TestRunner author while running tests

> The fact that the test runner sets the author name to 'TestRunner'
> while running the tests is really annoying.
>
> I write a lot of code in the debugger while running tests and then all
> the methods end up with the wrong author name.
>
> I would vote for changing that back to the old behavior.
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> Looks like the code of Adrian Lienhard ;-)
>
> I guess that it tries to fix tests so that they do not pup-up a dialog
> when they compile code.
>
> Initially the code in SUnit was like this:
>
> TestCase>>runCase
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> Then it was change to this (this is where I discovered the issue):
>
> TestCase>>runCase
> Author
> useAuthor: 'TestRunner'
> during: [
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> In the latest version it was changed again:
>
> TestCase>>runCase
> Author
> ifUnknownAuthorUse: 'TestRunner'
> during: [
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> In any case the author name is 'TestRunner' if you edit source code
> anywhere in the system when a debugger on a test is open. Also I am
> not sure what happens if you have multiple debuggers on tests open,
> that might give very strange situations.
>
> I suggest that we revert to the earliest presented version of the
> method in question, and fix the particular tests that compile code.
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> Looks like the code of Adrian Lienhard ;-)
>
> I guess that it tries to fix tests so that they do not pup-up a dialog
> when they compile code.
>
> Initially the code in SUnit was like this:
>
> TestCase>>runCase
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> Then it was change to this (this is where I discovered the issue):
>
> TestCase>>runCase
> Author
> useAuthor: 'TestRunner'
> during: [
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> In the latest version it was changed again:
>
> TestCase>>runCase
> Author
> ifUnknownAuthorUse: 'TestRunner'
> during: [
> [self setUp.
> self performTest]
> ensure: [
> self tearDown.
> self cleanUpInstanceVariables ] ]
>
> In any case the author name is 'TestRunner' if you edit source code
> anywhere in the system when a debugger on a test is open. Also I am
> not sure what happens if you have multiple debuggers on tests open,
> that might give very strange situations.
>
> I suggest that we revert to the earliest presented version of the
> method in question, and fix the particular tests that compile code.
>
> Lukas
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> The reason for this change (which I did with Oscar at Lille) was to  
> avoid the author dialog when running tests. It is very annoying  
> because when you run all the tests, get a coffee, and later come back,  
> the test runner would wait for your input halfway through.
>
> Now the first solution below had the annoying problem that  
> 'TestRunner' ends up as your initials when you compiled a method and  
> the testrunner has not returned, e.g., because you have a halt in the  
> code. To improve this, I slightly changed the behavior as follows:  
> (comment from issue #1379):
>
> A preliminary fix is in #10494. Now the name 'TestRunner' is only used  
> if no author was set before starting the tests. The implementation of  
> temporarily setting another author name is too simple to cope with the  
> suspending of a process and opening
> a debugger from a failed assertion. In this case, one would ideally  
> want to use the real author. The desired behavior could be implemented  
> using a dynamic variable to request the author. The test runner could  
> then catch the notification and return the temporary author. This  
> would then not interfere with compiling methods in the debugger  
> because it runs in a different call stack.
>
> As Lukas suggests, instead fixing the offending tests is also a  
> possible solution. If possible, though, I would prefer a generic  
> solution.
>
> Adrian
>
> On Jan 10, 2010, at 09:40 , Lukas Renggli wrote:
>
>> Looks like the code of Adrian Lienhard ;-)
>>
>> I guess that it tries to fix tests so that they do not pup-up a dialog
>> when they compile code.
>>
>> Initially the code in SUnit was like this:
>>
>> TestCase>>runCase
>> [self setUp.
>> self performTest]
>> ensure: [
>> self tearDown.
>> self cleanUpInstanceVariables ] ]
>>
>> Then it was change to this (this is where I discovered the issue):
>>
>> TestCase>>runCase
>> Author
>> useAuthor: 'TestRunner'
>> during: [
>> [self setUp.
>> self performTest]
>> ensure: [
>> self tearDown.
>> self cleanUpInstanceVariables ] ]
>>
>> In the latest version it was changed again:
>>
>> TestCase>>runCase
>> Author
>> ifUnknownAuthorUse: 'TestRunner'
>> during: [
>> [self setUp.
>> self performTest]
>> ensure: [
>> self tearDown.
>> self cleanUpInstanceVariables ] ]
>>
>> In any case the author name is 'TestRunner' if you edit source code
>> anywhere in the system when a debugger on a test is open. Also I am
>> not sure what happens if you have multiple debuggers on tests open,
>> that might give very strange situations.
>>
>> I suggest that we revert to the earliest presented version of the
>> method in question, and fix the particular tests that compile code.
>>
>> Lukas
>>
>> --
>> Lukas Renggli
>> http://www.lukas-renggli.ch
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> 2010/1/10 Adrian Lienhard <[hidden email]>:
>> The reason for this change (which I did with Oscar at Lille) was to
>> avoid the author dialog when running tests. It is very annoying
>> because when you run all the tests, get a coffee, and later come back,
>> the test runner would wait for your input halfway through.
>
> what about ensuring that the author name is set before running the
> tests? i.e., when #runCase is executed, the system asks for the author
> name if it is not already set.
>
>
> --
> Damien Cassou
> http://damiencassou.seasidehosting.st
>
> "Lambdas are relegated to relative obscurity until Java makes them
> popular by not having them." James Iry
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project