* New macro with-demoted-errors
@ 2007-07-07 6:07 Stefan Monnier
2007-07-07 20:47 ` Richard Stallman
0 siblings, 1 reply; 3+ messages in thread
From: Stefan Monnier @ 2007-07-07 6:07 UTC (permalink / raw)
To: emacs-devel
I often want to use CL's `ignore-errors' but am annoyed regularly that its
use tends to hide bugs and make them harder to track down.
So I sugest we introduce a new macro with-demoted-errors which does the same
thing except that errors are reported (via message) to the user rather than
being silently ignored. Also errors are not ignored if debug-on-error is
set, in the hope to make debugging easier.
The patch below includes a sample use of it in VC where it makes sure that
any error in a vc-BACKEND-registered function will not prevent the user from
visiting a file.
Stefan
--- orig/lisp/subr.el
+++ mod/lisp/subr.el
@@ -2552,6 +2552,20 @@
(or (input-pending-p)
,@body))))))
+(defmacro with-demoted-errors (&rest body)
+ "Run BODY and demote any errors to simple messages.
+If `debug-on-error' is non-nil, run BODY without catching its errors.
+This is to be used around code which is not expected to signal an error
+but which should be robust in the unexpected case that an error is signalled."
+ (declare (debug t) (indent 0))
+ (let ((err (make-symbol "err")))
+ `(lexical-let ((f (lambda () ,@body)))
+ (if debug-on-error
+ (funcall f)
+ (condition-case ,err
+ (funcall f)
+ (error (message "Error: %s" ,err) nil))))))
+
(defmacro combine-after-change-calls (&rest body)
"Execute BODY, but don't call the after-change functions till the end.
If BODY makes changes in the buffer, they are recorded
--- orig/lisp/vc-hooks.el
+++ mod/lisp/vc-hooks.el
@@ -834,7 +834,7 @@
(when buffer-file-name
(vc-file-clearprops buffer-file-name)
(cond
- ((vc-backend buffer-file-name)
+ ((with-demoted-errors (vc-backend buffer-file-name))
;; Look for pending-partial-checkin file and deal with it if found.
(vc-after-partial-checkin buffer-file-name)
;; Compute the state and put it in the modeline.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: New macro with-demoted-errors
2007-07-07 6:07 New macro with-demoted-errors Stefan Monnier
@ 2007-07-07 20:47 ` Richard Stallman
2007-07-08 14:08 ` Stefan Monnier
0 siblings, 1 reply; 3+ messages in thread
From: Richard Stallman @ 2007-07-07 20:47 UTC (permalink / raw)
To: Stefan Monnier; +Cc: emacs-devel
+ `(lexical-let ((f (lambda () ,@body)))
lexical-let is a cl macro, and I'd rather not make subr.el use them.
Perhaps we should add a way to make condition-case operate only if
(null debug-on-error). It could use some sort of marker symbol
in the handler:
(condition-case ,err
(funcall f)
((debug error) (message "Error: %s" ,err) nil))))))
would mean to catch all errors but only if debug-on-error is nil.
That would be a lot more convenient, whether for writing the macro,
for for direct use.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: New macro with-demoted-errors
2007-07-07 20:47 ` Richard Stallman
@ 2007-07-08 14:08 ` Stefan Monnier
0 siblings, 0 replies; 3+ messages in thread
From: Stefan Monnier @ 2007-07-08 14:08 UTC (permalink / raw)
To: rms; +Cc: emacs-devel
> + `(lexical-let ((f (lambda () ,@body)))
> lexical-let is a cl macro, and I'd rather not make subr.el use them.
It probably breaks bootstrapping indeed.
I can make it use make-symbol, if you prefer.
> Perhaps we should add a way to make condition-case operate only if
> (null debug-on-error). It could use some sort of marker symbol
> in the handler:
> (condition-case ,err
> (funcall f)
> ((debug error) (message "Error: %s" ,err) nil))))))
> would mean to catch all errors but only if debug-on-error is nil.
But that same syntax currently means "catch `error' or `debug' signals", so
it seems a bit dangerous to change its meaning. How 'bout
(condition-case ,err
(funcall f)
(((not debug) error) (message "Error: %s" ,err) nil))))))
Still, a problem here is that it requires extending `condition-case' in the
C code and in the byte-compiler. I'd rather just provide a new
macro condition-case-non-debug.
> That would be a lot more convenient, whether for writing the macro,
> for for direct use.
Indeed.
Stefan
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2007-07-08 14:08 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-07-07 6:07 New macro with-demoted-errors Stefan Monnier
2007-07-07 20:47 ` Richard Stallman
2007-07-08 14:08 ` Stefan Monnier
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).