[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] Re: sqlite3 egg patches for chicken experimental bra
From: |
Felix |
Subject: |
Re: [Chicken-users] Re: sqlite3 egg patches for chicken experimental branch |
Date: |
Sat, 08 Jan 2011 21:52:52 +0100 (CET) |
From: Thomas Chust <address@hidden>
Subject: [Chicken-users] Re: sqlite3 egg patches for chicken experimental branch
Date: Thu, 6 Jan 2011 04:05:44 +0100
> 2011/1/5 Alan Post <address@hidden>:
>> noop and milliseconds->time are removed (after being deprecated) in
>> the experimental branch of chicken. The following patches allow the
>> sqlite3 egg to compile against the chicken experimental branch.
>>
>> I'm least certain about the milliseconds-deprecate patch, as the
>> sqlite3 egg is declared to use fixnum arithmetic, which probably
>> degrades the resolution of the timer once this patch is applied.
>> [...]
>
> Hello Alan,
>
> the patch replacing noop by void is fine, of course. Concerning the
> change of thread-sleep!/ms I have the same doubts as you: It should
> really be possible to sleep less than a second waiting for database
> locks, so the computation inside thread-sleep!/ms would have to be
> changed to return a floating point number. I guess it is best to
> remove the declaration of fixnum arithmetic.
I recommend not to use fixnum mode, and to use fixnum-specific
operators instead. It can happen too easily that a library procedure
returns inexact results (for example for very large numbers that
exceed fixnum range).
Apologies for the ruthless cleaning up over the holidays, but I think
it is better go through this sooner than later.
cheers,
felix