-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
t/test62-subcaller1.t fails on older perls on Windows #153
Comments
Hi @jkeenan, I am currently preparing a CondaForge package recipe for
As a side note, on my system Perl,
If I can do anything to help find the cause of this issue, please let me know! |
I'm at a loss as to how to respond to this. I don't know what CondaForge is. I am aware of Docker but don't use it myself and I generally consider it outside of my remit to debug failures there that are not occuring outside of Docker. In this and other bug tickets we've discussed the Devel-NYTProf test suite's hypersensitivity to small changes in one's environment. By that we mean that the test suite is chock full of comparisons between a test run and a pre-computed corpus of data. With small changes in one's environment -- and I suspect changes due to Docker would be non-small -- you get output from the test run that is different from the corpus (and is therefore considered a test failure) without necessarily being wrong. (Indeed, the only reason I am involved in Devel-NYTProf maintenance at all is that changes in the Perl 5 core distribution sometimes trigger failures in Devel-NYTProf's test suite.)
The only thing I can suggest is to run the failing test file manually (i.e., outside your CI pipeline) and examine the results for plausibility. If the results are plausible, then simply skip the file within the pipeline. |
Hey @jkeenan, Some background on CondaForge: it is is a general-purpose package repository for the package manager Conda, which allows users to install arbitrary versions of a software into various virtual environments among other programs, and to activate, deactivate, and manage these environments and so on. This is useful if you need multiple versions of the same software, or if you want to provide a defined working environment to other people or for another computer, e.g. to reproduce scientific results. The purpose of my report was merely to inform you that this issue with the test exists in certain environments, and to provide some more information that may help to identify the component of the test suite that suffers from this hypersensitivy, as you call it. I do not expect you to provide personal support for my very specific use case, I was just trying to help. If I can do anything to support you in that matter, please let me know. |
Thanks very much for taking the time to provide your original report and this response. |
CPANtesters reports indicate that what, before Devel-NYTProf-6.07, was
t/test62-subcaller1.t
(and what is nowt/test62-subcaller1-a.t
) can experience test failures on older versions of Strawberry Perl. Here is output from a report generated on a Strawberry Perl 5.14.4 with NYTProf version 6.06.These error reports all appear to be emanating from a single CPANtester installation which first began generating reports when NYTProf was at version 6.05. Reports emanating from the same reporter start to PASS with perl-5.20.3.
The CPANtester has not provided a name, so it will be difficult to follow up on these reports. Other CPANtesters on Windows are reporting PASSes.
The text was updated successfully, but these errors were encountered: