R devel - Jul 2016 - Is .packageName part of the official API?

Hi, I've seen that some packages use .packageName internally to infer
their own name.  Is that officially supported?  I could not find it
documented anywhere.

There's utils::packageName(), which internally looks for .packageName.
However, if the latter is not found, it may return NULL whereas an
error would be more appropriate if a package name is expected.  Using
.packageName would give an error if it does not exist.  Also, which is
minor, using packageName() would add explicit dependency on the utils
package whereas .packageName doesn't.

So, should I use .packageName or utils::packageName() for this?

Thanks

Henrik

On 12/07/2016 2:55 PM, Henrik Bengtsson wrote:
> Hi, I've seen that some packages use .packageName internally to infer
> their own name.  Is that officially supported?  I could not find it
> documented anywhere.
>
> There's utils::packageName(), which internally looks for .packageName.
> However, if the latter is not found, it may return NULL whereas an
> error would be more appropriate if a package name is expected.  Using
> .packageName would give an error if it does not exist.  Also, which is
> minor, using packageName() would add explicit dependency on the utils
> package whereas .packageName doesn't.
>
> So, should I use .packageName or utils::packageName() for this?
You should use utils::packageName.  It is documented; .packageName is not.

Duncan Murdoch