-
Notifications
You must be signed in to change notification settings - Fork 1
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
Inconsistency of checks? #32
Comments
Oh yes, another very good question! @assignUser You may well have better suggestions than me, so please chime in. The meta-problem is #30, acknowledging that we need to restructure the documentation here. This issue indicates a clear need to explain the process for understanding failed checks. The action generates full information, but can't put it all directly on the workflow output. It nevertheless creates an "artefact" with all the information you need. To see it you need to:
Alternatively, just run the checks locally as I'll keep this issue open until we've clearly documented this general workflow. |
@mpadge maybe we should switch |
I've tried that, but it's not really intelligible that way - it's generally hundreds of lines of guff with very ilttle visual structure. Would it be possible to put a simple note in the workflow telling people to download the artefacts for details of failures? |
In my testing it had the same html summary fold out thingies... so it was a bit longer but not fa full CVS-receipt... let me check |
@mpadge like this: assignUser/octolog#14 this is 3 months old though so maybe something changed with the output? |
Oh yeah, sorry, my bad there. Yes, we definitely need jobs:
check:
runs-on: ubuntu-latest
steps:
- uses: ropensci-review-tools/pkgcheck-action@main
with:
summary-only: false Forget what i said above about artefacts; you don't need to worry about that. |
What I have done so far:
- changed the pkgcheck.yaml as indicated above ... added summary-only:
false (worked fine .. in my opinion as a newbie, this should be the default)
However, I still encounter differences:
On my local machine summary(checks) give me the following output on my
console:
-- concstats 0.1.2
---------------------------------------------------------------------------------------------------
√ Package is already on CRAN.
√ has a 'codemeta.json' file.
√ has a 'contributing' file.
√ uses 'roxygen2'.
√ 'DESCRIPTION' has a URL field.
√ 'DESCRIPTION' has a BugReports field.
√ Package has at least one HTML vignette
√ All functions have examples.
x Package has no continuous integration checks.
√ Package coverage is 79.6%.
√ R CMD check found no errors.
√ R CMD check found no warnings.
i Current status:
x This package is not ready to be submitted.
checks_to_markdown(checks, render = TRUE)
processing file: pkgcheck64c494d2745.Rmd (as an
attached file)
.................................................................................................................................
First, this structure is consistent with the README, but not with the
output file from "checks_to_markdown()
- the output (html) from the local machine has no "4. other checks" ...
(see attached file).
- the output (local machine, console) has no check "Function names are
duplicated in other packages" (also not mentioned in README)
- collateral: checks_to_markdown(.., render = TRUE) renders html, but also
(each and every line) on (my) the console.
( I don't know if this is an open ISSUE)
Second, the local output (console and html) throws an error for "CI check"
(see above), but on GITHUB (bot) the test passes ..
On GITHUB (bot) appear a test for "Function names ... " (failed) which does
not appear for the local check (console and html)
All this is somehow confusing ..
Well, in my particular case I suppose that I have to change/rename all my
function names? (and examples, documentation ... )
Something like this .... concstats_hhi, concstats_gini, ...?
[image: Mailtrack]
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
Remitente
notificado con
Mailtrack
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
17/06/22,
20:28:26
El vie, 17 jun 2022 a las 6:56, mark padgham ***@***.***>)
escribió:
… Oh yeah, sorry, my bad there. Yes, we definitely need summary_only: false.
@schneiderpy <https://github.com/schneiderpy> you just need to follow this
example from the docs
<https://github.com/ropensci-review-tools/pkgcheck-action#workflow-parameters>
and you'll get the full explanations of your failing checks in the issue:
jobs:
check:
runs-on: ubuntu-latest
steps:
- uses: ***@***.***
with:
summary-only: false
Forget what i said above about artefacts; you don't need to worry about
that.
—
Reply to this email directly, view it on GitHub
<#32 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ASTA3FYNUPQW4ZWJXZHKJB3VPRKWXANCNFSM5ZASBJNA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
---------- Forwarded message ---------
De: Andreas Schneider ***@***.***>
Date: vie, 17 jun 2022 a las 20:44
Subject: Re: [ropensci-review-tools/pkgcheck-action] Inconsistency of
checks? (Issue #32)
To: ropensci-review-tools/pkgcheck-action <
***@***.***>
Cc: ropensci-review-tools/pkgcheck-action <
***@***.***>, Mention ***@***.***>
What I have done so far:
- changed the pkgcheck.yaml as indicated above ... added summary-only:
false (worked fine .. in my opinion as a newbie, this should be the default)
However, I still encounter differences:
On my local machine summary(checks) give me the following output on my
console:
-- concstats 0.1.2
---------------------------------------------------------------------------------------------------
√ Package is already on CRAN.
√ has a 'codemeta.json' file.
√ has a 'contributing' file.
√ uses 'roxygen2'.
√ 'DESCRIPTION' has a URL field.
√ 'DESCRIPTION' has a BugReports field.
√ Package has at least one HTML vignette
√ All functions have examples.
x Package has no continuous integration checks.
√ Package coverage is 79.6%.
√ R CMD check found no errors.
√ R CMD check found no warnings.
i Current status:
x This package is not ready to be submitted.
checks_to_markdown(checks, render = TRUE)
processing file: pkgcheck64c494d2745.Rmd (as an
attached file)
.................................................................................................................................
First, this structure is consistent with the README, but not with the
output file from "checks_to_markdown()
- the output (html) from the local machine has no "4. other checks" ...
