Failing PetitSmalltalk tests
Locked
 
|
Hi.
In Pharo 2.0 PetitSmalltalk tests are failing. Firs of all there are real problems with parsing as some test fail because parser finishes to parse before the end of the stream. Right now I don't know what's causing the problem and how to solve it. Then there is a ton of deprecation warnings as PPToken>>#value became deprecated and it's recommended to use #inputValue. It's now hard to fix that, but there are two problems. 1) At places where #value message is sent, not only PPToken is used as a receiver, but sometimes RBToken is used instead. And RBToken does not understand the #inputValue message. This can be hacked by adding a protocol-extension that will return "self value" then, but it's not very nice. 2) Some #value messages are sent from methods in AST-Core package, and I don't know how we should deal with that. Any ideas for solutions? Uko -- View this message in context: http://moose-dev.97923.n3.nabble.com/Failing-PetitSmalltalk-tests-tp4026442.html Sent from the moose-dev mailing list archive at Nabble.com. _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Failing PetitSmalltalk tests
|
|
Hi,
There was a missing version from Lukas' repository that got lost during the migration. I thought I copied it afterwards to STHub, but perhaps something went wrong because I did not find it anymore in the repository. I copied it now and merged it with your latest version. It solved most of the problems, but there are still 7 failures and 6 errors. I do not quite figure out where the issues come from because these tests work fine in Pharo 1.4. Could you take a look now? Cheers, Doru On Mar 12, 2013, at 2:01 PM, Yuriy Tymchuk <[hidden email]> wrote: > Hi. > > In Pharo 2.0 PetitSmalltalk tests are failing. > > Firs of all there are real problems with parsing as some test fail because > parser finishes to parse before the end of the stream. Right now I don't > know what's causing the problem and how to solve it. > > Then there is a ton of deprecation warnings as PPToken>>#value became > deprecated and it's recommended to use #inputValue. It's now hard to fix > that, but there are two problems. > > 1) At places where #value message is sent, not only PPToken is used as a > receiver, but sometimes RBToken is used instead. And RBToken does not > understand the #inputValue message. This can be hacked by adding a > protocol-extension that will return "self value" then, but it's not very > nice. > > 2) Some #value messages are sent from methods in AST-Core package, and I > don't know how we should deal with that. > > Any ideas for solutions? > Uko > > > > -- > View this message in context: http://moose-dev.97923.n3.nabble.com/Failing-PetitSmalltalk-tests-tp4026442.html > Sent from the moose-dev mailing list archive at Nabble.com. > _______________________________________________ > Moose-dev mailing list > [hidden email] > https://www.iam.unibe.ch/mailman/listinfo/moose-dev -- www.tudorgirba.com "From an abstract enough point of view, any two things are similar." _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
|
|
Super. Now the only failing tests are the ones that I've mentioned at the
beginning. Maybe in Pharo 2.0 something changed with streams, but all the failing test have the same problem: parser finishes parsing before the end of the string. Maybe Lukas can take a look at it, because I'm quite new to the PP implementation -- View this message in context: http://moose-dev.97923.n3.nabble.com/Failing-PetitSmalltalk-tests-tp4026442p4026467.html Sent from the moose-dev mailing list archive at Nabble.com. _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Failing PetitSmalltalk tests
|
|
Hi,
I think Lukas does not have time. Could you take a closer look? Cheers, Doru On Mar 13, 2013, at 11:07 AM, Yuriy Tymchuk <[hidden email]> wrote: > Super. Now the only failing tests are the ones that I've mentioned at the > beginning. Maybe in Pharo 2.0 something changed with streams, but all the > failing test have the same problem: parser finishes parsing before the end > of the string. Maybe Lukas can take a look at it, because I'm quite new to > the PP implementation > > > > -- > View this message in context: http://moose-dev.97923.n3.nabble.com/Failing-PetitSmalltalk-tests-tp4026442p4026467.html > Sent from the moose-dev mailing list archive at Nabble.com. > _______________________________________________ > Moose-dev mailing list > [hidden email] > https://www.iam.unibe.ch/mailman/listinfo/moose-dev -- www.tudorgirba.com "Being happy is a matter of choice." _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Failing PetitSmalltalk tests
|
|
In reply to this post by Uko2
Hi I was checking at the code quickly and there is problem that is related to unary expression parsing. Something like this: whatever + 1. variable := 2+2
will be recognized as: "whatever" "+" "1. variable" so the second term is not just "1" but includes everything up to ":". Did something changed with unary expression evaluation? I tried to play with it but I couldn't fix it. I will investigate further tomorrow. Cheers, Fabrizio 2013/3/13 Yuriy Tymchuk <[hidden email]> Super. Now the only failing tests are the ones that I've mentioned at the _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
|
|
Hoorah. Busted 8 failures in one swipe: https://ci.inria.fr/moose/job/moose-latest-dev-4.8/86/testReport/
So now about details. After Fabrizio's example it wasn't hard to find out that "1." is treated as a number literal and "1. variable" is treated as a message send to a literal. After a fast talk with Marcus it turned out that Number class>>readFrom: started to use ExtendedNumberParser instead of SqNumberParser. So to fix the problem I've changed "Number readFrom: stream" to "(SqNumberParser on: stream) nextNumber" in a number parser of PPSmalltalkGrammar. Cheers, Uko On 14 бер. 2013, at 21:58, Fabrizio Perin-3 [via moose-dev] <[hidden email]> wrote:
View this message in context: Re: Failing PetitSmalltalk tests Sent from the moose-dev mailing list archive at Nabble.com. _______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Failing PetitSmalltalk tests
|
|
Excellent catch! You see, Yuriy: never underestimate the power of sticking to a problem. It is not mastery that solves the problem. It is solving the problem that leads to mastery.
Keep it up. Cheers, Doru On Mon, Mar 18, 2013 at 11:22 PM, Yuriy Tymchuk <[hidden email]> wrote: Hoorah. Busted 8 failures in one swipe: https://ci.inria.fr/moose/job/moose-latest-dev-4.8/86/testReport/ "Every thing has its own flow"
_______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
Re: Failing PetitSmalltalk tests
|
|
+1
On Tue, Mar 19, 2013 at 7:07 AM, Tudor Girba <[hidden email]> wrote:
_______________________________________________ Moose-dev mailing list [hidden email] https://www.iam.unibe.ch/mailman/listinfo/moose-dev |
«
Return to Moose
|
1 view|%1 views
| Free forum by Nabble | Edit this page |