Is there a reason to prefer one of these approaches over the other? Maybe one is more efficient and should be preferred if doing a large number of appends.
Approach #1
s := ''. s := s, 'foo'. s := s, 'bar'
Approach #2
stream := WriteStream on: ''. stream nextPutAll: 'foo'. stream nextPutAll: 'bar'. s := stream contents
Are there other approaches I should consider?
--- Mark Volkmann
"Mark" == Mark Volkmann mark@ociweb.com writes:
Mark> stream := WriteStream on: ''. Mark> stream nextPutAll: 'foo'. Mark> stream nextPutAll: 'bar'. Mark> s := stream contents
s := String streamContents: [:stream | stream nextPutAll: 'foo'; nextPutAll: 'bar'. ].
Then you don't even have to name the stream, because it has that temporary name inside the block.
Look for the many senders of #streamContents: as examples.
Hello Mark,
Friday, October 3, 2008, 4:52:39 PM, you wrote:
MV> Is there a reason to prefer one of these approaches over the other? MV> Maybe one is more efficient and should be preferred if doing a large MV> number of appends.
speed is the reason for the stream approach. The #, approach copies the first string once for each #, send.
MV> Are there other approaches I should consider?
Don't know of any but then I'm not an expert.
Cheers,
Herbert
Just check it by yourself. ;)
|result| 100 timesRepeat: [ result := String new. 4000 timesRepeat: [ result := result, 'abcdefg' ]]. ^result
|writer| 100 timesRepeat: [ writer := WriteStream on: String new. 4000 timesRepeat: [ writer nextPutAll:'abcdefg' ]]. ^writer contents Cheers,
That reminds me. I know I ran across a method that runs a block and then tells you how long it took. Anyone remember what that's called?
On Oct 3, 2008, at 3:09 PM, Mariano Abel Coca wrote:
Just check it by yourself. ;)
|result| 100 timesRepeat: [ result := String new. 4000 timesRepeat: [ result := result, 'abcdefg' ]]. ^result
|writer| 100 timesRepeat: [ writer := WriteStream on: String new. 4000 timesRepeat: [ writer nextPutAll:'abcdefg' ]]. ^writer contents
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 2:02 PM, Herbert König herbertkoenig@gmx.net wrote: Hello Mark,
Friday, October 3, 2008, 4:52:39 PM, you wrote:
MV> Is there a reason to prefer one of these approaches over the other? MV> Maybe one is more efficient and should be preferred if doing a large MV> number of appends.
speed is the reason for the stream approach. The #, approach copies the first string once for each #, send.
MV> Are there other approaches I should consider?
Don't know of any but then I'm not an expert.
Cheers,
Herbert
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
--- Mark Volkmann
I think you must be thinking in #timeToRun.
About that... In VisualAge there are the EsbTracer and EsbSampler, which allows you to meassure the time that involves every message sent. It's useful to tune some code... But don't know in Squeak. Any ideas?
Cheers,
On Oct 3, 2008, at 3:59 PM, Mariano Abel Coca wrote:
I think you must be thinking in #timeToRun.
No, but you saying that reminded me of where I saw it. It's in the Time class.
ms := Time millisecondsToRun: [ do some stuff here ]
Using this on the code below, the comma approach took 8478 milliseconds and the stream approach took 185 milliseconds. Big difference!
About that... In VisualAge there are the EsbTracer and EsbSampler, which allows you to meassure the time that involves every message sent. It's useful to tune some code... But don't know in Squeak. Any ideas?
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 5:44 PM, Mark Volkmann mark@ociweb.com wrote: That reminds me. I know I ran across a method that runs a block and then tells you how long it took. Anyone remember what that's called?
On Oct 3, 2008, at 3:09 PM, Mariano Abel Coca wrote:
Just check it by yourself. ;)
|result| 100 timesRepeat: [ result := String new. 4000 timesRepeat: [ result := result, 'abcdefg' ]]. ^result
|writer| 100 timesRepeat: [ writer := WriteStream on: String new. 4000 timesRepeat: [ writer nextPutAll:'abcdefg' ]]. ^writer contents
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 2:02 PM, Herbert König herbertkoenig@gmx.net wrote: Hello Mark,
Friday, October 3, 2008, 4:52:39 PM, you wrote:
MV> Is there a reason to prefer one of these approaches over the other? MV> Maybe one is more efficient and should be preferred if doing a large MV> number of appends.
speed is the reason for the stream approach. The #, approach copies the first string once for each #, send.
MV> Are there other approaches I should consider?
Don't know of any but then I'm not an expert.
Cheers,
Herbert
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Mark Volkmann
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
--- Mark Volkmann
Am 03.10.2008 um 14:13 schrieb Mark Volkmann:
On Oct 3, 2008, at 3:59 PM, Mariano Abel Coca wrote:
I think you must be thinking in #timeToRun.
No, but you saying that reminded me of where I saw it. It's in the Time class.
ms := Time millisecondsToRun: [ do some stuff here ]
which is the same as
ms := [ do some stuff here ] timeToRun.
- Bert -
Actually, the definition of timeToRun in BlockContext is:
BlockContext>>timeToRun "Answer the number of milliseconds taken to execute this block."
