Fix PropertyVector<T*> for multi-component case #1441
Conversation
| @@ -185,6 +197,20 @@ friend class Properties; | |||
| return t; | |||
| } | |||
|
|
|||
| //! Returns the value for the given component stored in the given tuple. | |||
| T const& getComponent(std::size_t tuple_index, | |||
| std::size_t component) const | |||
There was a problem hiding this comment.
At least on github the indentation seems to be not correct. ✅
|
|
||
| void initPropertyValue(std::size_t group_id, std::vector<T> const& values) | ||
| { | ||
| T* p = new T[values.size()]; |
There was a problem hiding this comment.
Maybe a check values.size() == _n_components could be inserted? ✅
| @@ -155,13 +156,24 @@ friend class Properties; | |||
|
|
|||
| void initPropertyValue(std::size_t group_id, T const& value) | |||
| { | |||
| _values[group_id] = new T(value); | |||
| T* p = new T[1]; | |||
There was a problem hiding this comment.
Maybe a check _n_components == 1 could be inserted? ✅
|
If tests are ok: 👍 |
There was a problem hiding this comment.
I am OK with this PR.
I am only wondering:
Is PropertyVector<T*> used anywhere? If I remember correctly the material group parameter will not be based on this class.
So what is a use-case for the multi-component extension?
And what will be the fate of this class in the medium/long run? Especially since its current implementation cannot store abstract types (am I correct?), nor unique_ptrs.
| OGS_FATAL("Single-component version of initPropertyValue() is called " | ||
| "for a multi-components PropertyVector<T*>"); | ||
| T* p = new T[1]; | ||
| p[0] = value; |
There was a problem hiding this comment.
Note: This will not work if T is an abstract class or T is not copyable. However, it did not work either before.
|
@chleh As far as I know, |
|
OpenGeoSys development has been moved to GitLab. |
Currently
PropertyVector<T*>works only with a single component.