[Bioc-devel] R CMD build error for 3.10 (but not 3.11)

Tue Mar 10 13:21:05 CET 2020

I don't think you were pushing to the devel builders correctly.   The last commit we have on the builder is Sat Feb 29th.

commit b810b6a81198438d6ad26b7d86b14eece00dfa59
Author: lrutter <lindsayannerutter using gmail.com>
Date:   Sat Feb 29 23:44:08 2020 +0900

    Fix version issue

So you never actually pushed the changes to the git.biocnductor  upstream server.

Based on what you have below please do the following and if you have any ERRORs please paste the command with the output so we can better assist you.

git fetch --all
git pull upstream master

# make sure your version number is 1.3.2
# I'm not sure if the version number higher in the history will cause problems.  If you get an error about version number I suggest then doing 1.3.6
# which is one higher than it looks like

git push upstream master

The upstream in these commands says to pull/push to the upstream remote which is pointed at the bioconductor git servers.
The master says to pull/push from the master branch of the given repository.

Cheers,

Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263

________________________________
From: L Rutter <lindsayannerutter using gmail.com>
Sent: Tuesday, March 10, 2020 4:11 AM
To: Shepherd, Lori <Lori.Shepherd using RoswellPark.org>
Cc: bioc-devel using r-project.org <bioc-devel using r-project.org>
Subject: Re: [Bioc-devel] R CMD build error for 3.10 (but not 3.11)

Dear Lori:

Thank you very much for your very helpful response. I am sorry to say that I still am having difficulty seeing errors I think I resolved locally get successfully pushed to my devel branch. I would really appreciate your feedback to help me possibly identify the problem.

First, I checked that my git remotes are set up correctly. Below is my output:

origin https://github.com/lindsayrutter/bigPint.git (fetch)
origin https://github.com/lindsayrutter/bigPint.git (push)
upstream git using git.bioconductor.org:packages/bigPint.git (fetch)
upstream git using git.bioconductor.org:packages/bigPint.git (push)

In an attempt to push changes to my devel branch, I updated my DESCRIPTION file to contain version 1.3.2. My last commit can be seen here<https://github.com/lindsayrutter/bigPint/commit/1f61951168d5ba35be6464c3f082e006ff2dd220>.

I waited a few days and see that my builder<http://bioconductor.org/checkResults/devel/bioc-LATEST/bigPint/malbec2-checksrc.html> is still failing. The ERROR seems to be coming from my R/convertSEPair.R script in the line:

vapply(colnames(oldData), function(x) strsplit(x, "[.]")[[1]][1])

I had updated this command, changing $B!H(Bvapply$B!I(B to $B!H(Blapply$B!I(B. It now reads:

lapply(colnames(oldData), function(x) strsplit(x, "[.]")[[1]][1])

I double-checked the R/convertSEPair.R file on my latest GitHub commit<https://github.com/lindsayrutter/bigPint/blob/1f61951168d5ba35be6464c3f082e006ff2dd220/R/convertSEPair.R> and indeed the $B!H(Bvapply$B!I(B command is no longer present.

I also note that my $B!H(BLast commit$B!I(B field on the builder website<http://bioconductor.org/checkResults/devel/bioc-LATEST/bigPint/malbec2-checksrc.html> is b810b6a from February 29, whereas my latest intended GitHub commit for the devel branch was 1f61951 from March 8.

I should also state that I had tried a few times and kept bumping up my version number. So, on March 7 (Commit c75da1e), I had version 1.3.5. In my latest commit, I reduced my version number to 1.3.2 just to check that the reason my devel branch was still failing was not simply because my z (x.y.z) number was increased by more than a value of 1. However, it seems my pushed changes on GitHub are not updating to the devel builder website.

I really thank you for sharing your advice with me. I wanted to ask if you may have any ideas on what I may be missing at this stage. Thank you again for your support.

Sincerely,
Lindsay

2020$BG/(B3$B7n(B3$BF|(B($B2P(B) 21:30 Shepherd, Lori <Lori.Shepherd using roswellpark.org<mailto:Lori.Shepherd using roswellpark.org>>:
It can take up to 24 from a push to be reflected on the build report.

