[OctDev] Re: octave-forge category index on sourceforge broken

[Paul K. <pki...@ja...>]

On Fri, Mar 28, 2003 at 10:55:57AM +0100, David Bateman wrote:
> With one issue... I changed main/comm/INDEX to include operators (eg
> "+ - = Addition and subtraction in a Galois Field.") and all of the
> unimplemented Matlab functions. However, only the functions that are
> either m-files or oct-files are referenced on the webpages. It however
> seems to have worked for the Signal Processing Toolbox. I imagine it
> is something I have done wrong, but I can't see it...

I'm not sure what you mean because it works for me now.

It might be that I neglected to do "cvs -q update -dP" before building
the index.

It might be that you were trying single character functions by themselves.
When it changed from:

   /^\s+(\S+)\s*=\s*(\S.*\S)\s*$/

to 

   /^\s+(\S.*\S)\s*=\s*(\S.*\S)\s*$/

the following was no longer recognized:

   ! = use system("command")

This has now been fixed.

BTW,  Should the categorical index show the latest release or the
latest CVS?


There is still a problem with the index in that it does not
handle overloaded functions.  For example,

	Galois Fields
	  ...
	  size = The size of a Galois array

does not do anything since the note is ignored if a
function by that name is documented.  I now say that the
note is being ignored.  What we should do is more 
complicated.  The category specific note should show
up in the category list, but all notes should show up
in the alphabetical list.  The current make_index
data structures will not handle this so somebody with
more time than me will have to work on this.

There is also a potential issue with combined notes such as the following:

	Galois Fields
	  ...
	  + - = Addition and subtraction in a Galois Field.
	  * / \ = Matrix multiplication and division of Galois arrays.
	  .* ./ .\ = Element by element multiplication and division of Galois arrays.
	  ** ^ =  Matrix exponentiation of Galois arrays.
	  .** .^ = Element by element matrix exponentiation of Galois arrays.
	  ' .' = Matrix transpose of Galois arrays.
	  == ~= != > >= < <= = Logical operators on Galois arrays.

All other type extensions (such as Symbolic and Sparse) will want to use
the same sets with the same spacing and ordering within the set since the
note is triggered off the entire string.  I don't think anything bad will
happen if the strings are different, so perhaps it won't be a problem.

Paul Kienzle
pki...@us...