Added SRFI 170 to the CC: list due to this discussing its timeline.

From: "Arthur A. Gleckler" <xxxxxx@speechcode.com>
Date: Wednesday, June 24, 2020 2:09 PM

On Wed, Jun 24, 2020 at 9:13 AM <xxxxxx@ancell-ent.com> wrote:

I am very comfortable in saying both John and I are done with SRFI 170, and it only exposes a not very controversial bit of SRFI 198, syscall-error? to determine if an object is one, and the accessors for its components.  I'm pretty sure there's nothing that can conceivably be added to this object now that syscall-name is in it.  The design space is rather finite. in part thanks to the memory and I/O limitations (slow 110/300 baud terminals) of the systems Original Formula UNIX ran on.

I have no informed opinion on SRFI 198, pretty sure I'll be finished with my part by the end of the week if not today, but given that it's summer, 60 days, or even the end of October to get past the crush of the beginning of the academic year would not be out of the question. 

I'm not willing to go that far. Given that SRFI 170 is already 328 days older than the maximum technically allowed by the SRFI process, I'm going to keep the actual limit at whatever gets SRFI 198 to 60 days.  SRFI 170 describes an API more complex than most, and with more historical and technical constraints than most, so extra time is warranted.  Reviewers should have plenty of time to review it thoroughly.  But there are limits.

So 60 - 8 days for closing off SRFI 170?  That's OK for me, I for example just fixed a couple of bugs I now realize crept in from our initial overloading of errno-error to also signal Scheme level detected errors; lies, but accurate ones.  Wasn't caught because to (chibi test) an exception is an exception, including screwing up error handling code.  Which of course is a long known locus of not so well tested for bugs.

In the coming weeks I'll see if I can improve the testing to check for the right sort of objects raised in exceptions, and who knows, maybe more.  I am using my SRFI 170 in production for my under development backup program.

- Harold