What's the best way to write complex emacs extension

What's the best way to write complex emacs extension

[scopeview .]

[-- Attachment #1: Type: text/plain, Size: 992 bytes --]

Hi all,

When I wrote my simple extension, pure elisp code is ok. But I found it is
hard to write complex elisp extension. Because when elisp predication
failed (for example stringp) it hard to find out the failed predication
place. (no file:function:line_number or backtrace)

So my complex extension is elisp + c++ code and I make elisp as simple
as possible. We can link c code to emacs binary, but I don't think it's a
good solution. In xemacs, we can extend elisp by writing c/c++ shared
library and load it dynamicly. If emacs has this feature, my complex emacs
extension can be much simpler.

Does there any other good way to extend emacs?
How about other emacs-alike editor like Hemlock written in common lisp
which can use many mature and complex common lisp library? Does anybody use
it?
I think xemacs's c/c++ shared library is more suitable. But it seems that
xemacs had stop its development.


following link is a simple extension of mine
http://www.emacswiki.org/emacs/dir-list

[-- Attachment #2: Type: text/html, Size: 1135 bytes --]

What's the best way to write complex emacs extension

[Stephen J. Turnbull]

scopeview . writes:

 > failed (for example stringp) it hard to find out the failed predication
 > place. (no file:function:line_number or backtrace)

edebug.  XEmacs has a native Lisp debugger as well (maybe Emacs does too).

 > But it seems that xemacs had stop its development.

The rumors of our death are greatly exaggerated, but this is not the
place to discuss that.

Re: What's the best way to write complex emacs extension

[Stefan Monnier]

> When I wrote my simple extension, pure elisp code is ok. But I found it is
> hard to write complex elisp extension. Because when elisp predication
> failed (for example stringp) it hard to find out the failed predication
> place. (no file:function:line_number or backtrace)

Options => Enter Debugger on Error


        Stefan

Re: What's the best way to write complex emacs extension

[Stephen Leake]

"scopeview ." <scopeview126@gmail.com> writes:

> So my complex extension is elisp + c++ code and I make elisp as simple
> as possible. We can link c code to emacs binary, but I don't think it's a
> good solution. In xemacs, we can extend elisp by writing c/c++ shared
> library and load it dynamicly. If emacs has this feature, my complex emacs
> extension can be much simpler.

This is being worked on, in the branch at
https://github.com/aaptel/emacs-dynamic-module

Note that Emacs is transitioning to use git for the main repository
soon, so this branch will change to use that soon.

-- 
-- Stephe

Re: What's the best way to write complex emacs extension

[scopeview .]

[-- Attachment #1: Type: text/plain, Size: 686 bytes --]

Thanks

I have used edebug for developing my elisp extension.

I just feel elisp + XML + c binary for complex extension is not so
convenient. Even with edebug, when some predication failed, I should guess
where it could be then set edebug function entry.

On Sun, Nov 2, 2014 at 1:27 PM, Stefan Monnier <monnier@iro.umontreal.ca>
wrote:

> > When I wrote my simple extension, pure elisp code is ok. But I found it
> is
> > hard to write complex elisp extension. Because when elisp predication
> > failed (for example stringp) it hard to find out the failed predication
> > place. (no file:function:line_number or backtrace)
>
> Options => Enter Debugger on Error
>
>
>         Stefan
>

[-- Attachment #2: Type: text/html, Size: 1147 bytes --]

Re: What's the best way to write complex emacs extension

[Alexis]

scopeview . writes:

> I have used edebug for developing my elisp extension.
>
> I just feel elisp + XML + c binary for complex extension is not so
> convenient. Even with edebug, when some predication failed, I should guess
> where it could be then set edebug function entry.

Are you already aware of the `debug-on-message` variable?

    Documentation:
    If non-nil, debug if a message matching this regexp is displayed.

It would let you do something like:

    (setq debug-on-message "Wrong type argument: stringp")


Alexis.