Skip to content
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

revdep results #3233

Closed
mattdowle opened this issue Dec 18, 2018 · 9 comments
Closed

revdep results #3233

mattdowle opened this issue Dec 18, 2018 · 9 comments
Milestone

Comments

@mattdowle
Copy link
Member

@mattdowle mattdowle commented Dec 18, 2018

Not bad: just 21 out of 586 CRAN revdeps to investigate. Most of the Bioconductor ones are usual suspects: they don't tend to fix warnings on Bioconductor and are usually unrelated to data.table. Still need to be investigated though.

Installed data.table to be tested against: 1.11.9 
CRAN:
 ERROR   :  17 : antaresRead batchtools behavr damr DeclareDesign ggetho maditr popEpi PROSPER rENA rtsdata scopr sdcMicro sleepr timeseriesdb toxboot zeitgebr 
 WARNING :   4 : assertable ctsem SpaDES.core tidytext 
 NOTE    : 183 
 OK      : 382 
 TOTAL   : 586 / 586 
 
BIOC:
 ERROR   :  12 : CAGEr facopy GenVisR ImmuneSpaceR LowMACA LymphoSeq MAST phantasus RGMQL scater TIN TitanCNA 
 WARNING :  18 : bsseq CONFESS CRISPRseek crossmeta eQTL GenoGAM iCNV IrisSpatialFeatures methylPipe minfi MinimumDistance paxtoolsr RiboProfiling rTANDEM S4Vectors SISPA TFutils Uniquorn 
 NOTE    :  87 
 OK      :  18 
 TOTAL   : 135 / 135 
 
TOTAL          : 721 

fail.log

@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 2, 2019

To include on CRAN upload message.

Communication with downstream maintainers has happened in advance :

Already in error/warning on CRAN and unrelated to data.table :

  • DeclareDesign I see the same error (expect_equal(ncol(draw_estimates(des)), 7) did not produce any warnings) locally. Seems unrelated to data.table.
  • ctsem Passes now locally after installing biber. The warnings and errors on CRAN seem to be system specific and unrelated to data.table
  • worrms Fails locally with same error as CRAN (all(grepl(format(Sys.Date(), "%Y"), aa$modified)) isn't true). Looks unrelated to data.table; e.g. related to 2018 now being 2019.

@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 4, 2019

Current status after full rerun :

Installed data.table to be tested against: 1.11.9 
CRAN:
 ERROR   :   2 : DeclareDesign worrms 
 WARNING :   0 :  
 NOTE    : 191 
 OK      : 403 
 TOTAL   : 596 / 596 
 
BIOC:
 ERROR   :   6 : GenVisR ImmuneSpaceR LymphoSeq MAST QuartPAC scater 
 WARNING :  18 : bsseq CONFESS CRISPRseek crossmeta eQTL GenoGAM iCNV IrisSpatialFeatures methylPipe minfi MinimumDistance paxtoolsr RiboProfiling rTANDEM S4Vectors SISPA TitanCNA Uniquorn 
 NOTE    :  91 
 OK      :  19 
 TOTAL   : 134 / 134 
 
TOTAL          : 730

@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 5, 2019

Rerun of Bioc fails with CRAN data.table 1.11.8 :

BIOC:
 ERROR   :   5 : GenVisR ImmuneSpaceR LymphoSeq MAST scater 
 WARNING :  18 : bsseq CONFESS CRISPRseek crossmeta eQTL GenoGAM iCNV IrisSpatialFeatures methylPipe minfi MinimumDistance paxtoolsr RiboProfiling rTANDEM S4Vectors SISPA TitanCNA Uniquorn 
 NOTE    :  92 
 OK      :  19 
 TOTAL   : 134 / 134 

So, other than QuartPAC, they're all already in error/warning with CRAN version. This is long-standing on Bioconductor and not receiving attention.

Bioc status diff to investigate :

  • QuartPAC When I reran with v1.11.9, it passed fine with just notes. I hadn't kept a log of the previous fail. Sometimes a rerun works if internet access is needed for a package's tests so I'll put it down to that.

Email to 23 Bioconductor package maintainers sent on 4th Jan 2018 :

Dear maintainer of Bioconductor package,

Please could you fix the error or warning status. Your package uses data.table in some way. Since your package is already in error/warning status, it makes it much harder for me to test the impact each data.table update has. In some cases the error/warning status has been left unfixed on Bioconductor for years. In some cases the problem might be local to my machine -- please look at the log attached for your package and let me know.

