This patch is something we're using in production in a local fork at the moment. Would it be possible to get this merged into mainline so we can retire our local fork?

Would be great if this PR was accepted - also having to maintain my own fork right now. :(

@giampaolo is there any chance of getting this PR reviewed and merged please?
Thanks!

The reason why I didn't merge this yet was because reading /proc/PID/status on __init__ has a performance penalty.

Before patch:

~/svn/psutil {master}$ python3 -m timeit -s "import psutil" "psutil.Process()"
50000 loops, best of 5: 9.33 usec per loop

After patch:

~/svn/psutil {master}$ python3 -m timeit -s "import psutil" "psutil.Process()"
20000 loops, best of 5: 17.7 usec per loop

It's not big but still...
I wanted to check if there is some other way to identify whether PID is a TID, but I am afraid there isn't. :(

Hi @giampaolo, thanks for the reply.

We could look at a refactor that removes the check for the PID identity and whether it is a TID into an API that is only run when the caller explicitly asks for it to be checked? That means the __init__ performance would stay the same, however the option would be there is a caller wanted to know whether a PID was a TID or not?