Re: Problems with edebug-set-conditional-breakpoint.

[Matt Hodges]

>>>>> Juri Linkov writes:

 > > I've been reading through lispref/edebug.texi and have come
 > > across two related problems with
 > > edebug-set-conditional-breakpoint.

 > Thanks for your report. The cause of both problem is unusual
 > handling of the history and default values (pushing the default
 > value temporarily to the history before reading an expression).

 > The standard way is to use either initial-contents or default-value
 > argument. In this case initial-contents looks better, because if
 > the default-value was used then too many parens inside the
 > top-level parens of "(default ...)" would be too confusing, and
 > secondly the current break condition might be too long so the
 > complete prompt wouldn't fit to the screen.

Thanks, Juri. The patch eliminates the errors for me.

My only minor comment is that if read-expression-history is the same
between invocations of edebug-set-conditional-breakpoint, such that
(format "%s" edebug-break-condition) and (car read-expression-history)
are equal, then two presses of M-p are required to get a different
entry. This feels a bit strange, but maybe that's just me -- I do
think presenting a default value is a good idea, though.

Matt