For release 3.10 (RELEASE_3_10  branch)  we see a version number of 1.2.0
http://bioconductor.org/checkResults/release/bioc-LATEST/bigPint/
and
For devel 3.11 (master branch)  we see a version number of 1.3.1  as well
http://bioconductor.org/checkResults/devel/bioc-LATEST/bigPint/

The landing page http://bioconductor.org/packages/3.11/bioc/html/bigPint.html  will only be updated on a successful build/check of the package so we don't propagate broken packages to users. If the fixes did not allow for a clean build and check it would not be reflected.

Also remember that in order for a change to propagate you need a valid version bump.  So your release version 1.2.0  in order to take effect would have needed to be 1.2.1.  And for devel master branch you would need a version of 1.3.2.

Never apologize for asking questions.  We are here to help.  And we appreciate that you are taking the time to fix the package and learn how to push the changes up correct.

I'll answer your specific questions below:

Firstly make sure your   git remotes are set up correctly  where origin is your github repository and upstream is pointing to the bioconductor repositories http://bioconductor.org/developers/how-to/git/sync-existing-repositories/

origin  (your github)
origin  (your github)
upstream git using git.bioconductor.org:packages/bigPint.git (fetch)
upstream git using git.bioconductor.org:packages/bigPint.git (push)

1) Is it acceptable for me to push what is on my GitHub master branch (1.3.1) to the development branch of Bioconductor?
Yes. however you will want to bump the version number in the DESCRIPTION to 1.3.2 because we have version 1.3.1 on our builders as indicated  http://bioconductor.org/checkResults/devel/bioc-LATEST/bigPint/.  This page reflects 1.3.0 because 1.3.1 has not built without ERROR and we would not propagate broken packages to users  http://bioconductor.org/packages/3.11/bioc/html/bigPint.html.

2) Is it acceptable for me to likewise push what is on my GitHub master branch (1.3.1) to the release branch of Bioconductor? I am assuming I am only supposed to push the "cherry-pick" commits that resolve what had been causing an error originally? This may be difficult for me as I have since added many more changes to the master branch on GitHub after resolving that release error.
Yes although not recommended but we would rather have a working package than one ERRORing so if you don't think you can cherry-pick than yes.  The best option for that would be to checkout the RELEASE_3_10 branch and do a  git merge master.  You will have to change the version number in the DESCRIPTION to be consistent with release version numbers.  (1.2.1) AFTER the merge.

3) Is it possible to ask if there is anything else identifiably "out of line" with my setup (on my GitHub master, development branch, or release branch)? I am sure that once I complete this process successfully I will have be better positioned to critically think of what is happening - but right now I worry that due to my earlier mistakes, something may be "off" with my setup that will make it difficult for me to successfully follow any set of instructions.
I think you were on the right track so no further comments.
Some helpful hints :

  1.  git fetch --all will fetch all branches from your origin github and the upstream bioconductor location (this will be important to do after the next release in april to get the next RELEASE_3_11 branch)
  2.  it is always a good idea to do a   git pull upstream (branch name)  to pull any changes we have made. occasionally if we made an infrastructure change bioconductor will apply patches.  We also do the version bump changes at release time. So for your master devel branch   git pull upstream master  (git pull upstream RELEASE_3_10  for the current release)
  3.  If you do git merge  read the out message carefully and make sure there are no conflicts.  Conflicts means there is code that was changes in both files (locally and on the branch you want to merge)  so you will have to check the file to make sure you don't include merge conflicts -   The code conflicts are indicated with  >>>>>      and  <<<<<.  you delete those and keep the changes you want to commit.  Then add and commit like normal and git should give you messages about resolving conflicts.

Hope this helps.  Let me know if you run into any further trouble.

