Running some code at fogbugz CI
Hello, I had this patch (for issue 11992) for a while now at the bug system. The problem I'm having is that in order to to run the test suite properly I need to run some code before merging the slice and then just before running the test suite. And as a consequence, I'm not notified whenever some other patch that gets integrated breaks my code. I have read that you can use metacello to configure post and preLoad hooks for a project. I'm not sure if this could work for the fogbugz CI, also since you need to add a class in order to use metacello it would probably get integrated into the image which is not what I intend. Is there then some easy way so run some code before the CI merges the patch and just before the test suite runs?. PS1: Sorry is the answer for this turns out being pretty obvious, I'm not that familiar with metacello. PS2: The issue haven't receive a lot of attention probably because it's been at 'Work Needed' state since forever, maybe it would help if the test suite passes. PS3: Here's a link to the issue https://pharo.fogbugz.com/f/cases/11992/Remove-tokens-from-the-AST-Core-Node... . Thanks, Pablo
Hi Pablo, I presume you have not noticed when you <Edit> the Fogbugz issue, there are two header fields "Preload" & "Postload" , which may satisfy your need. Add your code just like you would type it into a Workspace. cheers -ben On Tue, Feb 10, 2015 at 9:19 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
Hello,
I had this patch (for issue 11992) for a while now at the bug system. The problem I'm having is that in order to to run the test suite properly I need to run some code before merging the slice and then just before running the test suite. And as a consequence, I'm not notified whenever some other patch that gets integrated breaks my code.
I have read that you can use metacello to configure post and preLoad hooks for a project. I'm not sure if this could work for the fogbugz CI, also since you need to add a class in order to use metacello it would probably get integrated into the image which is not what I intend.
Is there then some easy way so run some code before the CI merges the patch and just before the test suite runs?.
PS1: Sorry is the answer for this turns out being pretty obvious, I'm not that familiar with metacello. PS2: The issue haven't receive a lot of attention probably because it's been at 'Work Needed' state since forever, maybe it would help if the test suite passes. PS3: Here's a link to the issue
https://pharo.fogbugz.com/f/cases/11992/Remove-tokens-from-the-AST-Core-Node... .
Thanks, Pablo
I think I tried that before but it didn't work, maybe I did something wrong then. 2015-02-10 11:23 GMT-03:00 Ben Coman <btc@openinworld.com>:
Hi Pablo,
I presume you have not noticed when you <Edit> the Fogbugz issue, there are two header fields "Preload" & "Postload" , which may satisfy your need. Add your code just like you would type it into a Workspace.
cheers -ben
On Tue, Feb 10, 2015 at 9:19 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
Hello,
I had this patch (for issue 11992) for a while now at the bug system. The problem I'm having is that in order to to run the test suite properly I need to run some code before merging the slice and then just before running the test suite. And as a consequence, I'm not notified whenever some other patch that gets integrated breaks my code.
I have read that you can use metacello to configure post and preLoad hooks for a project. I'm not sure if this could work for the fogbugz CI, also since you need to add a class in order to use metacello it would probably get integrated into the image which is not what I intend.
Is there then some easy way so run some code before the CI merges the patch and just before the test suite runs?.
PS1: Sorry is the answer for this turns out being pretty obvious, I'm not that familiar with metacello. PS2: The issue haven't receive a lot of attention probably because it's been at 'Work Needed' state since forever, maybe it would help if the test suite passes. PS3: Here's a link to the issue
https://pharo.fogbugz.com/f/cases/11992/Remove-tokens-from-the-AST-Core-Node... .
Thanks, Pablo
I think for a while the monkey did not run them, but it was fixed. Have another ago :) On Tue, Feb 10, 2015 at 10:27 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
I think I tried that before but it didn't work, maybe I did something wrong then.
2015-02-10 11:23 GMT-03:00 Ben Coman <btc@openinworld.com>:
Hi Pablo,
I presume you have not noticed when you <Edit> the Fogbugz issue, there are two header fields "Preload" & "Postload" , which may satisfy your need. Add your code just like you would type it into a Workspace.
cheers -ben
On Tue, Feb 10, 2015 at 9:19 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
Hello,
I had this patch (for issue 11992) for a while now at the bug system. The problem I'm having is that in order to to run the test suite properly I need to run some code before merging the slice and then just before running the test suite. And as a consequence, I'm not notified whenever some other patch that gets integrated breaks my code.
I have read that you can use metacello to configure post and preLoad hooks for a project. I'm not sure if this could work for the fogbugz CI, also since you need to add a class in order to use metacello it would probably get integrated into the image which is not what I intend.
Is there then some easy way so run some code before the CI merges the patch and just before the test suite runs?.
PS1: Sorry is the answer for this turns out being pretty obvious, I'm not that familiar with metacello. PS2: The issue haven't receive a lot of attention probably because it's been at 'Work Needed' state since forever, maybe it would help if the test suite passes. PS3: Here's a link to the issue
https://pharo.fogbugz.com/f/cases/11992/Remove-tokens-from-the-AST-Core-Node...
.
Thanks, Pablo
2015-02-10 15:27 GMT+01:00 Pablo Herrero <pablodherrero@gmail.com>:
I think I tried that before but it didn't work, maybe I did something wrong then.
I changed the compiler to the old one. load your slice changed the compiler back to opal. - But now, runing OC* tests half of them fail. - Executing OpalCompiler recompileAll throws some errors and the images is really slow, constantly writing errors to the debugfile: [31mMessageNotUnderstood: RBArgumentNode>>start: [0mRBArgumentNode(Object)>>doesNotUnderstand: #start: RBArgumentNode(RBVariableNode)>>named:start: RBArgumentNode class(RBVariableNode class)>>named:start: RBArgumentNode class(RBVariableNode class)>>identifierNamed:at: RBExplicitVariableParser>>parsePrimitiveIdentifier RBExplicitVariableParser(RBParser)>>parseVariableNode RBExplicitVariableParser(RBParser)>>parseKeywordPattern RBExplicitVariableParser(RBParser)>>parseMessagePattern
2015-02-10 11:23 GMT-03:00 Ben Coman <btc@openinworld.com>:
Hi Pablo,
I presume you have not noticed when you <Edit> the Fogbugz issue, there are two header fields "Preload" & "Postload" , which may satisfy your need. Add your code just like you would type it into a Workspace.
cheers -ben
On Tue, Feb 10, 2015 at 9:19 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
Hello,
I had this patch (for issue 11992) for a while now at the bug system. The problem I'm having is that in order to to run the test suite properly I need to run some code before merging the slice and then just before running the test suite. And as a consequence, I'm not notified whenever some other patch that gets integrated breaks my code.
I have read that you can use metacello to configure post and preLoad hooks for a project. I'm not sure if this could work for the fogbugz CI, also since you need to add a class in order to use metacello it would probably get integrated into the image which is not what I intend.
Is there then some easy way so run some code before the CI merges the patch and just before the test suite runs?.
PS1: Sorry is the answer for this turns out being pretty obvious, I'm not that familiar with metacello. PS2: The issue haven't receive a lot of attention probably because it's been at 'Work Needed' state since forever, maybe it would help if the test suite passes. PS3: Here's a link to the issue
https://pharo.fogbugz.com/f/cases/11992/Remove-tokens-from-the-AST-Core-Node...
.
Thanks, Pablo
2015-02-10 12:17 GMT-03:00 Nicolai Hess <nicolaihess@web.de>:
- But now, runing OC* tests half of them fail.
- Executing OpalCompiler recompileAll throws some errors and the images is really slow, constantly writing errors to the debugfile:
[31mMessageNotUnderstood: RBArgumentNode>>start: [0mRBArgumentNode(Object)>>doesNotUnderstand: #start: RBArgumentNode(RBVariableNode)>>named:start: RBArgumentNode class(RBVariableNode class)>>named:start: RBArgumentNode class(RBVariableNode class)>>identifierNamed:at: RBExplicitVariableParser>>parsePrimitiveIdentifier RBExplicitVariableParser(RBParser)>>parseVariableNode RBExplicitVariableParser(RBParser)>>parseKeywordPattern RBExplicitVariableParser(RBParser)>>parseMessagePattern
Ok, I'll take a look at that.
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway). 2015-02-10 12:32 GMT-03:00 Pablo Herrero <pablodherrero@gmail.com>:
Yeap, I evidently something got broken again recently I'll try to fix it ASAP.
On 10 Feb 2015, at 23:20, Pablo Herrero <pablodherrero@gmail.com> wrote:
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway).
I really need to review your change⦠sometimes I wonder⦠so much todo and I feel I could work double and it would still not be enough. Marcus
I understand, this patch could be potentially dangerous given the particular classes it needs to modify. If I can make the CI let me know when is broken again at least I can prevent it to stale. 2015-02-11 7:15 GMT-03:00 Marcus Denker <marcus.denker@inria.fr>:
On 10 Feb 2015, at 23:20, Pablo Herrero <pablodherrero@gmail.com> wrote:
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway).
I really need to review your change⦠sometimes I wonder⦠so much todo and I feel I could work double and it would still not be enough.
Marcus
On 11 Feb 2015, at 13:05, Pablo Herrero <pablodherrero@gmail.com> wrote:
I understand, this patch could be potentially dangerous given the particular classes it needs to modify. If I can make the CI let me know when is broken again at least I can prevent it to stale.
Ci is down for today⦠hopefully later today it will work again.
2015-02-11 7:15 GMT-03:00 Marcus Denker <marcus.denker@inria.fr>:
On 10 Feb 2015, at 23:20, Pablo Herrero <pablodherrero@gmail.com> wrote:
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway).
I really need to review your change⦠sometimes I wonder⦠so much todo and I feel I could work double and it would still not be enough.
Marcus
One step at the time and pushing regularly. Stef
On 10 Feb 2015, at 23:20, Pablo Herrero <pablodherrero@gmail.com> wrote:
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway).
I really need to review your change⦠sometimes I wonder⦠so much todo and I feel I could work double and it would still not be enough.
Marcus
Ok, now is running again but failing, I don't really understand what's going on neither what's the problem. Also, how can you tell if the pre/posLoad scripts actually ran? Here's the report: https://ci.inria.fr/pharo/view/4.0-Analysis/job/Pharo-4.0-Issue-Validator/24... 2015-02-12 5:06 GMT-03:00 stepharo <stepharo@free.fr>:
One step at the time and pushing regularly.
Stef
On 10 Feb 2015, at 23:20, Pablo Herrero <pablodherrero@gmail.com> wrote:
Ok, thanks Nicolas for letting me now, the test suite should be green now. BTW, this is the list of packages I'm testing for this: AST-Core, AST-Interpreter, Collections-Strings, NodeNavigation, OpalCompiler, OpalDecompiler, Refactoring, SmartSuggestions and Shout. (There could be more that are also important anyway).
I really need to review your change⦠sometimes I wonder⦠so much todo and I feel I could work double and it would still not be enough.
Marcus
2015-02-13 12:34 GMT-03:00 Pablo Herrero <pablodherrero@gmail.com>:
Ok, now is running again but failing, I don't really understand what's going on neither what's the problem. Also, how can you tell if the pre/posLoad scripts actually ran?
Here's the report: https://ci.inria.fr/pharo/view/4.0-Analysis/job/Pharo-4.0-Issue-Validator/24...
Just in case: it does work fine if I run the test suite with a recent image on my laptop.
On Fri, Feb 13, 2015 at 5:07 PM, Pablo Herrero <pablodherrero@gmail.com> wrote:
2015-02-13 12:34 GMT-03:00 Pablo Herrero <pablodherrero@gmail.com>:
Ok, now is running again but failing, I don't really understand what's going on neither what's the problem. Also, how can you tell if the pre/posLoad scripts actually ran?
Here's the report:
https://ci.inria.fr/pharo/view/4.0-Analysis/job/Pharo-4.0-Issue-Validator/24...
Just in case: it does work fine if I run the test suite with a recent image on my laptop.
I think the monkey still does not execute pre and postscripts... I have put the issue on "review needed" with the noMonkey tag.
participants (5)
-
Ben Coman -
Marcus Denker -
Nicolai Hess -
Pablo Herrero -
stepharo