Mantis: 0000474 AccurateDateAndTime harvested?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Mantis: 0000474 AccurateDateAndTime harvested?

keith1y
Has this been included in a base squeak image yet?

RC2?

best regards

Keith
Send instant messages to your online friends http://uk.messenger.yahoo.com 

Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

stephane ducasse-2
Hi

I will release RC2.
But we should not add more to it because this will never stop.

So we should go 3.10 after and start to build the next release.

Stef


On 12 oct. 06, at 12:21, Keith Hodges wrote:

> Has this been included in a base squeak image yet?
>
> RC2?
>
> best regards
>
> Keith
> Send instant messages to your online friends http://
> uk.messenger.yahoo.com


Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

Giovanni Corriga
Il giorno gio, 12/10/2006 alle 21.06 +0200, stephane ducasse ha scritto:
> Hi
>
> I will release RC2.
> But we should not add more to it because this will never stop.

yes, Yes, YES. Unless it's a serious showstopper, it should be punted to
either 3.9.1 or 3.10.

        Giovanni


Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

timrowledge
In reply to this post by stephane ducasse-2

On 12-Oct-06, at 12:06 PM, stephane ducasse wrote:

> Hi
>
> I will release RC2.
> But we should not add more to it because this will never stop.

I claim we must fix the incorrect compiledmethod sharing currently  
affecting traits. As Andreas said "should always return the class it  
*installed* in, never the trait it is *defined* in." and the result  
ofsourceClass should be the same as methodClass - and then source  
class should be removed. It seems like it ought to be simple but I'm  
awaiting a reply from Adrian about it.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Futuristic: It will only run on a next generation supercomputer.



Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

keith1y
In reply to this post by Giovanni Corriga
Giovanni Corriga wrote:

> Il giorno gio, 12/10/2006 alle 21.06 +0200, stephane ducasse ha scritto:
>  
>> Hi
>>
>> I will release RC2.
>> But we should not add more to it because this will never stop.
>>    
>
> yes, Yes, YES. Unless it's a serious showstopper, it should be punted to
> either 3.9.1 or 3.10.
>
> Giovanni
>
>  
I agree,  but I think some improvements can be made...

cc from note in mantis.

I made a variant which effectively uses the millisecond clock for the
time, and the second clock for the date. I do this in anticipation of
the day when the vm primitives give us useful milliseconds (or
nanoseconds) since midnight.  [ Is this likely to happen? ]

I also discovered that avoiding the use of Duration in the instantiation
of DateAndTime improves performance by a factor of x3 on Andreas'
observation. Making an accessor to set the instance vars directly
without normalizing (I don't understand the purpose of normalising)
boosts that to a factor of x4.

Then removing the semaphore and use of the class var LastTicks improves
speed up to a pretty wizzy 90x the original.

Perhaps for those of us who dont care about this
  (DateAndTime now) <= (DateAndTime now) is always true*

would prefer the quicker version which I have christened.

DateAndTime-i-#rightNow

:-)

Keith
Send instant messages to your online friends http://uk.messenger.yahoo.com 

Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

stephane ducasse-2
In reply to this post by timrowledge
Adrian is also managing a company in his other life. So he will  
certainly reply but this will take more time.

Stef


On 12 oct. 06, at 23:54, tim Rowledge wrote:

>
> On 12-Oct-06, at 12:06 PM, stephane ducasse wrote:
>
>> Hi
>>
>> I will release RC2.
>> But we should not add more to it because this will never stop.
>
> I claim we must fix the incorrect compiledmethod sharing currently  
> affecting traits. As Andreas said "should always return the class  
> it *installed* in, never the trait it is *defined* in." and the  
> result ofsourceClass should be the same as methodClass - and then  
> source class should be removed. It seems like it ought to be simple  
> but I'm awaiting a reply from Adrian about it.
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Futuristic: It will only run on a next generation supercomputer.
>
>
>


Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

Marcus Denker
In reply to this post by keith1y

On 12.10.2006, at 12:21, Keith Hodges wrote:

> Has this been included in a base squeak image yet?
>

of course not.

I added this to mantis on 10-28-04 15:38

The very first comment on this bug is by you from 10-12-06 14:50

A fix that is so important that nobody looks at it for 2 years is for  
sure need
to be addt between Release Canditate 1 and 2. Yes. Obviously.


      Marcus



Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

Marcus Denker
In reply to this post by keith1y

On 13.10.2006, at 06:09, Keith Hodges wrote:

> Giovanni Corriga wrote:
>> Il giorno gio, 12/10/2006 alle 21.06 +0200, stephane ducasse ha  
>> scritto:
>>
>>> Hi
>>>
>>> I will release RC2.
>>> But we should not add more to it because this will never stop.
>>>
>>
>> yes, Yes, YES. Unless it's a serious showstopper, it should be  
>> punted to
>> either 3.9.1 or 3.10.
>>
>> Giovanni
>>
>>
> I agree,  but I think some improvements can be made...
>
> cc from note in mantis.
>
> I made a variant which effectively uses the millisecond clock for  
> the time, and the second clock for the date. I do this in  
> anticipation of the day when the vm primitives give us useful  
> milliseconds (or nanoseconds) since midnight.  [ Is this likely to  
> happen? ]
>
> I also discovered that avoiding the use of Duration in the  
> instantiation of DateAndTime improves performance by a factor of x3  
> on Andreas' observation. Making an accessor to set the instance  
> vars directly without normalizing (I don't understand the purpose  
> of normalising) boosts that to a factor of x4.
>
> Then removing the semaphore and use of the class var LastTicks  
> improves speed up to a pretty wizzy 90x the original.
>
> Perhaps for those of us who dont care about this
>  (DateAndTime now) <= (DateAndTime now) is always true*
>
> would prefer the quicker version which I have christened.
>
> DateAndTime-i-#rightNow
>

