gepsio
gepsio copied to clipboard
JeffFerguson
Test case ExecuteXBRLCONF20141210Testcases is failing
Test case is failing on develop branch following v2.1.0.14 :
Message: Assert.Fail failed. Instance: 320-00-BindCalculationInferPrecision-instance.xbrl Name: v00-CalculationWithPrecision3-Consistent Description: 320.00 - Valid
Validation Error Type: JeffFerguson.Gepsio.SummationConceptValidationError Validation Error Description: Fact A is based on an element that is named as the summation concept in a calculation link. The fact's value, after rounding, is 1600; however, the sum of the values of the contributing concepts, after rounding, is 1000. These values do not match; therefore, the rule specified by the fact's calculation link has been broken.
This is by design, and why Gepsio is a 2.1.0.* and not a 2.1.1.0 version (the initial 2.1 matches the version of the XBRL spec that is supports). The vision for Gepsio's versioning -- and I apologize for not making this more clear -- is that the versioning stays 2.1.0.* until it can pass all of the tests in the XBRL conformance suite. Once it is 2.1 compliant, then my idea was to move Gepsio's versioning to 2.1.1.0. I will continue to work on compatibility issues until all of the conformance tests pass.
Does that help clarify things?
Yes, it's clear now. I was a bit confused because I'm used to running unit tests as regression tests. I occasionaly do TDD but tag failing "work in progress" tests as to be ignored until work is done.
A list of yet unsupported XBRL specs would be helpfull for people to contribute.
On my branch "29", I sorted out passing from failing conformance tests. If you find it usefull I can make it clean and submit a pull request ?
I'm glad you called it out. I added a bit to the Wiki page at https://github.com/JeffFerguson/gepsio/wiki/Version-Numbering-Scheme to attempt to explain this further.
Tell me more about this branch "29" that you have. Do you mean a branch for Issue 29?
Yes, originally it was a branch to fix issue 29. I thought I would only have a few failing tests due to my configuration.
For this moment, I only managed to split the big bulk test into smaller groups of per category tests. I also changed code in order to log and continue in case of failing test in a batch. Finally, I tagged passing and failing test cases to sort things out.
Thank you! I'd be interested in taking a look at how you have organized things. Is there a pull request for it yet?
Looks like this test case should be fixed.
INVALID TEST CASE ......\XBRL-CONF-2014-12-10\Common\300-instance\301-04-IdScopeContextRefToUnit.xml INVALID example:fixedAssets precision="4" **contextRef="u1"** unitRef="u1"5000</example:fixedAssets> VALID example:fixedAssets precision="4" **contextRef="ci1"** unitRef="u1"5000</example:fixedAssets>
