Re: CVS, GWorkspace.app and app wrappers

[Enrico Sersale]

On 2004-09-30 21:49:05 +0300 Charles Phlip Chan <cpchan@sympatico.ca> wrote:

On 30 Sep 2004, enrico@imago.ro wrote:

> > The first problem (the wrappers) is not GWorkspace related; I've not
> > changed anything there. And I suppose that Charles means that he has
> > updated the libraries, not gw.

> Yes, it is the libs and not GWorkspace.app's fault- I downgraded
> GWorkspace.app to 0.6.5 for testing with the same results. :-(

> Enrico, the GWorkspace.app in cvs is looking very nice. The "show
> menu" is very interesting, however I still think a "list view" ala
> Workspace.app would be nice. Also, I think implimenting the following
> will make GWorkspace.app even better:

For the list view I'm only waiting for NSTableView to become more stable.
Probably, 0.7 will have the list viewer.

> (1) Annotations for files and be able to search by it. This is the
>    main reason why I am still using TKdesk as my main file manager.

> (2) Colourisation of files by user defined catagories ala finder.app.

The new classes I'm using for the icons already allow colourisation of the icon 
or the file name. The problem is only where to store the added info about the 
file. (see below)

Regarding (1) and (2):
The problem is that annotations and colors must be saved somewhere.
But I've just started to think to a solution for the "Live Search Folders" I'm 
writing and the annotations could be saved in the same place.
Perhaps I'll use libdb with a btree access method but I'm not sure. Any 
suggestion?

> (3) A PS inspector.

I've had a PS inspector. This was when I used ghostscript to preview pdf files.
Now I've switched to PDFKit, beeing it a GNUstep library. I'll try to think a 
new solution for ps files...

> (4) Use "magic numbers" to determine file type if there is no
>     extension. I use Emacs-wiki extensively, thus I have a lot of text
>     files with no extensions- it would be nice to be able to preview
>     them in the "content inspector". Also, IMHO I think the text
>     inpector should be turned on for other types of text based files
>     with no "content inspector" yet, such as html and tex.

This is a interesting think. For a while I've used the output of the "file"
command but now I'd want to implement some of its features directly in
Inspector. Does anyone know where I can find a list of the sequences of bits
to read to identify a file type? I'd want something simplier than
/usr/share/magic used by file.