <br...@ph...>]

On Thu, 3 Jun 2004, Lars Hecking wrote:

> 
> > Before changing this, I'ld like a poll.  So: if anybody has strong
> > opinions in either direction, please voice them soon.
>  
>  If people who's posts are held receive a notification of the fact,
>  I'd like to keep things the way they are - 15.000 spams/month are
>  no fun.

So far, they don't get notification while it's on hold --- only if the
admin decides to reject one, they'll receive a notice stating the reason.  
Spams and worms are discarded, instead, i.e. they get no response at all.


-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: Should we keep this list subscriber-only?

[Ethan M. <merritt@u.washington.edu>]

On Thursday 03 June 2004 09:18 am, Hans-Bernhard Broeker wrote:
>
> Before changing this, I'ld like a poll.  So: if anybody has strong
> opinions in either direction, please voice them soon.

I'd rather keep the list subscriber-only, and if possible to 
auto-generate a reply to non-subscribed messages that
says something like:

This is a subscriber-only mailing list for discussion of 
gnuplot development.  

If you have a specific question, please post it to 
usenet group comp.graphics.apps.gnuplot.

If you want to report a bug or file a feature request, please use
the web-submission forms at http://sourceforge.net/projects/gnuplot

If you would like to participate in on-going discussion 
about gnuplot development, subscribe yourself to this
list by {.... instructions}

-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: Should we keep this list subscriber-only?

[Petr M. <mi...@ph...>]

> you may not have been aware of this, but the new gnuplot-beta mailing list
> at sourceforge restricts the right to post mails to its subscribers to the
> subscribers themselves.  This may or may not be a good thing.

I would prefer your suggestion -- the administrator could redirect mails of
non-subscribers to the newsgroup / gnuplot users lists, instead of replying
to them on this list.

People got accustomed to ask questions at this list as they have used
gnuplot beta's for so many years...

--
PM

Re: Should we keep this list subscriber-only?

[Lars H. <lhe...@us...>]

> Before changing this, I'ld like a poll.  So: if anybody has strong
> opinions in either direction, please voice them soon.
 
 If people who's posts are held receive a notification of the fact,
 I'd like to keep things the way they are - 15.000 spams/month are
 no fun.

 If it helps, you can add my SF address to the moderators list and we
 can share the load. I'm already moderating two lists at gnupg.org, and
 it would be little additional load.

Re: loose plan (was Re: K&R C)

[Petr M. <mi...@ph...>]

> I have five patchsets that I consider stable and ready to go into
> CVS whenever continued development is split off from 4.0 bugfixing.
>
>    fillbounds	   adds "above" and "below" bound-clipping to filledcurves
>    datastrings	   'plot with labels', tic and plot labels from datafile, etc
>    histograms	   plot with histograms
>    fillcurves	   use fill style for filledcurves (cgm gd pdf post svg x11)
>    XEvent  	   reorganizes X event loop to speed up refresh

OK with me

> The 'set mouse key' patch works, but I'm still cleaning it up and it
> sure would be nice to have string variables so that I could return
> the keypress as a character rather than as a key code.

If there are string variables, then both the keypress code (an integer
number) and string representation of the key will be returned.


Some other "cleaning" of interactive terminal APIs could be done by changing
    "term->set_cursor(int,int,int)"
to
    "term->set_interactive_terminal_function(int,int,int [,int?], string)"

to be reused for several purposes like "unset your default hotkeys", "change
menu", "do raise", etc.
  Just I don't have time to do this change, right now.

---
PM

Should we keep this list subscriber-only?

[Hans-Bernhard B. <br...@ph...>]

Hello, everyone,

you may not have been aware of this, but the new gnuplot-beta mailing list
at sourceforge restricts the right to post mails to its subscribers to the
subscribers themselves.  This may or may not be a good thing.

As it is right now, submissions by non-members will be held for approval
by an admin (i.e: myself), which may delay their processing by a day or
more, and thus encumber the flow of discussion.  I've just cut a path
through an absolutely enormous back-log of such mailing list
administration requests, over 99 percent of which was junk mail to the
dartmouth list which ended up in the admin interface at SF.net (thanks to
Dave Denholm and the guys at Dartmouth for setting up a better solution to
this, which will be used from now on).  The flurry of belated mails you're
seeing today are the remaining less than 1 percent as accumulated in about
a month (plus replies to those).

OTOH, if I open up the list for posting by everybody, that means all
that's standing guard between your inbox and the unwashed masses of junk
mailers out there waiting to flood it is the automatic spam filtering
system used by SourceForge.  That's quite a successful one, granted, but
still not 100% leakproof.

Before changing this, I'ld like a poll.  So: if anybody has strong
opinions in either direction, please voice them soon.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: Errors in Batch mode

[Petr M. <mi...@ph...>]

> Apologies if this in an FAQ, but I'm in the process of integrating version 4
> into our gnuplot-enabled web applications and have the same question I had
> back in 2000 (on ver 3.7).
>
> When running on MS Windows, is there any mechanism whereby the errors that
> would be displayed in interactive mode (with LOAD), can be stored to a file
> when running in batch mode?

