Re: Couple comments about image_viewer and path behavior

[Daniel J Sebald]

Quentin Spencer wrote:

> I have always wondered why we had a separate path for images. I'm all 
> for getting rid of it.

I'm on board for that.  The only issue I wonder about is that if someone has a 
path in which there are very large numbers of database files it might slow down 
searching for M files?

John W. Eaton wrote:

> In addition to .m files, Matlab searches "MATLABPATH" for files read
> by load, opened by fopen, imread, etc.  I suppose doing the same for
> Octave would be OK, but to avoid confusion, I don't think we should do
> so without a warning for files that have relative names and that are
> not found in the current directory.

I sort of see what you are getting at, but on further thought, why exactly?  If the user enters 

imread("./image1.jpg");
imread("../image2.jpg");

it is almost implicit that he or she has the current working directory in mind. 
 But not so perhaps for

imread("xray/tibia2.jpg");

The rule is always look from working directory first, right?  So Octave would look for 
"./xray/tibia2.jpg" then any other directories (sub "xray") afterward.

Dan


Dan