Beta of FreeMat 4

[Eugene]

I posted the first beta version of FreeMat4 for win32.
http://sourceforge.net/project/showfiles.php?group_id=91526&package_id=322238&release_id=682268

One issue I already found is that after installation you have to run:

"c:\Program Files\FreeMat\bin\FreeMat.exe" -i "c:\Program Files
\FreeMat"

to set the root path correctly.

I need volunteers to test the installer and FreeMat 4 itself. Please
post any issues you discover on this list or file bug reports...

Eugene

[adrian]

> I posted the first beta version of FreeMat4 for win32.http://sourceforge.net/project/showfiles.php?group_id=91526&package_i...

>
> One issue I already found is that after installation you have to run:
>
> "c:\Program Files\FreeMat\bin\FreeMat.exe" -i "c:\Program Files
> \FreeMat"
>
> to set the root path correctly.
> I need volunteers to test the installer and FreeMat 4 itself. Please
> post any issues you discover on this list or file bug reports...

So far, quickly I observed only the followings:
- Default window size at first start is very small (almost unusable).
This is so for editor, and helpwin window too.
- The installation does not seem to contain the links to the
documentation and the manual in Start menu (like it did with 3.6)
- the "plug-ins" directory seems to be missing (maybe it's not
required in 4.0 anymore?)
- there's a "tmp" directory in \help\ - I suppose it's a leftover from
the build
- typing "edit" opens the editor but without the "empty" file (but
this problem was there in 3.6 too)
- Ctrl+C doesn't seem to copy the selected text in the console, only
using the menu item for copy.
- trying to run all the tests, I get:
------------------------
In C:/Programs/FreeMat4b/toolbox/test/wbtest_abs_1.m(wbtest_abs_1) at
line 33
In C:/Programs/FreeMat4b/toolbox/test/run_tests.m(run_tests) at
line 1
In docli(builtin) at line 2
In base(base)
In base()
In global()
Error: Undefined function or variable run_path-->
------------------------
However if I try to run FreeMat with the variable you specified above
than it won't start at all, but no error message is displayed. Also I
can't find where FreeMat is putting it's log file (I suppose there
would be more information why it's not starting).

[Eugene]

> However if I try to run FreeMat with the variable you specified above
> than it won't start at all, but no error message is displayed. Also I
> can't find where FreeMat is putting it's log file (I suppose there
> would be more information why it's not starting).

You need to run the command I gave (with flag -i) only once - it only
sets path to installation root directory. It doesn't start FreeMat.
After you run this command you can start FreeMat the usual way (from
start menu, or command line without the flag).

Thank you for reporting other issues.
Eugene

[adrian]

Even if I do that (only once from the command line), if I run
"run_tests", I still get the error from above (as if it can't find
that path).

What about a beta2 that fixes this problem.

Thank you.
P.S. I tried to build it myself under ubuntu 9, but despite of the
upgrade of FreeMat to CMake, the build is still too fragile and
couldn't get it work (but that's another thread on this forum :) ).

[Timothy Cyders]

When I do a simple plot and maximize the window, it is restricted to a smaller space than my entire screen. Windows animates the size transition as if it will fill the entire screen, but the resulting window ends up being a good deal smaller (1120x647 pixels, according to an image save from the plot). I've done this in XP 64-bit and Vista 64-bit with the same results. I've uploaded two screenshots (beforemax.jpg and aftermax.jpg) and the resulting maximized image save, maxplot.png. Dynamic resizing by clicking a corner and dragging results in the same window size maximums.

TJ 

[Samit Basu]

Yes -- I have a branch (HandleRework) where I am fixing this and a few
other problems with the plotting in FreeMat4. I plan to merge those
changes in before we release.

Samit

[Chuong]

Glad to know we come this close to the release.

I have an problem with a program using import() function. It works
with v3.6 but in v4 beta, there is an error with string input in the
imported function. It looks like a problem about how the data is
transfered to the imported function.

Chuong

On May 22, 1:35 pm, Samit Basu <basu.sa...@gmail.com> wrote:
> Yes -- I have a branch (HandleRework) where I am fixing this and a few
> other problems with the plotting in FreeMat4.  I plan to merge those
> changes in before we release.
>
> Samit
>

> On Thu, May 21, 2009 at 1:20 PM, Timothy Cyders <t.cyd...@gmail.com> wrote:
> > When I do a simple plot and maximize the window, it is restricted to a
> > smaller space than my entire screen. Windows animates the size transition as
> > if it will fill the entire screen, but the resulting window ends up being a
> > good deal smaller (1120x647 pixels, according to an image save from the
> > plot). I've done this in XP 64-bit and Vista 64-bit with the same results.
> > I've uploaded two screenshots (beforemax.jpg and aftermax.jpg) and the
> > resulting maximized image save, maxplot.png. Dynamic resizing by clicking a
> > corner and dragging results in the same window size maximums.
> > TJ
>

> > On Wed, May 20, 2009 at 1:17 PM, Eugene <gene...@gmail.com> wrote:
>
> >> I posted the first beta version of FreeMat4 for win32.
>

> >>http://sourceforge.net/project/showfiles.php?group_id=91526&package_i...

[Timothy Cyders]

I've found that ode45 is printing out some debug var's - for example:

--> [t, y] = ode45('secondode', [0 1], [1 1])k =  2 y =     1.0000    1.0010     1.0000    1.0010 t =    1.0e-003 *          0    1.0000
k =  3

y =     1.0000    1.0010    1.0060     1.0000    1.0010    1.0060

t =    1.0e-003 *          0    1.0000    6.0000

k =  4

y =     1.0000    1.0010    1.0060    1.0315     1.0000    1.0010    1.0060    1.0315

t =    1.0e-002 *          0    0.1000    0.6000    3.1000

k =  5

y =     1.0000    1.0010    1.0060    1.0315    1.1400     1.0000    1.0010    1.0060    1.0315    1.1400

t =          0    0.0010    0.0060    0.0310    0.1310

.....(you get the idea)


t =  Columns 1 to 11         0    0.0010    0.0060    0.0310    0.1310    0.2310    0.3310    0.4310    0.5310    0.6310    0.7310 

Columns 12 to 14    0.8310    0.9310    1.0000

y =     1.0000   
1.0000    
1.0010   
1.0010    
1.0060   
1.0060    
1.0315   
1.0315   
1.1400   
1.1400    
1.2599   
1.2599   
1.3924   
1.3924   
1.5388   
1.5388    
1.7006   
1.7006    
1.8795   
1.8795   
2.0772  
2.0772    
2.2956   
2.2956    
2.5370   
2.5370    
2.7183   
2.7183


Also, I believe MATLAB's standard output for t is a column matrix, not a row matrix, for what it's worth.

TJ

[Timothy Cyders]

I don't know if you're looking at subplot in your handle rework, but it seems to be borked. Calling subplot creates a figure, but any subsequent plot commands don't plot anything, you just end up with a blank figure box. If there was already a plot active in the figure, it clears the figure leaving a blank set of axes, but doesn't add any data with a plot command, and doesn't separate the axes into the defined subplot areas.

TJ

[Samit Basu]

Fixed in the HandleRework branch. Thanks for pointing out the problem.

Samit

[Arnaud Amiel]

Cool! It is the first time I look at a beta of Freemat.

It tries to install in my %HOMEPATH% dicectory which happens to be H:\
here... I don't know if it is normal behaviour. I would expect C:
\Program Files

I look at Help -> About and I get "FreeMat version VERSION"
I also look Help -> Online Manual and I get "FreeMat v svn
Documentation"

Now I need to play around

Arnaud

[Arnaud Amiel]

The installer ties to install in my %HOMEDRIVE% whic on this PC
happens not to be c:\program files.
Not sure if this is the expected behaviour.

The about box does not report the version nor does the online manual.

Apart from that, starting to play with version 4. I already found a
log10(0) in one of my scripts that was not picked up by previous
versions.

Nice job anyway...

Arnaud