Re: Use aligned-on-x-parent instead of other callbacks for some grobs (issue 127860043 by address@hidden)

[dak]

On 2014/08/18 06:09:44, janek wrote:

> As far as i can see, the code for getting non-boolean properties
follows the
> convention:

> Type value = robust_scm2type (grob->get_property("prop-name"),
default)

> When writing code for issue 2245, I have initially tried to use
robust_scm2bool
> - which didn't exist.

And it's pointless since unset options are #f by default and convention.
 Somewhat irritatingly, separately to to_boolean there _does_ exist a
ly_scm2bool (which is equivalent to scm_is_true and thus defaults to #t
(!) when given a non-boolean).  It is used a few times in our code base.
 Most of the uses should likely be replaced by to_boolean, some
conceivably by scm_is_true.  And afterwards, all instances of to_boolean
should be renamed to ly_scm2bool and the behavior of _that_ changed to
be that of to_boolean.

> As i was not aware of to_boolean, i ended up with
> ly_scm2bool, which turned out to be a bad choice.

> So, my question is: can we do something to make this more intuitive?

The most important answer to that question in this context is: not as
part of *this* issue.  Please use to_boolean here.  I remember being
tripped up by ly_scm2bool at some point of time in the past.  I'm
definitely sympathetic to changing its behavior to that of to_boolean
and getting rid of to_boolean altogether (after making sure that all
current uses of ly_scm2bool are mapped to the most appropriate of either
to_boolean behavior under the guise of "new ly_scm2bool" or to
scm_is_true).  But not intermingled with this issue, and most certainly
not half-intermingled with this issue.

https://codereview.appspot.com/127860043/