Cheers

Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263

________________________________
From: L Rutter <lindsayannerutter using gmail.com<mailto:lindsayannerutter using gmail.com>>
Sent: Tuesday, March 3, 2020 5:55 AM
To: Shepherd, Lori <Lori.Shepherd using RoswellPark.org>
Cc: bioc-devel using r-project.org<mailto:bioc-devel using r-project.org> <bioc-devel using r-project.org<mailto:bioc-devel using r-project.org>>
Subject: Re: [Bioc-devel] R CMD build error for 3.10 (but not 3.11)

Dear Lori:

Thank you for this helpful advice. I did successfully now manage to change my keys and name on the git.credentials app you suggested on February 25.

At that time, my development version was building but my release version was failing. So, I followed the instructions in the link you provided on February 3 (http://bioconductor.org/developers/how-to/git/bug-fix-in-release-and-devel/) to update my release version. I was able to also follow the "Sync existing repositories" tab at the beginning of the guideline.

After waiting for the updates, when I checked into my package (bigPint), I noticed now both the development (https://bioconductor.org/packages/devel/bioc/html/bigPint.html) and release (https://bioconductor.org/packages/release/bioc/html/bigPint.html) version are not building. I was surprised because I did not think I was making any changes to the development version through these instructions. I also notice my development version is 1.3.0, my release version is 1.2.0, and my GitHub master branch (https://github.com/lindsayrutter/bigPint/blob/master/DESCRIPTION) is version 1.3.1. It also seems that the version on my local computer reversed to 1.2.0.

I tentatively plan to:

1) Clone my GitHub master branch (version 1.3.1) back to my local computer (to make sure I do not lose many recent changes I have made to my package).
2) Try to push my most current version (1.3.1) to the development branch of Bioconductor.
3) Try to push changes to the release branch of Bioconductor.

I am pretty stuck even to the point where it is hard for me to formulate questions but below are a few questions that might help me identify the point of confusion:
1) Is it acceptable for me to push what is on my GitHub master branch (1.3.1) to the development branch of Bioconductor?
2) Is it acceptable for me to likewise push what is on my GitHub master branch (1.3.1) to the release branch of Bioconductor? I am assuming I am only supposed to push the "cherry-pick" commits that resolve what had been causing an error originally? This may be difficult for me as I have since added many more changes to the master branch on GitHub after resolving that release error.
3) Is it possible to ask if there is anything else identifiably "out of line" with my setup (on my GitHub master, development branch, or release branch)? I am sure that once I complete this process successfully I will have be better positioned to critically think of what is happening - but right now I worry that due to my earlier mistakes, something may be "off" with my setup that will make it difficult for me to successfully follow any set of instructions.

I apologize for reaching out about this headache of an issue now. I am just concerned I may not be able to resolve this in a timely manner with my current scope of knowledge.

Sincerely,
Lindsay

2020$BG/(B2$B7n(B25$BF|(B($B2P(B) 21:14 Shepherd, Lori <Lori.Shepherd using roswellpark.org<mailto:Lori.Shepherd using roswellpark.org>>:
Thank you for reaching out.   Have you tried checking your keys and changing your name on the git.credentials app?

https://git.bioconductor.org/BiocCredentials/login/?next=/BiocCredentials/

Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263

________________________________
From: L Rutter <lindsayannerutter using gmail.com<mailto:lindsayannerutter using gmail.com>>
Sent: Tuesday, February 25, 2020 1:55 AM
To: Shepherd, Lori <Lori.Shepherd using RoswellPark.org>
Cc: bioc-devel using r-project.org<mailto:bioc-devel using r-project.org> <bioc-devel using r-project.org<mailto:bioc-devel using r-project.org>>
Subject: Re: [Bioc-devel] R CMD build error for 3.10 (but not 3.11)

Dear Lori:

Thanks for the suggestions. I tried to follow the linked instructions on February 4 and February 9, but was unsuccessful both times (i.e. did not see the changes pushed to the release branch). I tried following the instructions again today. I notice that when I issue the command:

git merge upstream/master

I receive the message:

merge: upstream/master - not something we can merge

So, I clicked on the "Sync existing repositories" tab and checked that the instructions there were working. After adding the upstream branch as follows:

git remote add upstream git using git.bioconductor.org:packages/bigPint.git

When I then try to fetch all (git fetch --all), I receive the following error:

