On Thu, 2020-04-16 at 22:16 +0200, Wolfram Sang wrote:
>
> > Wait, is the "RelWithDebInfo" as good a choice, and already gets
> > enforced in the absence of a user spec? And is accepting the user
> > spec -- when provided -- the most appropriate activity? Does it
> > mean that strictly speaking this specific change is not needed?
>
> OK, my take on this was "enforce a valid setting".
>
> Sure, it can be argued that the user is allowed to setup invalid
> settings, especially if the default value is working.
This specific change also actively _prevents_ users from specifying
other values which also were valid. Sure the result is _one_ valid
configuration, unconditionally, but users (here: developers) may
want to have a choice.
So my vote would be on not forcing anything. Assume developers
know what they do, especially if they specify build configurations.
> Maybe the best thing is to bail out if STACKTRACE is enabled and '-g' is
> not in CFLAGS?
Guess a warning for suspicious configurations is useful. If such
a test can get implemented with acceptable effort (and portably,
without assuming specific cmake internals in application rules).
When the warning cannot easily get implemented, I would not care
(again: assuming developers know what they do).
A hard error may be too strict, considering that an optional
feature may not work which is motivated by rare exceptional
conditions, while the application's main feature set does work.
virtually yours
Gerhard Sittig
--
If you don't understand or are scared by any of the above
ask your parents or an adult to help you.
_______________________________________________
sigrok-devel mailing list
sigrok-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sigrok-devel
