Support parsing elements that are not part of the schema #638
Conversation
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
jennuine
added this to In progress
in SDFormat Development TRI (Moved to https://github.com/orgs/gazebosim/projects/2)
via automation
jennuine
moved this from In progress
to In Review
in SDFormat Development TRI (Moved to https://github.com/orgs/gazebosim/projects/2)
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
src/Param_TEST.cc
Outdated
| @@ -75,11 +77,6 @@ TEST(Param, Bool) | |||
| strParam.Get<bool>(value); | |||
| EXPECT_TRUE(value); | |||
|
|
|||
| // Anything other than 1 or true is treated as a false value | |||
There was a problem hiding this comment.
Is there a reason for removing this test?
There was a problem hiding this comment.
Since this section was removed:
https://github.com/ignitionrobotics/sdformat/blob/f3dbe1d225939ff1bc038b0789fff884d763cfa7/include/sdf/Param.hh#L296-L315
It now reaches this error: https://github.com/ignitionrobotics/sdformat/blob/f0f9759ec9d4572db9c5fd9e0dd861e6f679a7cf/src/Param.cc#L329
To me, it made more sense to error when a random value such as % was set rather than defaulting to false. Should I change this back?
There was a problem hiding this comment.
Yes, I think making it an error makes sense, but I'd like to avoid changing behavior since this is a stable version and someone wrote a test to deliberately check this behavior. Would it be possible to assign the value regardless of the error? I know printing the error is still a change in behavior but I think it's small enough not to break anything.
There was a problem hiding this comment.
How's this? 650b544
This seemed to be the easiest fix. Changing ParamPrivate::ValueFromStringImpl bool section to assign the value regardless of the error and return true (since _valueToSet was successfully set) caused Param.InvalidBool test to fail
There was a problem hiding this comment.
That works. Can you add a comment there saying that the piece of code handling bool types is there to keep backward behavior?
I'm thinking we should change the behavior for Fortress. Would you agree?
There was a problem hiding this comment.
I agree. I'll add the comment and put a TODO to remove for Fortress
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
Signed-off-by: Jenn Nguyen <jenn@openrobotics.org>
j-rivero
removed this from Done
in SDFormat Development TRI (Moved to https://github.com/orgs/gazebosim/projects/2)
🦟 Bug fix
Fixes #603
Summary
sdf::Param::Getparses a value to a given type regardless of thetypeNameof thesdf::Param. Previously whentypeName = "string"thensdf::Param::Get<double>("inf")would return0, now it returns the expected valueinf.Test:
Checklist
codecheckpassed (See contributing)Note to maintainers: Remember to use Squash-Merge