Closed pboettch closed 5 years ago
You're fixing the iter objects - but AFAICT the problem is with the chip object. Was that on purpose?
Of course not. Sorry.
But seeing again where there are deleters in the binding makes me wonder whether it was correct to create a shared_ptr owning a nullptr in the first place.
Yes, this is my other concern - we should probably check if the C function returned a null pointer and then not initialize the chip object.
Yes, as you do in chip_iter& chip_iter::operator++(void)
It's maybe better like this now.
Yes, it looks good now. Can you send it to linux-gpio with [PATCH][libgpiod] tag and your Signed-off-by: tag? I use patchwork for patch management. This github repo is not really the official development tool for me.
Too bad. GIthub/lab review/pr/merge is so superior to mailing lists-review - IMHO.
But I will try to do it, git-send-email-galore. git-format-patch. Oh lala ;-).
Of course I messed up the mail. I didn't know how to add the [libgpiod]-tag to the subject-line. So I added it to the commit-message. Re-run format-patch and send send-email the wrong-file.
Could not resist, I'm so impatient :-/.