If it is for web application, i.e. for non-interactive work only, then use
the Cygwin/X11 binary -- that's the console application. You may even
compile your own binary (./configure; make procedure) disabling all
terminals except that one you are drawing to.

In summary: run gnuplot as you are use to run it on unix.

---
PM

Re: timefmt does not have provision for milliseconds

[Daniel J S. <dan...@ie...>]

Hans-Bernhard Broeker wrote:

>On Sat, 8 May 2004, Vijayaraj AK wrote:
>
>  
>
>>I was using GNU for plotting function and I wanted to
>>plot time in hour:min:sec:millisecond format. I
>>searched extensively for this kind of format. If this
>>is not defined it would be incredibly useful to have
>>such a format in future release.
>>    
>>
>
>I suspect it's nowhere near as useful as you think it is.  It's
>practically impossible to actually need both millisecond resolution and an
>entire hour of reach on a single axis --- no output device you're likely
>to use will be able to resolve 3 million positions along a single axis.
>

I'm guessing Vijayaraj basically meant that he wants millisecond 
resolution.  (Even still, if not I'm guessing there are a lot of 
applications that may be synchronized to a clock and watch for short 
transient events and time-stamp the event.  Hence, may want the 
hour/min/sec information along for the ride.)  In any case, if 
millisecond resolution is desired, this is a base 10 unit (as opposed to 
"base 60").  Normal plotting should suffice so long as you don't go 
across a one minute boundary.

Dan

Re: timefmt does not have provision for milliseconds

[Hans-Bernhard B. <br...@ph...>]

On Thu, 3 Jun 2004, Vijayaraj AK wrote:

> Well, the data need not be generated every
> millisecond.  And yeah as you point out, it is not
> logged for hours. I want to plot lets say, 20-30min
> duration log of data sampled at 20-30ms interval. 

Well, that's still an axis range of 20*60/0.02 = 60000 samples.  Do you
have an output device with sixty thousand pixels along the axis?  I doubt
it.

So I'ld suggest you just use extended 'using' specs and a format string 
for this:

	plot 'datafile' u ($1*3600+$2*60+$3):4 '%lf:%lf:%lf %lf'

It might even work ;-)

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: mono png

[Hans-Bernhard B. <br...@ph...>]

On Wed, 12 May 2004, Dan Jacobson wrote:

> echo "set terminal png;plot '/tmp/t'"|gnuplot -mono
> does not make a mono png. Oh, it is an x option... OK, at least make
> an error message.

An Error message about what? There's nothing really wrong about this
usage.

> You ought to put cross references in the Label, Title, etc. Info
> pages, to "Key", as I searched for hours before finally figuring out
> how to turn the filename off in the output.

I don't quite see how changing the main documentation in this way would
really help --- it'ld just bloat everything, for the benefit of what
I believe are rare cases.

Such questions are best left to the FAQ to answer, and they are.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: emf.trm update

[Allin C. <cot...@wf...>]

On Wed, 26 May 2004, Stephane Barbaray wrote:

> Ethan Merritt wrote:
>
> >The new version does not work for me (or rather, for Word2000).
> >
> I don't have Word2000... can someone confirm the problem? Working OK for
> me in Word97 and Word2002

Stephane's new version works for me too, on win98 and XP.

-- 
Allin Cottrell
Department of Economics
Wake Forest University, NC

Re: timefmt does not have provision for milliseconds

[Vijayaraj AK <vij...@ya...>]

Well, the data need not be generated every
millisecond.  And yeah as you point out, it is not
logged for hours. I want to plot lets say, 20-30min
duration log of data sampled at 20-30ms interval. 

To make the plot look good and ease of reference, hour
is also logged and displayed, though it might not be
useful. 

As for the usefulness, I work for Qualcomm and we
collect extensive amount of data and they are all in
milli-second granularity. 

As for now, I just convert the timestamp into a number
"seconds.milliseconds" and plot data vs xxx.000
number. It serves my purpose. It would just make
plotting easier if a format was created that takes
care of this. 

Thanks for you attention!
Vijay.

--- Hans-Bernhard Broeker
<br...@ph...> wrote:
> On Sat, 8 May 2004, Vijayaraj AK wrote:
> 
> > I was using GNU for plotting function and I wanted
> to
> > plot time in hour:min:sec:millisecond format. I
> > searched extensively for this kind of format. If
> this
> > is not defined it would be incredibly useful to
> have
> > such a format in future release.
> 
> I suspect it's nowhere near as useful as you think
> it is.  It's
> practically impossible to actually need both
> millisecond resolution and an
> entire hour of reach on a single axis --- no output
> device you're likely
> to use will be able to resolve 3 million positions
> along a single axis.
> 
> -- 
> Hans-Bernhard Broeker
> (br...@ph...)
> Even if all the snow were burnt, ashes would remain.
> 
> 


=====
Vijayaraj A.K.          
#304, Creekside Apartments,            \\\|///     o00o   
505, 27th WAY,                       \\  - -  //   (  )  
Boulder, CO 80305,                   (   0 0   )  (  )   
USA                             ***oOOo--(_)--*******  
Phone : (R) 303 499 1818
        (O) 303 247 5061                Hello!!


