Debugging 
Author Message
 Debugging

In Lisp style languages, if you leave off type declarations, you have to
wait until runtime to see if you made a typo (or it's mental equivalent,
a mento).

Lisp makes up for this by giving you a nice de{*filter*}, so that when you
are accidentally adding elephants together, it'll let you know, and
show you how you got there.

How do folks handle stuff like this in Dylan?

I suppose it depends on the implementation, and so Harlequin, I imagine,
has a nice de{*filter*}.

How about Gwydion?

In general, how do folks like to work with Dylan, re debugging? In
Smalltalk, for instance, some folks depend on the de{*filter*} to find
places where they need to add new methods and such. Does that happen in
Dylan, too?



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging

Hi All -

I am trying to compile the Airport Example from the Harlequin web site
under the version of Mindy that comes with GD 2.2.0. I've built it on
the Mac, and the code compiles fine. I then get the error "unknown
library Transcendental in module airport" when trying to run it. It
seems to be this line that cause the problem:

use transcendental, import: {sqrt};

If I comment out the import section to test, I get a complaint about
multiple definitions in/from transcendental.

I don't have easy access to a UNIX box, so can I just ask:
Has anyone else had this problem with the airport example using 2.2.0 on
UNIX?

If there's no problem it's just a problem with the Mac build and I can
continue in the de{*filter*}, I wanted to check this first, though.

- Rob.
__________________________________________________________________
                           __            __    _      __        

http://www.*-*-*.com/ ;/ /__/ _ \(_-</ __/  | |/ |/ / - `/\ \ /
                        /____/\___/___/\__/   |__/|__/\_,_//_\_\



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging


    Date: Tue, 23 Mar 1999 14:19:37 GMT

    In Lisp style languages, if you leave off type declarations, you have to
    wait until runtime to see if you made a typo (or it's mental equivalent,
    a mento).

    Lisp makes up for this by giving you a nice de{*filter*}, so that when you
    are accidentally adding elephants together, it'll let you know, and
    show you how you got there.

    How do folks handle stuff like this in Dylan?

    I suppose it depends on the implementation, and so Harlequin, I imagine,
    has a nice de{*filter*}.

Harlequin Dylan "makes up" for this by having a nice de{*filter*}.  It's
not really "making up", though, because if you left off the type
declarations, it's presumably for ease of initial implementation.

    How about Gwydion?

    In general, how do folks like to work with Dylan, re debugging? In
    Smalltalk, for instance, some folks depend on the de{*filter*} to find
    places where they need to add new methods and such. Does that happen in
    Dylan, too?

After using Dylan for aobut 5 years now, I find that I almost always
include type declarations for inputs and outputs even in the earliest
implementation stages.  I do this for the exact reason given by
advocates of statically typed languages -- the compiler can tell you
early on when type constraints are being violated.  Then I get to
debug actual problems instead of just messing about with type errors.

I used Lisp almost exclusively for 15 years, and I have to say that
there's a lot about Dylan I like better.



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging


    Date: Tue, 23 Mar 1999 15:07:36 +0000

    Hi All -

    I am trying to compile the Airport Example from the Harlequin web site
    under the version of Mindy that comes with GD 2.2.0. I've built it on
    the Mac, and the code compiles fine. I then get the error "unknown
    library Transcendental in module airport" when trying to run it. It
    seems to be this line that cause the problem:

    use transcendental, import: {sqrt};

The library and module are actually called 'transcendentals' (plural).
Is your typo in your code or in this mail message?

    If I comment out the import section to test, I get a complaint about
    multiple definitions in/from transcendental.

    I don't have easy access to a UNIX box, so can I just ask:
    Has anyone else had this problem with the airport example using 2.2.0 on
    UNIX?

    If there's no problem it's just a problem with the Mac build and I can
    continue in the de{*filter*}, I wanted to check this first, though.

    - Rob.
    __________________________________________________________________
                               __            __    _      __        

    http://www.*-*-*.com/ ;/ /__/ _ \(_-</ __/  | |/ |/ / - `/\ \ /
                            /____/\___/___/\__/   |__/|__/\_,_//_\_\



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging

Quote:

>     use transcendental, import: {sqrt};

> The library and module are actually called 'transcendentals' (plural).
> Is your typo in your code or in this mail message?

