In my use case, I'm trying to build a temp file for a MySQL "LOAD DATA LOCAL INFILE".  I do not want the file to persist after the script has run, and there's no clear way to remove an SplFileObject at the end of the script.  My recourse was to use SplTempFileObject calling:

$tempFile = new \SplTempFileObject(0);

According to the documentation:

"public SplTempFileObject::__construct ([ int $max_memory ] )"
and
"If max_memory is zero, no memory will be used"

which forces the tempFile to be written to disk.  While I can see my temp file during script execution, I have no way of feeding the file URI to my SQL call.

SplTempFileObject is implemented as simple wrapper over php://temp
and php://memory, respectively,[1] and as such it's not possible
to get the actual filename AFAIK.  This is why ::getRealPath()
returns FALSE (it's not an empty string).

[1] <https://github.com/php/php-src/blob/php-7.3.0beta2/ext/spl/spl_directory.c#L2314-L2323>

On having a closer look, the behavior is unlikely to be
intentional.  SplTempFileObject::getFileName() returns the name of
the php:// wrapper, and ::getRealPath() returns false, because
realpath(php://temp) returns false; ::getPath(), however, returns
an empty string.

Still, changing this would be a BC break per se, but worse, the
change would need to be quite intrusive, since php://temp's
writing to temp is hidden in the stream wrapper and the actual
path to the temp file is ignored[1].  Given that the
implementation behaves this way since "forever"[2], we need to
document that behavior anyway.  Thus, changing to doc problem.
Could switch back to bug or feature request, after the docs have
been fixed.  I'm not sure how to fix the docs, though, since none
of these methods are actually overloaded, nor are there properties
to document.

[1] <https://github.com/php/php-src/blob/php-7.4.22/main/streams/memory.c#L384>
[2] <https://3v4l.org/sX6CK>