On Thursday, 22 December 2016 at 18:49:02 UTC, Stefan Koch wrote:
On Thursday, 22 December 2016 at 18:04:51 UTC, Observer wrote:
(1) Serve as a convenient breakpoint handle in the debugger,
perhaps
as a kind of centralized this_cannot_ever_happen()
function.
(2) conditionally_die(conditions);
(3) Sleep for some run-time-computable length of time.
(4) Yield the thread so other threads can take execution
priority.
(5) Yield the entire process so other processes can take
execution
priority.
(6) Wait for an external trigger (perhaps a hardware
interrupt, for
instance).
(7) Invoke a pass of garbage collection.
A function that does any of that cannot be pure.
You judge too quickly. (1) is certainly possible:
% cat pure.d
pure void cannot_happen_breakpoint () {
return;
}
int main () {
cannot_happen_breakpoint();
return 0;
}
% dmd --version
DMD64 D Compiler v2.071.0
Copyright (c) 1999-2015 by Digital Mars written by Walter Bright
% dmd pure.d
% nm -pr pure | fgrep cannot
0000000000422a60 T _D4pure24cannot_happen_breakpointFNaZv
% gdb pure
GNU gdb (Ubuntu 7.7.1-0ubuntu5~14.04.2) 7.7.1
Copyright (C) 2014 Free Software Foundation, Inc.
...
Reading symbols from pure...(no debugging symbols found)...done.
(gdb) break _D4pure24cannot_happen_breakpointFNaZv
Breakpoint 1 at 0x422a64
(gdb) run
Starting program: /home/anon/dlang/pure
warning: the debug information found in "/lib64/ld-2.19.so" does
not match "/lib64/ld-linux-x86-64.so.2" (CRC mismatch).
[Thread debugging using libthread_db enabled]
Using host libthread_db library
"/lib/x86_64-linux-gnu/libthread_db.so.1".
Breakpoint 1, 0x0000000000422a64 in
pure.cannot_happen_breakpoint() ()
With respect to the others, I see in testing that the compiler
objects to a pure function calling an impure function such as
sched_yield(). I was misled in my thinking by the
nearly-exclusive
focus on memory mutation in the TDPL discussion of pure functions.
Note to Andrei: in a revised TDPL, this aspect should be
broadened,
to include explicit mention of other sorts of things that a pure
function is not allowed to do. Yes, I know the text says "a
function
is considered pure if returning a result is its only effect", but
from a teaching point of view, that's not enough emphasis to catch
the reader's attention when the entire rest of the discussion is
about mutation.
(Though frankly, I don't know why sched_yield() shouldn't be
marked
as pure in core.sys.posix.sched. It's not like a pure function
cannot be interrupted by a timeslice expiration at the OS level,
so calling sched_yield() ought to open no new doors.)