	
		
__________________________________
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/ 

Re: Plotting with linespoints

[Hans-Bernhard B. <br...@ph...>]

On Wed, 12 May 2004, [utf-8] S=C3=B8ren Madsen wrote:

> Is it possible to make a graph from a datafile with "linespoints" w=
here
> the line is drawn for every data-point and the points are drawn for
> every nth datapoint? If not, then this should be implemented... :D

No, because it's not strictly necessary.  You can just plot the file
twice, instead --- once with lines, ones with points.

Creeping featurism is like cholera, you have to stop it before it tur=
ns
epidemic.

--=20
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: Windows EMF

[Hans-Bernhard B. <br...@ph...>]

On Thu, 3 Jun 2004, BBands wrote:

> Things have changed for EMFs in Windows. A Windows
> security patch to fix an EMF buffer overflow
> vulnerability, KB 835732, prevents many EMFs from
> displaying. 

We're aware of that.  A patch for gnuplot is on our patch tracker, but
according to Ethan, it still doesn't work.

I.e. gnuplot-generated EMFs are incorrect, so the MS fix is somewhat 
justified to block them.  But we haven't managed to fully fix them, yet.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: timefmt does not have provision for milliseconds

[Hans-Bernhard B. <br...@ph...>]

On Sat, 8 May 2004, Vijayaraj AK wrote:

> I was using GNU for plotting function and I wanted to
> plot time in hour:min:sec:millisecond format. I
> searched extensively for this kind of format. If this
> is not defined it would be incredibly useful to have
> such a format in future release.

I suspect it's nowhere near as useful as you think it is.  It's
practically impossible to actually need both millisecond resolution and an
entire hour of reach on a single axis --- no output device you're likely
to use will be able to resolve 3 million positions along a single axis.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: (No subject)

[Hans-Bernhard B. <br...@ph...>]

[Note: this had been sitting in limbo for 4 weeks between Dartmouth and
SF.net]

On Thu, 6 May 2004 sch...@me... wrote:

> I would be very grateful if you could help me
> with a gnuplot problem. Using splot to plot
> a data file with hidden line removal is
> very slow.  

I may have answered that elsewhere, but for the record: that's because the
default build of 3.8j didn't enable either of the two speed-ups of the
hidden3d code, which has been completely rewritten since 3.7.
 
Build your own binary of version 4.0 and check out the configure option
--enable-h3d-quadtree, and this should improve by quite a bit.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: Plotting with linespoints

[Petr M. <mi...@ph...>]

> Is it possible to make a graph from a datafile with "linespoints" where the line
> is drawn for every data-point and the points are drawn for every nth datapoint?

yes

plot 'bla.dat' with line, '' every 20 with point

--
PM

Re: loose plan (was Re: K&R C)

[Ethan A M. <merritt@u.washington.edu>]

On Thursday 03 June 2004 03:10 am, Petr Mikulik wrote:
>
> Currently, I build my gnuplot with these patch:
> 	raise
> 	with image
> 	set mouse key
> 	$#
>
> I would see them in "regular development version of" gnuplot (soon?).

I have five patchsets that I consider stable and ready to go into
CVS whenever continued development is split off from 4.0 bugfixing.

   fillbounds	   adds "above" and "below" bound-clipping to filledcurves
   datastrings	   'plot with labels', tic and plot labels from datafile, etc 
   histograms	   plot with histograms
   fillcurves	   use fill style for filledcurves (cgm gd pdf post svg x11)
   XEvent  	   reorganizes X event loop to speed up refresh

The 'set mouse key' patch works, but I'm still cleaning it up and it
sure would be nice to have string variables so that I could return
the keypress as a character rather than as a key code.

-- 
Ethan A Merritt
Department of Biochemistry & Biomolecular Structure Center
University of Washington, Seattle

Windows EMF

[BBands <bb...@ya...>]

Dear gnuplotters,

Per an earlier discussion on EMFs and Windows. 

Things have changed for EMFs in Windows. A Windows
security patch to fix an EMF buffer overflow
vulnerability, KB 835732, prevents many EMFs from
displaying. See MS Knowledge Base article KB 840997.
The hotfix did not work for me, I had to back out the
security patch to get gnuplot generated EMFs to
display.

Pass this along if you think it germane.

    --jab

=====
John Bollinger, CFA, CMT
www.BollingerBands.com

If you advance far enough, you arrive at the beginning.


	
		
__________________________________
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/