Re: [RFA] Implement show | set can-call-inferior-functions [on|off]

[Tom Tromey <tom@tromey.com>]

>>>>> "Eli" == Eli Zaretskii <eliz@gnu.org> writes:

>> From: Philippe Waroquiers <philippe.waroquiers@skynet.be>
>> Cc: gdb-patches@sourceware.org
>> Date: Wed, 24 Apr 2019 23:25:27 +0200
>> 
>> Generally, returning 0 (or whatever value) can then later on 
>> cause problems in a script.
>> For example, evaluating some function calls sometimes imply to first
>> call malloc in the inferior.  When can-call-inferior-functions is off,
>> returning 0 (or whatever) from malloc will then cause further problems
>> (such as a SEGV).

Eli> Then perhaps a built-in variable to test whether inferior calls are
Eli> allowed would be of help?  Then script writers could look at that
Eli> variable and avoid calling inferior functions if they want the script
Eli> to continue running regardless, perhaps with reduced functionality.
Eli> WDYT?

An error definitely seems like the correct thing to do here.  Other
choices will give silently mysterious behavior.

A convenience variable is fine, too, though I suspect it will be little
used.  I wouldn't require it for this patch to go in.

gdb could use better control over the handling of errors.  There's a
catch/throw patch sitting in bugzilla that would be good to resurrect.
Or, the ignore-errors command (implemented in Python) could be checked in.

Tom