[YANGTOOLS-789] Errors on augments to nodes in groupings Created: 19/Jun/17  Updated: 10/Apr/22  Resolved: 17/Jul/17

Status: Resolved
Project: yangtools
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug
Reporter: Peter Verthez Assignee: Peter Kajsa
Resolution: Cannot Reproduce Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

Attachments: Zip Archive augments.zip     File error.txt.gz    
External issue ID: 8711

 Description   

On the parsing of the models of one of our devices, we are getting errors on augments not finding their targets. This is always in combination with nested groupings, where the target is somewhere in the nested grouping.

I've attached a minimal model to this bug report. It is the original model that failed (with some IETF and BBF models), but I've removed everything that is not related to the problem, to have a minimal model to investigate with. I couldn't see an issue with the models however.

Note also that in some occasions the parsing succeeds, and in others it fails. I've not been able to find the environment difference between the 2 systems that would influence this.

I've also attached a file with debug logging from the error.

 Comments   
Comment by Peter Verthez [ 19/Jun/17 ]

Attachment augments.zip has been added with description: minimized model showing the bug

Comment by Peter Verthez [ 19/Jun/17 ]

Attachment error.txt.gz has been added with description: debug logging

Comment by Peter Kajsa [ 06/Jul/17 ]

Hi,

I have been trying to replicate this bug with the attached yang models, however without any success. I run the test in loop thousand times, but SchemaContext is always created successfully. Please are you sure, there is not something else what has influence on the bug occurrence (e.g. loading of an another revision of a module under some conditions etc.) or something what is missing in the attached set of yang models ?

Thanks.

Comment by Peter Verthez [ 06/Jul/17 ]

We have several modules with independent models.

However, for this parsing we are not using the SharedSchemaRepository, but

CrossSourceStatementReactor.BuildAction reactor = YangInferencePipeline.RFC6020_REACTOR.newBuild(supportedFeatures);
reactor.addSources(moduleSources);
schemaContext = reactor.buildEffective();

So I was assuming that there could be no interference between the modules. Or is this not correct? Can the parsing result depend on earlier parse results? If so, how can we avoid that?

Comment by Peter Kajsa [ 07/Jul/17 ]

The code you posted is also standard way how to use parser, so I think there should not be any interference between models. Parsing result should not depend on earlier parse results, otherwise there is a bug.

I will try to perform further investigation. If you find some new info that could help to replicate the bug, please let me know.

Thanks.

Comment by Peter Verthez [ 10/Jul/17 ]

OK, I will investigate further from my end as well.

Comment by Peter Verthez [ 11/Jul/17 ]

I and some other people have tried to investigate the issue, but apparently we can't reproduce it anymore either (although it was reproduced by multiple people before). We don't exactly know why however.

But in any case, I think you can close this issue and stop further investigating it. If it reoccurs, we will investigate first on our side.

Comment by Peter Kajsa [ 17/Jul/17 ]

Ok, thanks.

Generated at Wed Feb 07 20:54:19 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.