The typo is in the code, as seen in the downloaded archive ( which is
untampered with :-) ), and in the library used with mindy, here is the
.lid file from CVS:

library: Transcendental
unique-id-base: 3000

library.dylan
transcendental.dylan

I don't think a typo is the problem ( it is consistent ), but I'm at a
loss as to what is.

- Rob.

__________________________________________________________________
                           __            __    _      __        

http://www.lostwax.com/  / /__/ _ \(_-</ __/  | |/ |/ / - `/\ \ /
                        /____/\___/___/\__/   |__/|__/\_,_//_\_\



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging

Quote:

> library: Transcendental
> unique-id-base: 3000

> library.dylan
> transcendental.dylan

> I don't think a typo is the problem ( it is consistent ), but I'm at a
> loss as to what is.

For some unknown reason, the library does appear to be called
'transcendental'. Don't ask me why. :-(

As for your problem with Mindy: the transcendental math functions are not
built on all platforms. You may want to subscribe to the gd-hackers list
and help us find this problem. See:

  http://www.gwydiondylan.org/lists.phtml

Good luck!

Cheers,
Eric



Sat, 08 Sep 2001 03:00:00 GMT  
 Debugging

Quote:

> As for your problem with Mindy: the transcendental math functions are not
> built on all platforms.

You're right, and changing my script that recurses through the "common"
directory structure compiling every .lid file it finds solved half the
problem.
I solved the other half of the problem, and got the program to run, by
adding:
        use transcendental;
to the
        define library airport
section of
        airport-lib.dylan
. It runs quite happily now. Can somebody please comment on this, and if
anyone has a UN*X copy of Mindy to hand, please can you try the Airport
demo from the Harlequin web site, as I *think* I'm doing everything else
right.

Thanks.

- Rob.

__________________________________________________________________
                           __            __    _      __        

http://www.lostwax.com/  / /__/ _ \(_-</ __/  | |/ |/ / - `/\ \ /
                        /____/\___/___/\__/   |__/|__/\_,_//_\_\



Sun, 09 Sep 2001 03:00:00 GMT  
 Debugging


Quote:

>    Date: Tue, 23 Mar 1999 14:19:37 GMT

>    In Lisp style languages, if you leave off type declarations, you have to
>    wait until runtime to see if you made a typo (or it's mental equivalent,
>    a mento).

>    Lisp makes up for this by giving you a nice de{*filter*}, so that when you
>    are accidentally adding elephants together, it'll let you know, and
>    show you how you got there.

>    How do folks handle stuff like this in Dylan?

>    I suppose it depends on the implementation, and so Harlequin, I imagine,
>    has a nice de{*filter*}.

>Harlequin Dylan "makes up" for this by having a nice de{*filter*}.  It's
>not really "making up", though, because if you left off the type
>declarations, it's presumably for ease of initial implementation.

This doesn't really make up for it once your product is delivered,
either in Dylan or in Common Lisp.  It's a major problem.

Dylan is somewhat better about this than Common Lisp, but it's
still much easier than it should be to deliver a program that gets
no-applicable-method errors and other type errors at run time.



Tue, 11 Sep 2001 03:00:00 GMT  
 
 [ 12 post ] 

 Relevant Pages 

1. Debug/Kill + Debug/Run = RB Crash

2. can someone email me Debug.com not debug.exe

3. Debug/Anti-debug/compression help wanted

4. can someone please send me debug.com and not debug.exe

5. LE, SYM,SEPARATE, and debugging withOUT the debug tool

6. debug Python interpreter work with the debug swig dll

7. Python-2.1 Debug libs zip is missing tcl &tk debug libs

8. Debug won't debug

9. Anyone have some debugging aids for debugging pkgIndex.tcl problems?

10. Debugging J Release 2

11. Any suggestions for how to debug this?

12. Walkback Trying to Debug STBFileIn Problem

 

 
Powered by phpBB® Forum Software