Re: An engine for octave

[Jesse Bennett]

On Thu, 22 Apr 1999, Min Xu wrote:

> Hi, Bennett:
> 
> > The only problem I had with this approach is that there is not a good way
> > in Octave to "load" a file from stdin (needed by engPutFull).  IIRC, the
> > load will not terminate without an EOF, and the EOF character causes
> > Octave to close stdin (or something like that, I never got around to
> > investigating the problem in detail).  In any case, only the first load
> > succeeds with this method.  
> 
> I took a similar approach and also had problems with put an array into octave.
> The data is transfered using load -mat-binary from stdin of octave and the 
> program just hangs there.

Based on my experience, I believe reading files (ascii or binary) from
stdin will require modifications to Octave.

> > The workaround I used was to create a fifo,
> > and then pass the data through it. This is not a portable approach, since
> > some UNIX systems will not allow ordinary users to create fifos.  A
> > temporary file could also be used, but this is messy as well.  
> 
> But then the octave can only be at the same host as the caller, isn't it?

True.  Another good reason this is a bad approach.

> I also tried fifo, I encountered strange error messages when I transfer
> data in mat-binary format and the program totally hangs whenever a huge
> matrix is being transfered. 

I initially used ascii format only because it facilitates debugging (I can
read the data stream).  I will test my implementation with a binary format
and see if I have the problem you describe.

> > It would be
> > nice if someone created a patch for Octave that allows `load -ascii "-"'
> > to work (at least for a single matrix). 
> 
> Sure. But load -mat-binary - is more desired, which preserves accuracy.

True.

> > Since it seems others are interested in this, I will try to get it
> > organized this weekend and post the code on the sources list.
> 
> I am very interested. And others too. 

I will try to post the code sometime this weekend.

Jesse