From enaeher at gmail.com  Fri Jan  3 20:58:55 2014
From: enaeher at gmail.com (Eli Naeher)
Date: Fri, 3 Jan 2014 14:58:55 -0600
Subject: Durations and intervals
Message-ID: <CAJVWyAyMyo95TAgZa=Jd4Hm1q4-3rKyZo7YYNLcJd7reD65j=Q@mail.gmail.com>

Hello,

I'm interested in adding durations, intervals, and some arithmetic support.
I see that this stuff is all present in the TODO.

Would such a patch be welcomed? Is there any particular way that the
maintainers would prefer this be implemented?

-Eli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/local-time-devel/attachments/20140103/405205c5/attachment.html>

From enaeher at gmail.com  Tue Jan  7 18:49:32 2014
From: enaeher at gmail.com (Eli Naeher)
Date: Tue, 7 Jan 2014 12:49:32 -0600
Subject: local-time-duration
Message-ID: <CAJVWyAyso5qExtAis_b1gVwwVt10Gi8=yUi0=-Q_ymHdFoNQAQ@mail.gmail.com>

Hello,

After looking at the time-interval package, I determined that it did some
things that I didn't need (like handling intervals denominated in months
and years) and wasn't able to do some things I did need (like comparing
intervals expressed in different units), so I have written
local-time-duration, available at
https://github.com/enaeher/local-time-duration. There is room for
optimization but it does what I need and may be useful to others.

I re-used the %defcomparator macro from local-time to implement things like
duration>, duration<, etc., but had to modify it slightly since the
original asserts that its arguments are timestamps. Would a patch to
local-time so that %defcomparator can be passed an expected argument type
be accepted? If so, I will prepare a pull request.

-Eli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/local-time-devel/attachments/20140107/99065cc7/attachment.html>

From psilord at cs.wisc.edu  Fri Jan 31 22:07:47 2014
From: psilord at cs.wisc.edu (Peter Keller)
Date: Fri, 31 Jan 2014 16:07:47 -0600
Subject: Patch to support Allegro Common Lisp
Message-ID: <20140131220747.GA21821@cs.wisc.edu>

Hello,

I have produced a patch which allows LOCAL-TIME to support subsecond
resolution of time in Allegro Common Lisp. I produced it with git
format-patch and attached it to this email.  Hopefully, it'll be good
enough without further changes. Let me know if you need anything else.

Thank you.

Peter Keller
-------------- next part --------------
>From ae6325f7aadaf771f6ed86ad7c4893818351031e Mon Sep 17 00:00:00 2001
From: Peter Keller <psilord at cs.wisc.edu>
Date: Fri, 31 Jan 2014 15:57:26 -0600
Subject: [PATCH] Add support for subsecond resolution under Allegro.

---
 src/local-time.lisp | 27 ++++++++++++++++++++++++++-
 1 file changed, 26 insertions(+), 1 deletion(-)

diff --git a/src/local-time.lisp b/src/local-time.lisp
index 831660d..508160a 100644
--- a/src/local-time.lisp
+++ b/src/local-time.lisp
@@ -947,8 +947,33 @@ elements."
   (declare (type timestamp timestamp))
   (timestamp-values-to-unix (sec-of timestamp) (day-of timestamp)))
 
+#+allegro
+(eval-when (:compile-toplevel :load-toplevel :execute)
+  ;; Allegro common lisp requires some toplevel hoops through which to
+  ;; jump in order to call unix's gettimeofday properly.
+  (ff:def-foreign-type timeval
+      (:struct (tv_sec :long)
+               (tv_usec :long)))
+
+  (ff:def-foreign-call
+      (allegro-ffi-gettimeofday "gettimeofday")
+      ((timeval (* timeval))
+       ;; and do this to allow a 0 for NULL
+       (timezone :foreign-address))
+    :returning (:int fixnum)))
+
 (defun %get-current-time ()
   "Cross-implementation abstraction to get the current time measured from the unix epoch (1/1/1970). Should return (values sec nano-sec)."
+  #+allegro
+  (flet ((allegro-gettimeofday ()
+           (let ((tv (ff:allocate-fobject 'timeval :c)))
+             (allegro-ffi-gettimeofday tv 0)
+             (let ((sec (ff:fslot-value-typed 'timeval :c tv 'tv_sec))
+                   (usec (ff:fslot-value-typed 'timeval :c tv 'tv_usec)))
+               (ff:free-fobject tv)
+               (values sec usec)))))
+    (multiple-value-bind (sec usec) (allegro-gettimeofday)
+      (values sec (* 1000 usec))))
   #+cmu
   (multiple-value-bind (success? sec usec) (unix:unix-gettimeofday)
     (assert success? () "unix:unix-gettimeofday reported failure?!")
@@ -967,7 +992,7 @@ elements."
     (let ((err (ccl:external-call "gettimeofday" :address tv :address (ccl:%null-ptr) :int)))
       (assert (zerop err) nil "gettimeofday failed")
       (values (ccl:pref tv :timeval.tv_sec) (* 1000 (ccl:pref tv :timeval.tv_usec)))))
-  #-(or cmu sbcl (and ccl (not windows)))
+  #-(or allegro cmu sbcl (and ccl (not windows)))
   (values (- (get-universal-time)
              ;; CL's get-universal-time uses an epoch of 1/1/1900, so adjust the result to the Unix epoch
              #.(encode-universal-time 0 0 0 1 1 1970 0))
-- 
1.8.1.2


From attila.lendvai at gmail.com  Fri Jan 24 03:20:22 2014
From: attila.lendvai at gmail.com (Attila Lendvai)
Date: Fri, 24 Jan 2014 09:20:22 +0600
Subject: local-time-duration
In-Reply-To: <CAJVWyAyso5qExtAis_b1gVwwVt10Gi8=yUi0=-Q_ymHdFoNQAQ@mail.gmail.com>
References: <CAJVWyAyso5qExtAis_b1gVwwVt10Gi8=yUi0=-Q_ymHdFoNQAQ@mail.gmail.com>
Message-ID: <CAE4vfc+F93k009VrspO1kRpqMV+1si-TDeOm4i+Yq1QY31MFhA@mail.gmail.com>

> intervals expressed in different units), so I have written
> local-time-duration, available at
> https://github.com/enaeher/local-time-duration. There is room for
> optimization but it does what I need and may be useful to others.


thanks for the note, i've added this to the local-time readme/TODO.

Daniel, can you please flip me the commit bit on github for
local-time?

i don't have any big plans, just small semantic-noop's like
this... and even if i'll have some plans i'll open a discussion and a
branch in my own fork.


> I re-used the %defcomparator macro from local-time to implement things like
> duration>, duration<, etc., but had to modify it slightly since the original
> asserts that its arguments are timestamps. Would a patch to local-time so
> that %defcomparator can be passed an expected argument type be accepted? If
> so, I will prepare a pull request.


that sounds like a reasonable change *in the current setup*, but i
don't think it makes sense to have a separate library for your code;
it should be part of local-time.

but i can sympathize with your decision given the level of feedback
you received... sorry for that! i personally will not have much time
for local-time in the future either.

-- 
? attila lendvai
? PGP: 963F 5D5F 45C7 DFCD 0A39
--
?Democracy: It's not freedom, it's just smarter human farming.?
	? Stefan Molyneux


From enaeher at gmail.com  Fri Jan  3 21:31:00 2014
From: enaeher at gmail.com (Eli Naeher)
Date: Fri, 3 Jan 2014 15:31:00 -0600
Subject: Durations and intervals
In-Reply-To: <CAJVWyAyMyo95TAgZa=Jd4Hm1q4-3rKyZo7YYNLcJd7reD65j=Q@mail.gmail.com>
References: <CAJVWyAyMyo95TAgZa=Jd4Hm1q4-3rKyZo7YYNLcJd7reD65j=Q@mail.gmail.com>
Message-ID: <CAJVWyAySWxT0CwDFSADm-6RwkCDrwEHhmXBN-QNmzoqj4x2w3A@mail.gmail.com>

I have been directed to the existing time-interval package which looks like
it will do what I need to do.

Thank you,
-Eli


On Fri, Jan 3, 2014 at 2:58 PM, Eli Naeher <enaeher at gmail.com> wrote:

> Hello,
>
> I'm interested in adding durations, intervals, and some arithmetic
> support. I see that this stuff is all present in the TODO.
>
> Would such a patch be welcomed? Is there any particular way that the
> maintainers would prefer this be implemented?
>
> -Eli
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/local-time-devel/attachments/20140103/67a80d6e/attachment.html>