[R-pkg-devel] Package accepted, then an error to fix by 3/22

[Michael Barr]

Thanks Simon - that definitely was a possible explanation. However I�ve ruled it out as the explanation - the same record comes through in one_rec since I last ran during initial development. I and fairly confident that the explanation must involve a discrepancy between the check process for a submission (and prior to submission) and the auto check after acceptance.

I don�t know how I raise this issue for investigation so I�m just adding back to the list in case anyone has an interest.

Thanks
Mike


________________________________
From: Simon Urbanek <simon.urbanek using R-project.org>
Sent: Tuesday, March 8, 2022 8:00 PM
To: Michael Barr <mike using bikeactuary.com>
Subject: Re: [R-pkg-devel] Package accepted, then an error to fix by 3/22



> On Mar 9, 2022, at 1:54 PM, Simon Urbanek <simon.urbanek using R-project.org> wrote:
>
> Michael,
>
> responses inline.
>
>
>> On Mar 9, 2022, at 1:29 PM, Michael Barr <mike using bikeactuary.com> wrote:
>>
>> Hi - thanks for the reply. Yes I built this vignette probably over 150 times through the course of working out other NOTES, via rhub::check_for_cran, devtools::check, devtools::build_vignette, R CMD CHECK, some other devtools function which checks for win devel the exact name of which escapes me atm�
>>
>> Also this made it past the CRAN submission process, newbies pipeline == higher touch as I understand.
>>
>> I�m curious why I make it past so many different checks prior to publishing.
>
>
> I would suspect it's because the results in your vignette are not fixed, but rely on web queries, so it just so happens that the first entry that you print changed in the NIH database and now triggers the problem. This issue has not been picked up because presumably the previous result was shorter.
>
>
>
>> Not surprised it is latex related though. I will look up the character limit for latex to fix this.
>>


I forgot to add that this has rally nothing to do with latex limits - it's simply a bug in the example that tries to produce an insanely long line - I don't think it makes any sense to try to print an abstract as one line. I would suggest either use str() just to illustrate the idea or if you really want to print it, you have to split it into lines and use cat()  [or let knitr soft-wrap the output].

Cheers,
Simon



	[[alternative HTML version deleted]]