Menu
▾
▴
Re: [Gnuplot-info] proporsal of pgnuplot patch
|
From: Tatsuro M. <tma...@ya...> - 2007-12-03 07:28:02
|
Dear Hans-Bernhard Broeker
****************************************************************************************
Sorry for the previous mail had again a lot of errors and the duplicated posting due to my
misoperation.
Please ignore the previous mails.
****************************************************************************************
Tatsuro> The current octave for windows distribution, the Michael's modified pgnuplot was used.
Tatsuro> That is independent of wgnuplot and which has both abilities of pipe receiving and the
Tatsuro> plotting
Tatsuro> system.
Hans> Sorry, but I'm afraid your English is confused enough that I have almost
Hans> no idea what you're talking about. Sounds like that "Michael's
Hans> pgnuplot" isn't actually pgnuplot at all. It's regular gnuplot command
Hans> line, built on Windows.
Sorry for my poor abilities of English. I do not want to say that the Micheal's pgnuplot is irregular.
I think the Michael's way is much better than that I have proposed.
I only would like to point out the possibity of another simply way that enable us to use the non-
specially prepared set of the wgnuplot and pgnuplot.
Tatsuro> I think that the current pgnuplot is not suitable for this purpose and not effective.
Tatsuro> The command will be better to be interpreted by line to line.
Hans> Unfortunately, the attached patch doesn't achieve that. It just inserts
Hans> a rather arbitrary pause of one millisecond per input line. There's no
Hans> particular reason to assume that 1 millisecond would be sufficient or
Hans> necessary.
You are correct.
I tested on three computers, it worked well. But it was only three.
It is dengarous to generalize the success on the only three computers.
But I suspected that bottleneck came from each command line interpritation.
First I forgot to explain what 'the sucessive plots' indiacate.
Istead of wrting poor English, I will write a example code of octave.
x=0:0.01:2; # genarate 200 x axis points
for n=1:100;
plot(x.^2+n/0.05);
endfor
(This was not actual code which was used in the test.)
The plot command from ocatave like this,
***********
gnuplot> plot "-" using ($1):($2) axes x1y1 title "2800: t= 2.6667 s" with lines
linestyle 1 \
> ;
input data ('e' ends) > 0 30
input data ('e' ends) > 0.1 30.5172603938434
input data ('e' ends) > 0.2 31.0339103781298
input data ('e' ends) > 0.3 31.5493410356967
input data ('e' ends) > 0.4 32.0629464247251
input data ('e' ends) > 0.5 32.5741250428799
input data ('e' ends) > 0.6 33.0822812634419
input data ('e' ends) > 0.7 33.5868267344726
input data ('e' ends) > 0.8 34.0871817323741
input data ('e' ends) > 0.9 34.5827764616014
input data ('e' ends) > 1 35.0730522927464
:
:
input data ('e' ends) > e
gnuplot> plot "-" using ($1):($2) axes x1y1 title "2900: t= 3.000 s" with lines
linestyle 1 \
> ;
input data ('e' ends) > 0 30
input data ('e' ends) > 0.1 30.5172603938434
input data ('e' ends) > 0.2 31.0339103781298
input data ('e' ends) > 0.3 31.5493410356967
input data ('e' ends) > 0.4 32.0629464247251
input data ('e' ends) > 0.5 32.5741250428799
input data ('e' ends) > 0.6 33.0822812634419
input data ('e' ends) > 0.7 33.5868267344726
input data ('e' ends) > 0.8 34.0871817323741
input data ('e' ends) > 0.9 34.5827764616014
input data ('e' ends) > 1 35.0730522927464
:
:
input data ('e' ends) > e
**************************
First I suspect that the end of each plot lines was the bottleneck.
So that I placed a rather long wait factor at each " e\n".
But nothing was imporved.
So I simply placed a wait after each line.
After that the successive plot was successful.
So I suspect that interpertating time of each line is slower than incoming commands from the pipe.
Anyway my way will not to be accepted. But I do not mind the fact. That is mere a proporsal.
However, will my results indicate the clue to kick out the problem ?
Buy the way, please allow me to distribute this experimentary binary of the modified pgnuplot under
the gnuplot copyright.
*********************************************************************************
* Permission to modify the software is granted, but not the right to
* distribute the complete modified source code. Modifications are to
* be distributed as patches to the released version. Permission to
* distribute binaries produced by compiling modified sources is granted,
* provided you
* 1. distribute the corresponding source modifications from the
* released version in the form of a patch file along with the binaries,
* 2. add special version identification to distinguish your version
* in addition to the base release version number,
* 3. provide your name and address as the primary contact for the
* support of your modified version, and
* 4. retain our contact information in regard to use of the base
* software.
* Permission to distribute the released version of the source code along
* with corresponding source modifications in the form of a patch file is
* granted with same provisions 2 through 4 for binary distributions
*********************************************************************************
* 2. add special version identification to distinguish your version
* in addition to the base release version number,
I executed pgnuplot --help and pgnuplot --version.
****************
C:\Program Files\gp422win32\gnuplot\bin>pgnuplot --help
Usage: gnuplot [OPTION] [FILE] [-]
-V, --version show gnuplot version
-h, --help show this help
-persist don't close the plot after executing FILE
-noend, /noend like -persist (non-portable Windows-only options)
- allow work in interactive mode after executing FILE
Only on Windows, -persist and - have the same effect.
This is gnuplot 4.2 patchlevel 2
Report bugs to <http://sourceforge.net/projects/gnuplot>
***************
C:\Program Files\gp422win32\gnuplot\bin>pgnuplot --version
gnuplot 4.2 patchlevel 2
*************
But where is the proper space 'add special version identification'.
I cannot find a proper space for pgnuplot.
The message
'gnuplot 4.2 patchlevel 2 (The pgnuplot is a testing binary for using from 'octave'.)'
is proper?
If you would like suggest me, it would be grateful for me.
***********
Tatsuro> The reason is the gnuplot 4.2.0 suffured from the pipe line problem.
Tatsuro> It was solved at the gnuplot 4.2.2.
Hans> I have no idea what "the pipe line problem" is supposed to be.
Sorry also here my poor English
What I want to say the below,
At 4.2.0, the thread of 'unexpected error when very long line given'
http://sourceforge.net/tracker/index.php?func=detail&aid=1673352&group_id=2055&atid=102055
appeared.
Perhaps similar phenomenon occurred at the octave, 'Fwd: Re: gnuplot error when using hist'
http://www.nabble.com/Fwd:-Re:-gnuplot-error-when-using-hist-t4547312.html
In the thread
gnuplot 0.970272 0
^
line 366: invalid command
gnuplot 0.970272 280
^
line 366: invalid command
were written.
In the Unix-octave user, some people still uses the gnuplot 4.0.0.
I have not hear the claim from 4.0.0 user.
I suspect that from gnuplot 4.2, the syntax will be more complex that that of 4.0.0.
Therefore interpretating time of each line will be slower.
>From the gnuplot 4.2.2, such problem have not been reported by Unix Octave user.
For Unix, the problem was solved from 4.2.2.
When I obseved the wgnuplot command window of 4.2.2 in succesive plots with pgnuplot in which my patch
was not applied, the message like
*********************
gnuplot 0.970272 0
^
line 366: invalid command
**********************************
was often appeared.
Acually for octave, Michael's binary solved this.
However, is this proglem only for the octave?
The case : the thread of 'unexpected error when very long line given'
http://sourceforge.net/tracker/index.php?func=detail&aid=1673352&group_id=2055&atid=102055
was for the perl.
Is the currently distributed pgnuplot (not Michael's one) the safe for such case ?
I feel that the currently distributed pgnuplot is neccesary to be revised for such cases.
I cannot say all what I want to say in my poor English. But in the above description if you feel that
what I would like to say, it will be grateful.
Finally I really appreciate you for your reply even though my poor English.
Sincerely yours
Tatsuro Matsuoka
--------------------------------------
New Design Yahoo! JAPAN 2008/01/01
http://pr.mail.yahoo.co.jp/newdesign/
|