^ Time millisecondsToRun: self
So, your example it's the same than:
[ do some stuff here ] timeToRun
Remember: let the objects do the work for you :-)
On Fri, Oct 3, 2008 at 6:13 PM, Mark Volkmann mark@ociweb.com wrote:
On Oct 3, 2008, at 3:59 PM, Mariano Abel Coca wrote:
I think you must be thinking in #timeToRun.
No, but you saying that reminded me of where I saw it. It's in the Time class.
ms := Time millisecondsToRun: [ do some stuff here ]
Using this on the code below, the comma approach took 8478 milliseconds and the stream approach took 185 milliseconds. Big difference!
Nice... I've never measured it ;)
About that... In VisualAge there are the EsbTracer and EsbSampler, which allows you to meassure the time that involves every message sent. It's useful to tune some code... But don't know in Squeak. Any ideas?
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 5:44 PM, Mark Volkmann mark@ociweb.com wrote:
That reminds me. I know I ran across a method that runs a block and then tells you how long it took. Anyone remember what that's called?
On Oct 3, 2008, at 3:09 PM, Mariano Abel Coca wrote:
Just check it by yourself. ;)
|result| 100 timesRepeat: [ result := String new. 4000 timesRepeat: [ result := result, 'abcdefg' ]]. ^result
|writer| 100 timesRepeat: [ writer := WriteStream on: String new. 4000 timesRepeat: [ writer nextPutAll:'abcdefg' ]]. ^writer contents Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 2:02 PM, Herbert König herbertkoenig@gmx.netwrote:
Hello Mark,
Friday, October 3, 2008, 4:52:39 PM, you wrote:
MV> Is there a reason to prefer one of these approaches over the other? MV> Maybe one is more efficient and should be preferred if doing a large MV> number of appends.
speed is the reason for the stream approach. The #, approach copies the first string once for each #, send.
MV> Are there other approaches I should consider?
Don't know of any but then I'm not an expert.
Cheers,
Herbert
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Mark Volkmann
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Mark Volkmann
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Thanks to all! Another thing to add to my growing set of notes.
On Oct 3, 2008, at 4:41 PM, Mariano Abel Coca wrote:
Actually, the definition of timeToRun in BlockContext is:
BlockContext>>timeToRun "Answer the number of milliseconds taken to execute this block."
^ Time millisecondsToRun: selfSo, your example it's the same than:
[ do some stuff here ] timeToRun
Remember: let the objects do the work for you :-)
On Fri, Oct 3, 2008 at 6:13 PM, Mark Volkmann mark@ociweb.com wrote: On Oct 3, 2008, at 3:59 PM, Mariano Abel Coca wrote:
I think you must be thinking in #timeToRun.
No, but you saying that reminded me of where I saw it. It's in the Time class.
ms := Time millisecondsToRun: [ do some stuff here ]
Using this on the code below, the comma approach took 8478 milliseconds and the stream approach took 185 milliseconds. Big difference!
Nice... I've never measured it ;)
About that... In VisualAge there are the EsbTracer and EsbSampler, which allows you to meassure the time that involves every message sent. It's useful to tune some code... But don't know in Squeak. Any ideas?
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 5:44 PM, Mark Volkmann mark@ociweb.com wrote: That reminds me. I know I ran across a method that runs a block and then tells you how long it took. Anyone remember what that's called?
On Oct 3, 2008, at 3:09 PM, Mariano Abel Coca wrote:
Just check it by yourself. ;)
|result| 100 timesRepeat: [ result := String new. 4000 timesRepeat: [ result := result, 'abcdefg' ]]. ^result
|writer| 100 timesRepeat: [ writer := WriteStream on: String new. 4000 timesRepeat: [ writer nextPutAll:'abcdefg' ]]. ^writer contents
Cheers,
-- Mariano.
On Fri, Oct 3, 2008 at 2:02 PM, Herbert König herbertkoenig@gmx.net wrote: Hello Mark,
Friday, October 3, 2008, 4:52:39 PM, you wrote:
MV> Is there a reason to prefer one of these approaches over the other? MV> Maybe one is more efficient and should be preferred if doing a large MV> number of appends.
speed is the reason for the stream approach. The #, approach copies the first string once for each #, send.
MV> Are there other approaches I should consider?
Don't know of any but then I'm not an expert.
Cheers,
Herbert
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Mark Volkmann
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
Mark Volkmann
Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
-- Mariano. www.egola.com.ar _______________________________________________ Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
--- Mark Volkmann
Or you can pick "tally it" from the menu in a waorkspace, instead of "do it".
Hello Mark,
Saturday, October 4, 2008, 8:17:46 AM, Aidan wrote:
AG> Or you can pick "tally it" from the menu in a waorkspace, instead of "do it".
to be a bit more explicit, you might want to look at MessageTally's class side #tallySends and TimeProfileBrowser #spyOn: and TheWorldMenu #startThenBrowseMessageTally. The latter is what gets invoked from the World menu, debug start/browse MessageTally.
To understand what you see there please search the archives (of Squeak dev) this is a FAQ.
BTW #tallySends IMHO uses the simulator, you would not want to use it on the examples of Mariano due to its looow speed.
beginners@lists.squeakfoundation.org
-
Aidan Gauland -
Bert Freudenberg -
Herbert König -
Mariano Abel Coca -
Mark Volkmann -
merlyn＠stonehenge.com