Re: CML2-2.3.0 is available

["Eric S. Raymond" <esr@thyrsus.com>]

Jean-Luc Leger <reiga@dspnet.fr.eu.org>:
> On Fri, Feb 15, 2002 at 09:06:24AM -0500, Eric S. Raymond wrote:
> > Symbol type is inferred from use in a menu.
> 
> what about those "legend" symbols that were in menus list but with
> no menu declaration ?
> 
> ex:
> menu usb        # USB? support
>         USB_DEBUG
>         usb_options_legend
>         USB_DEVICEFS USB_BANDWIDTH USB_LONG_TIMEOUT
>         usb_controllers_legend
>         h:USB_UHCI? h:USB_UHCI_ALT? h:USB_OHCI?
> ...
> 
> I see no difference in use for usb_options_legend and USB_DEBUG.

Good catch!  Messages had been temporarily disabled, but I finished 
adding a message declaration suffix (!) just before your mail came in.
In 2.3.1 your example will look like this.

menu usb        # USB? support
         USB_DEBUG
         usb_options_legend!
         USB_DEVICEFS USB_BANDWIDTH USB_LONG_TIMEOUT
         usb_controllers_legend!
         h:USB_UHCI? h:USB_UHCI_ALT? h:USB_OHCI?
 ...

The reason for this change in design is to make it easier to automatically
generate per-directory distributed symbols files.  But the good thing about
it is that it simplifies the language -- I was able to remove three constructs
and only have to add back one.
-- 
		<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>