On Tue, Mar 3, 2009 at 12:20 AM, Phil Berry <2009pb@...> wrote:
> I'm trying to understand the startup options. Working through the Java code
> the options listed below seem to be the definitive list. I've put some
> summary comments next to some of them, but there are some which I can't work
> out what they are for.
> --noinit If you have defined an initialisation file you can use this option
> to prevent this file from being processed when the application starts.
> --noinform Suppresses the information that is normally printed when the
> application starts.
> --batch No idea what this is for.
Neither do I, but I imagine it could be an option to terminate the
application upon an unrecoverable error instead of asking for user
> --eval No idea what this is for. It requires an argument, but using a
> filename fails, and putting a Lisp form there fails also.
May I ask what Lisp form you gave it? Because what you describe is
exactly what this option is for.
> --load This takes a filename as an argument. The file is executed when the
> system is started.
exactly although the wording is "loaded", not executed :-)
> --load-system-file This requires a filename as an argument. It seems to use
> a Lisp variable called *lisp-home*, or at least look in the home directory
> of the application. Not sure where this is.
I imagine --load-system-file is used by the ABCL build process or for
some other internal means. I don't think it shoud be used by users.
The *lisp-home* variable is used to find where to load the system
files from, if not from the ABCL jar file. It shouldn't normally
apply, since normal users are expected to run the program from the
> Can anyone please fill in the gaps?
> I must admit that as I work through the Java code I'm surprised that there
> are so few comments. This is making it very diffiuclt to understand what
> goes on.
Exactly. It's something I'm hoping we'll be able to change some day.
In your search, did you take notes? If so, could you rewrite them to
code-comments and submit them? It would be really great to have a
contribution like that! Unfortunately, we inherited the code base with
as few comments as it is. The only thing we can do is add them after
the fact. You could really help us there.
> The code itself is very well written, but it is rather dismaying to
> find so little documentation within the source code. It would be good to
> have Javadoc associated at least with the public methods, and the best time
> to write this is when the method is written -- or at least this is my view
> on writing good code.
Absolutely agreed. However, as mentioned before, the code was
inherited as-is for the most part. I've declared that all changes to
the Java code base should at least have documentation associated in
terms of Javadoc. However, it could take forever before we've reached
a usefull level of documentation if we only add comments upon
> Oh well, I'll keep plugging away...
Thanks for your comments!
If you have notes you want to see added to the documentation, please
consider whether adding a page to the wiki
(http://trac.common-lisp.net/armedbear/wiki) would be appropriate. If
you can't access/create the page yourself, there's always the
possibility of submitting the changes to this list so others can add