This is the mail archive of the
ecos-discuss@sourceware.cygnus.com
mailing list for the eCos project.
RE: building something real
- To: Paul Kinzelman <PaulK at isdcorp dot com>
- Subject: RE: [ECOS] building something real
- From: Gary Thomas <gthomas at cygnus dot co dot uk>
- Date: Wed, 19 Apr 2000 11:16:31 +0100 (BST)
- Cc: ecos-discuss at sourceware dot cygnus dot com
- Organization: Cygnus Solutions, UK
On 17-Apr-00 Paul Kinzelman wrote:
> Could you please point me to something that describes how to build
> something real? In other words, I've built all the tests, but haven't a
> clue how to do something on my own or to use some of the packages.
> Did I just miss the documentation for this?
>
> For instance, the 'hello' application has only a 'main' routine. The
> other tests seem to all have a 'cyg_user_start' which looks like
> it's probably the entrypoint.
>
> For instance, how would I create a simple eCos program to
> watch for characters to be typed and just show them on the
> screen? The EDB7xxx kbd_support.c, etc., seem to have
> just their own entrypoints that some overall null process
> must call or something. Dare I ask about a tcpip stack?
>
> What am I missing?
> TIA!
> -PaulK
>
There is a progression of "entry" points in eCos:
cyg_start
cyg_user_start
main
Each of these has a default behaviour and can be separately overridden.
Most of our tests use 'cyg_user_start' as the entry point, since 'main'
is only used if you are linking in the C library functions. If you don't
want/need the 'main()' functionality, in particular the argument stuff
which is common in Unix style applications, using 'cyg_user_start' is
just as good.
As for "real" programs, actually looking at the TCP/IP tests would be
a good idea since these are stand alone programs, and not created in
the same fashion as the other tests in the system.