Fetching origin
Fetching upstream
FATAL: R any packages/bigpint lrutter DENIED by fallthru
(or you mis-spelled the reponame)
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
error: Could not fetch upstream

I changed my GitHub userName about one year ago from "lrutter" to "lindsayrutter". By doing "git remote -v", I can see that my origin branches are correctly at "lindsayrutter", but I am unable to see about my upstream branches on git.bioconductor. This is my output from "git remote -v":

origin https://github.com/lindsayrutter/bigPint (fetch)
origin https://github.com/lindsayrutter/bigPint (push)
upstream git using git.bioconductor.org:packages/bigpint.git (fetch)
upstream git using git.bioconductor.org:packages/bigpint.git (push)

Do you have any advice on how I can fix this issue with fetch upstream possibly expecting my older userName? Thank you for sharing advice.

Sincerely,
Lindsay

2020$BG/(B2$B7n(B3$BF|(B($B7n(B) 21:54 Shepherd, Lori <Lori.Shepherd using roswellpark.org<mailto:Lori.Shepherd using roswellpark.org>>:
It looks like you only pushed changes to the master branch which is our devel branch (3.11).

Please see the instructions for
http://bioconductor.org/developers/how-to/git/bug-fix-in-release-and-devel/

where RELEASE_X_Y   would be RELEASE_3_10  for pushing changes to the release 3.10 branch.

Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263

________________________________
From: Bioc-devel <bioc-devel-bounces using r-project.org<mailto:bioc-devel-bounces using r-project.org>> on behalf of L Rutter <lindsayannerutter using gmail.com<mailto:lindsayannerutter using gmail.com>>
Sent: Saturday, February 1, 2020 4:18 AM
To: bioc-devel using r-project.org<mailto:bioc-devel using r-project.org> <bioc-devel using r-project.org<mailto:bioc-devel using r-project.org>>
Subject: [Bioc-devel] R CMD build error for 3.10 (but not 3.11)

Dear all:

I have been receiving messages that my Bionconductor package bigPint is
producing errors on 'R CMD build' on malbec1. I believe I discovered the
source of the error, fixed it, and pushed changes to my GitHub account (
github.com/lindsayrutter<http://github.com/lindsayrutter>) on December 10, 2019. I continued to receive this
error notification and was advised to ensure I had pushed the changes to
Bioconductor. I pushed changes following these instructions (
http://bioconductor.org/developers/how-to/git/) on January 4, 2020. I am
still receiving notifications that my package produces errors on 'R CMD
build' on malbec1 (
https://master.bioconductor.org/checkResults/3.10/bioc-LATEST/bigPint/malbec1-buildsrc.html).
However, it seems my package is no longer failing daily build reports for
version 3.11 as of January 4, 2020 (
http://bioconductor.org/checkResults/3.11/bioc-LATEST/).

I am seeking advice on what a maintainer should try in this situation? Is
it possible my changes were only recognized in version 3.11 but not version
3.10? Thank you for any advice you may have!

Sincerely,
L. Rutter

        [[alternative HTML version deleted]]

_______________________________________________
Bioc-devel using r-project.org<mailto:Bioc-devel using r-project.org> mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

This email message may contain legally privileged and/or confidential information. If you are not the intended recipient(s), or the employee or agent responsible for the delivery of this message to the intended recipient(s), you are hereby notified that any disclosure, copying, distribution, or use of this email message is prohibited. If you have received this message in error, please notify the sender immediately by e-mail and delete this email message from your computer. Thank you.

This email message may contain legally privileged and/or confidential information.  If you are not the intended recipient(s), or the employee or agent responsible for the delivery of this message to the intended recipient(s), you are hereby notified that any disclosure, copying, distribution, or use of this email message is prohibited.  If you have received this message in error, please notify the sender immediately by e-mail and delete this email message from your computer. Thank you.
	[[alternative HTML version deleted]]