(see attached file).
- the output (local machine, console) has no check "Function names are
duplicated in other packages" (also not mentioned in README)
- collateral: checks_to_markdown(.., render = TRUE) renders html, but also
(each and every line) on (my) the console.
( I don't know if this is an open ISSUE)
Second, the local output (console and html) throws an error for "CI check"
(see above), but on GITHUB (bot) the test passes ..
On GITHUB (bot) appear a test for "Function names ... " (failed) which does
not appear for the local check (console and html)
All this is somehow confusing ..
Well, in my particular case I suppose that I have to change/rename all my
function names? (and examples, documentation ... )
Something like this .... concstats_hhi, concstats_gini, ...?
[image: Mailtrack]
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
Remitente
notificado con
Mailtrack
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
17/06/22,
20:28:26
El vie, 17 jun 2022 a las 6:56, mark padgham ***@***.***>)
escribió:
Oh yeah, sorry, my bad there. Yes, we definitely need summary_only: false.
@schneiderpy <https://github.com/schneiderpy> you just need to follow this
example from the docs
<https://github.com/ropensci-review-tools/pkgcheck-action#workflow-parameters>
and you'll get the full explanations of your failing checks in the issue:
jobs:
check:
runs-on: ubuntu-latest
steps:
- uses: ***@***.***
with:
summary-only: false
Forget what i said above about artefacts; you don't need to worry about
that.
—
Reply to this email directly, view it on GitHub
<#32 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ASTA3FYNUPQW4ZWJXZHKJB3VPRKWXANCNFSM5ZASBJNA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
[image: Mailtrack]
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
Remitente
notificado con
Mailtrack
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
11/07/22,
10:54:44
|
@schneiderpy I'm not really sure what your questions are in the above? Please note that GitHub issues respect all markdown formatting, so you should please use that. In other words, please run The inconsistencies between standardised checks via The inconsistences between what you see directly on your console screen and the markdown-formatted versions are an unavoidably consequence of compromises necessary in trying to convey a lot of information in compact ways. The most complete version is always the markdown/html version generated by Finally, regarding function names, I would suggest in your case that yes, you ought to rename a lot of them as many of the names are quite generic. A standard package-specific prefix is the generally recommended approach. This is often some abbreviated form of a full package name, although in your case the function names are generally quite short, so yes, prefixing with |
Thank you for your prompt reply. I thought you were on vacation ...
Actually these are not questions .. it's more like an observation. But I'll
come to this later, if I'll still encounter these issues (using Github
issues?)
(e.g. just look at the readme file of the pkgcheck() package, section:
"what is checked" ... unique_fs_names is not listed, and so in the
summary() which appear on the console screen .... but appears in the html
version ..)
However, I have another question: I have almost changed all functions
(consistently prefixed all with concstats_xxx) ..
Am I right that I have to do a new CRAN / Zenodo submission? (prob. with a
new version number) .. at least that sounds somehow logical to me
Regards,
Andreas
[image: Mailtrack]
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
Remitente
notificado con
Mailtrack
<https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&>
12/07/22,
20:07:04
El mar, 12 jul 2022 a las 7:41, mark padgham ***@***.***>)
escribió:
… @schneiderpy <https://github.com/schneiderpy> I'm not really sure what
your questions are in the above? Please note that GitHub issues respect all
markdown formatting, so you should please use that. In other words, please
run checks_to_markdown(), and paste the markdown-formatted result
directly in any issue comments.
The inconsistencies between standardised checks via pkgcheck-action and
your local checks could be due to any number of things. If you're not
seeing local CI checks, that likely suggests that you don't have your
GitHub authorization token set up properly. That's explained on the
pkgcheck README <https://github.com/ropensci-review-tools/pkgcheck#setup>.
The inconsistences between what you see directly on your console screen
and the markdown-formatted versions are an unavoidably consequence of
compromises necessary in trying to convey a lot of information in compact
ways. The most complete version is always the markdown/html version
generated by checks_to_markdown(), while the screen version will
generally be only an abbreviated version of that.
Finally, regarding function names, I would suggest in your case that yes,
you ought to rename a lot of them as many of the names are quite generic. A
standard package-specific prefix is the generally recommended approach.
This is often some abbreviated form of a full package name, although in
your case the function names are generally quite short, so yes, prefixing
with concstats_ would seem like a reasonable approach. Note also
ropensci-review-tools/pkgcheck#144
<ropensci-review-tools/pkgcheck#144> - once
that's been implemented these namespace conflicts won't necessarily be an
outright fail, but again in the case of your package i would suspect that
reviewers would likely suggest what I'm suggesting here regardless.
—
Reply to this email directly, view it on GitHub
<#32 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ASTA3F655IZ2T7H5UZQWY2TVTVKU5ANCNFSM5ZASBJNA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
That's entirely up to you, and not related either to this issue, nor to |
I am stuck with a "failed check".
Checking my package concstats with pkgcheck on my local machine a came across the following issue:
In the summary popped up a failed check due to .. continuous integration, however, I have github actions for rmd check and (now) for pkgcheck
On github the bot (of pkgcheck?) the continuous integration check passed, however, it popped up a failed check for .. "Function names are duplicated in other packages"? I don't know where to look at or what function names are duplicated (I cannot check thousands of packages). This failed check does not appear on the local check (console).
The text was updated successfully, but these errors were encountered: