Re: FW: error: load: ../some image directory/..: inconsistent number of colum

[Quentin Spencer]

Joe Koski wrote:
> ------ Forwarded Message
> From: "John W. Eaton" <address@hidden>
> Date: Tue, 5 Sep 2006 21:42:24 -0400
> To: Joe Koski <address@hidden>
> Cc: shoker_17 <address@hidden>, <address@hidden>
> Subject: Re: error: load: ../some image directory/..: inconsistent number of
> colum
>
> On  5-Sep-2006, Joe Koski wrote:
>
> | Having said that, I have been seeing the same problem with octave-2.9.8 on
> | my Mac. I haven¹t asked about the problem because octave-2.9.x is still
> | under development, and it may just be a matter of time until the problem
> is
> | corrected.
>
> Maybe you should report the problem, along with enough information to
> allow someone to debug it.  It's hard to fix problems when they
> haven't been reported.
>
> jwe
>
> ------ End of Forwarded Message
> John,
>
> OK, I hear you, but just playing around, I can't get octave-2.9.8 to read
> any image on my machine, regardless of whether they are .jpg, .tiff, etc. I
> have tried both imread and loadimage, and neither seems to be working as
> they have in the past. The test is simple - just try opening any .jpg file.
>
> For imread, I get
>
> octave:1> A=imread("Full-fire-later-bw.jpg");
> error: `IMAGEPATH' undefined near line 41 column 23
> error: evaluating argument list element number 1
> error: evaluating assignment expression near line 41, column 8
> error: called from `imread' in file
> `/usr/local/share/octave/2.9.8/site/m/octave-forge/image/imread.m'
> error: evaluating assignment expression near line 1, column 2
>
> How do I define IMAGEPATH for octave-forge?
>   

This is fixed in CVS (and discussed a month or so ago on one of the 
lists). The problem is that IMAGEPATH is now IMAGE_PATH. If you add that 
one character in your imread function, it will work.

> For loadimage, I get
>
> octave:2> A=loadimage("Full-fire-later-bw.jpg");
> error: load: /Users/jakoski/Codes/octave_test_files/Full-fire-later-bw.jpg:
> inconsistent number of columns near line 2
> error: load: unable to extract matrix size from file
> `/Users/jakoski/Codes/octave_test_files/Full-fire-later-bw.jpg'
> error: called from `loadimage' in file
> `/usr/local/share/octave/2.9.8/m/image/loadimage.m'
> error: evaluating assignment expression near line 2, column 2
>
> Which is the error that shoker_17 is reporting. I have read that same image
> before with octave-2.1.73.
>   

With imread or loadimage? As far as I know, loadimage has always been 
for octave's image format, and imread for all of the others. I would 
expect imread would work correctly with 2.1.73.


> Is there an image reading test that is part of octave or octave-forge? Since
> every part of octave-2.9.8 that I really need is working, I had assumed that
> image reading was a work in progress, and would eventually get fixed. I want
> to help, but for me to stand around saying that things just don't work is
> not useful to anyone, especially when people are already aware of the
> problems. I'm not sure which part of image reading falls onto octave turf
> (loadimage?), and which is part of octave-forge (imread?), and why they both
> exist (history?).
>
> To check things correctly, I should probably be working with the latest CVS
> octave-forge. When I attempt to use the CVS instructions on the octave
> repository web page, I get:
>
> Joe-Koskis-Computer:~ jakoski$ cvs
> -d:pserver:address@hidden:/cvsroot/octave login
> Logging in to 
> :pserver:address@hidden:2401/cvsroot/octave
> CVS password: 
> cvs login: CVSROOT may only specify a positive, non-zero, integer port (not
> `cvsroot').
> cvs login: Perhaps you entered a relative pathname?
> cvs login: warning: skipping invalid entry in password file at line 1
>
> Do I need to export CVSROOT=something before I attempt login?
>   

I wouldn't think so. I just tried an anonymous checkout using the 
command you used above and did not get the error messages. I wonder if 
this is something Mac-specific?

> Once I get past that, I assume the modulename that I want is octave-forge.
> Correct?
>   

Yes.

Quentin