And there is http://bugs.impara.de/view.php?id=4669

0004669: [ENH] TimeForSpeed

"
"DateAndTime now" is about 6 times faster and "Date today" 4 times.  
The full chronology test suite runs almost 2 times faster. See  
preamble for a do-it that shows this (run before and after installing  
cs).
"

     Marcus


Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

keith1y
In reply to this post by Marcus Denker
Marcus Denker wrote:

>
> On 12.10.2006, at 12:21, Keith Hodges wrote:
>
>> Has this been included in a base squeak image yet?
>>
>
> of course not.
>
> I added this to mantis on 10-28-04 15:38
>
> The very first comment on this bug is by you from 10-12-06 14:50
>
> A fix that is so important that nobody looks at it for 2 years is for
> sure need
> to be addt between Release Canditate 1 and 2. Yes. Obviously.
>
>
>      Marcus
>
>
>
>
Was that sarcasm?

I asked the question, simply and only because I am still using 3.8 and
have never even looked at 3.9 to know what is or isn't in there.

Other people have looked at it since, just not in that particular report.
 
The whole point of using testing in the first place is to reduce the
'cost' of releasing, to release frequently and often etc. If the test
suite is comprehensive enough to cover the bases, and the test suite
passes, then including it is a 'no brainer'. I don't think it matters
what release, or when. In this case where you are fortunate enough to
have a test suite, then why not go for it.

So, although I hadn't suggested putting this in RC2, and I don't think
that my fixes/ideas will have been sufficiently tested for a couple of
days, now you come to mention it, I think that that it would be a great
idea to include it once the testing has been done.

Keith








Send instant messages to your online friends http://uk.messenger.yahoo.com 

Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

Marcus Denker

On 13.10.2006, at 15:10, Keith Hodges wrote:

> Marcus Denker wrote:
>>
>> On 12.10.2006, at 12:21, Keith Hodges wrote:
>>
>>> Has this been included in a base squeak image yet?
>>>
>>
>> of course not.
>>
>> I added this to mantis on 10-28-04 15:38
>>
>> The very first comment on this bug is by you from 10-12-06 14:50
>>
>> A fix that is so important that nobody looks at it for 2 years is  
>> for sure need
>> to be addt between Release Canditate 1 and 2. Yes. Obviously.
>>
>>
>>      Marcus
>>
>>
>>
>>
> Was that sarcasm?
>

Yes, sorry.

> I asked the question, simply and only because I am still using 3.8  
> and have never even looked at 3.9 to know what is or isn't in there.
>
> Other people have looked at it since, just not in that particular  
> report.
> The whole point of using testing in the first place is to reduce  
> the 'cost' of releasing, to release frequently and often etc. If  
> the test suite is comprehensive enough to cover the bases, and the  
> test suite passes, then including it is a 'no brainer'. I don't  
> think it matters what release, or when. In this case where you are  
> fortunate enough to have a test suite, then why not go for it.
>
> So, although I hadn't suggested putting this in RC2, and I don't  
> think that my fixes/ideas will have been sufficiently tested for a  
> couple of days, now you come to mention it, I think that that it  
> would be a great idea to include it once the testing has been done.
>

The idea of gamma (which was the state we where in *before* RC1) is  
defined as "Show Stopping Bugs only". The reason for that is that we  
need to find
a way to stop and *release* 3.9. There are lots of cool changes,  
sure. Especially because they where not added to 3.9 after it went  
gamma. If we
now start to add again everything that's cool, we will *never* be  
able to release.

Why don't we release first (fixing those bugs that are real show  
stoppers), and then start again to add stuff with 3.10?

     Marcus

Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

keith1y

>
> Why don't we release first (fixing those bugs that are real show
> stoppers), and then start again to add stuff with 3.10?
>
>     Marcus
>
>
I have no problem with that at all, particularly since I can feel
stephane's pain

Keith
Send instant messages to your online friends http://uk.messenger.yahoo.com 

Reply | Threaded
Open this post in threaded view
|

Re: Mantis: 0000474 AccurateDateAndTime harvested?

stephane ducasse-2
This is not pain this is just that we should stop and celebrate then  
restart :)

Stef

On 13 oct. 06, at 15:39, Keith Hodges wrote:

>
>>
>> Why don't we release first (fixing those bugs that are real show  
>> stoppers), and then start again to add stuff with 3.10?
>>
>>     Marcus
>>
>>
> I have no problem with that at all, particularly since I can feel  
> stephane's pain
>
> Keith
> Send instant messages to your online friends http://
> uk.messenger.yahoo.com