ERROR : 5 : GenVisR ImmuneSpaceR LymphoSeq MAST scater
WARNING : 18 : bsseq CONFESS CRISPRseek crossmeta eQTL GenoGAM iCNV IrisSpatialFeatures methylPipe minfi MinimumDistance paxtoolsr RiboProfiling rTANDEM S4Vectors SISPA TitanCNA Uniquorn

Best, Matt

fail.log


@mattdowle mattdowle closed this Jan 5, 2019
@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 8, 2019

I needed to update repo of Bioconductor from 3.7 to 3.8. Thanks to several Bioconductor maintainers for taking a look and letting me know.

@mattdowle mattdowle reopened this Jan 8, 2019
@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 8, 2019

After updating to Bioc 3.8 and rerunning, email to 26 Bioc maintainers on 8th Jan 2018 :


Dear maintainer of Bioconductor package,

Please could you fix the error or warning status in the log attached, or let me know how I can fix it locally. Your package uses data.table in some way. Since your package is already in error/warning status with the released version of data.table (1.11.8), it makes it much harder for me to test the impact each data.table update has.

There are no errors or warnings for the 598 packages on CRAN that use data.table :

CRAN:
 ERROR   :   0 :  
 WARNING :   0 :  
 NOTE    : 192 
 OK      : 406 
 TOTAL   : 598 / 598 

There are 26 Bioc packages that use data.table that have error or warning. I don't think these are related to data.table but these make it much harder to spot errors or warnings that data.table has caused when it is updated.
I'm using Bioc 3.8 and session info is at the top of the attached log.

BIOC:
 ERROR   :   2 : GENESIS ImmuneSpaceR 
 WARNING :  24 : CEMiTool CONFESS crossmeta dada2 ELMER eQTL flowWorkspace GenoGAM GOTHiC
                 iCNV MAST methylPipe MinimumDistance netSmooth OUTRIDER RiboProfiling
                 rTANDEM S4Vectors SCnorm SISPA TENxBrainData TitanCNA Ularcirc Uniquorn 
 NOTE    : 102 
 OK      :  24 
 TOTAL   : 152 / 152 

Thanks, Matt


fail.log

@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 8, 2019

Testing 1.11.9 and looking for diff in status.

  • rerun ok for 7 due to resource/internet issues: BiocParallel, MAGeCKFlute, SWATH2stats, GSALightning, RegParallel, TCGAbiolinksGUI, RCAS

  • geneXtendeR -- true break caused by data.table in dev: fixed #3264

@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 10, 2019

checkmate 1.9.0 update today is broken by data.table 1.12.0
Caused by checking a bug isn't fixed but it is now: mllg/checkmate#155

@mattdowle mattdowle reopened this Jan 10, 2019
@mattdowle mattdowle closed this Jan 10, 2019
@mattdowle
Copy link
Member Author

@mattdowle mattdowle commented Jan 11, 2019

Final rerun for #3211 given #3273 (fread colClasses fixes).

  • damr Can't use NULL in colClasses when select or drop is used as well. due to #3273. Passes ok now with fix in data.table f0bd6e3
  • prediction WARNING Unknown package ‘mnlogit’ in Rd xrefs mnlogit removed from CRAN today "as check issues were not corrected despite reminder." so prediction will need to remove its xref to mnlogit accordingly.
  • segregation Error: between + within = total (@test_mutual_total.R#53). New error on CRAN today too (here). I see the same locally. So not due to data.table update. Just coincidence there was a recent change to between() and the messages refer to between.
  • tosca GitHub API rate limit exceeded
  • xlm Error handling for bad public key. Also on CRAN (new error today) so not due to data.table update.
  • bsseq Duplicates detected in drop PR #3277

@jangorecki
Copy link
Member

@jangorecki jangorecki commented Jan 12, 2019

GitHub API rate limit exceeded might be client side issue. I am recently getting this when using devtools install_github for dplyr in db-benchmark and I need to wait up to one hour. It seems that gh reduced limit of API calls when not using gh token auth.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Linked pull requests

Successfully merging a pull request may close this issue.

None yet
2 participants