ConfigurationOfSpy load gives me 'No version found for ''default'' of ConfigurationOfMemoryMonitor'
I tried both, from Spy and MetacelloRepository. thanks mariano |
Arg. There is no such a problem on the Hudson server: http://hudson.moosetechnology.org/
I tried yesterday, and it works. I will have a look at it on Sunday. Alexandre Le 11 mars 2011 à 07:20, Mariano Martinez Peck <[hidden email]> a écrit : > ConfigurationOfSpy load gives me 'No version found for ''default'' of ConfigurationOfMemoryMonitor' > > I tried both, from Spy and MetacelloRepository. > > thanks > > mariano > |
On Fri, Mar 11, 2011 at 11:53 AM, Alexandre Bergel <[hidden email]> wrote: Arg. There is no such a problem on the Hudson server: http://hudson.moosetechnology.org/ Thanks. Then it seems I don't know how to use it. I want to do a simple TimeProfileBrowser spyOn: [Transcript show: 'zaraza'] with your cool profiler. It seems I cannot do it neither using the UI or script. Any help is appreciated. Alexandre |
Look on the class side of the class KaiProfiler. In any case, I am now working on a very cool new profiler. It takes a bit of time to pull it out but it will come. Maybe for esug. Alexandre
|
On Fri, Mar 11, 2011 at 2:05 PM, Alexandre Bergel <[hidden email]> wrote:
Any of these, answers me an empty profiler: KaiProfiler viewProfiling: [ 1 to: 999 do: [: each | Transcript show: 'aaaa'] ] KaiProfiler viewProfiling: [ 1 to: 999 do: [: each | Transcript show: 'aaaa'] ] inClassCategory: 'classCategory' :(
|
Kai uses the output of messagetally to obtain method execution time and the coverage. What has been covered is then instrumented for a second run of you code. Note that the Pharo kernel is not instrumented (this is why you have an empty result). Additional metrics are then obtained. So, my advice is to not use a 'hello world' like example. On the class side of KaiProfiler there are some examples. But essentially, where messagetally does not give result then Kai will be of no help. In case where message tally gives some results, then Kai will help you to find bottleneck and will give hints on how to resorb it. I know this is a serious limitation of Kai. As I said, I am working on a new generation of code profilers that are not based on execution sampling. I recently shown that counting messages is as efficient as sampling the execution to identify bottleneck. But message counting has some wonderful properties: it works for very short execution, it is much more stable and it is not subject to the memory management. More will come soon. Alexandre
|
Free forum by Nabble | Edit this page |