Giter Club home page Giter Club logo

revdepcheck's Introduction

❗ This package is not actively maintained

Use it at your own risks. For alternative reverse dependency check tools, see

Features

  • To avoid false positives due to existing failures, revdepcheck runs R CMD check twice for each revdep, once with the CRAN version of your package, and once with the local development version. revdepcheck reports the difference, so you can see exactly what has changed.

  • To speed up installation of revdeps and their dependencies, revdepcheck relies on crancache. You can see what packages are currently cached with crancache::crancache_list().

  • revdepcheck is carefully designed to make long running checks as pleasant as possible. You run checks in parallel, check time is limited to 10 minutes, and an elegant progress bar keeps you up-to-date with what's happening (including an estimate of how much time is remaining).

Installation

pak::pkg_install("r-lib/revdepcheck")

Usage

library(revdepcheck)

Check package in working directory, creating "revdep/" directory if it doesn't already exist:

revdep_check(num_workers = 4)

If the run fails to complete, run again and it will pick up where it left off:

revdep_check(num_workers = 4)

During execution, run these in a separate R process to view status completed checks:

revdep_summary()                 # table of results by package 
revdep_details(".", "<package>") # full details for the specified package

Generate human-friendly summary documents in revdep/:

revdep_report()
## Writing *partial* report
## Writing summary to 'revdep/README.md'
## Writing problems to 'revdep/problems.md'
## Writing failures to 'revdep/failures.md'

Manage a "todo" list of packages to examine:

revdep_add(pkg = ".", <packages>)  # add <packages> to the list
revdep_rm(pkg = ".", <packages>).  # remove <packages> from list

revdep_add_broken()  # add all broken packages
revdep_add_new()     # add newly available packages
revdep_todo()        # list packages in the todo list

Clear out all previous results

revdep_reset()

We recommend running revdep_check() in a separate process (e.g. new terminal under RStudio). That way, while it runs in a background tab, you can easily use your revdep_details(revdep = "pkg") to see what's gone wrong with "pkg".

Status Flags:

  • install newly fails: i-
  • install still fails: i+
  • install/check newly timeouts: t-
  • install/check still timeouts: t+
  • No new failures, success: +
  • Some new failures: -

License

MIT © Gábor Csárdi, R Consortium, RStudio Inc

revdepcheck's People

Contributors

adamhsparks avatar ateucher avatar bertcarnell avatar davisvaughan avatar fmichonneau avatar gaborcsardi avatar hadley avatar henrikbengtsson avatar jennybc avatar jimhester avatar krlmlr avatar lbergelson avatar lionel- avatar llrs avatar maksymiuks avatar romainfrancois avatar stevenolen avatar vspinu avatar warnes avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar

revdepcheck's Issues

Install all deps of revdeps first?

I think there are two advantages:

  • crancache will have maximal information, so can easily use parallel downloads etc
  • you're more robust to bad internet (i.e. #15)

You'd still do the install phase for each package to create the individual libraries, and to possible catch new updates in long running rev dep processes.

Resume by default

Currently, if you run revdep_check() after a previous partial run, it prompts you to overrwrite or resume. I think it should resume by default, verifying each step to see whether or not it's necessary.

Add all results to database

revdep_report_summary() seems to require access to the revdep/checks directory, which I'd rather avoid archiving. Would it be feasible to rely only on the database for displaying summaries?

Progress bar

  • elapsed time,
  • ETA
  • number of packages checked
  • total number of packages
  • number of packages checking right now

Move debugme to Suggests?

Not particularly urgent, but I'm looking into adding debugme to other packages used by revdepcheck, and Suggests seems more appropriate to me.

Email template is duplicated when there are multiple results

The your_results field may have length > 1 if there were issues in both examples and tests. Then glue_data() will duplicate the email template.

I worked around this by sticking this in email_build():

data$your_results <- glue::collapse(data$your_results)

Warnings after revdepcheck

Probably from crancache or cranlike. Example from tidyr

1: call dbDisconnect() when finished working with a connection
2: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES.new' to '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES', reason 'No such file or directory'
3: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES.gz.new' to '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES.gz', reason 'No such file or directory'
4: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES.rds.new' to '/home/muelleki/.cache/R-crancache/cran-bin/src/contrib/PACKAGES.rds', reason 'No such file or directory'
5: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES', reason 'No such file or directory'
6: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.gz.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.gz', reason 'No such file or directory'
7: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.rds.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.rds', reason 'No such file or directory'
8: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES', reason 'No such file or directory'
9: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.gz.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.gz', reason 'No such file or directory'
10: In file.rename(file.path(dir, pkgs_new), file.path(dir,  ... :
  cannot rename file '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES.new' to '/home/muelleki/.cache/R-crancache/cran/src/contrib/PACKAGES', reason 'No such file or directory'

... and ~40 more in this vein.

Graphics output

Some packages fail, because they try to use tcltk, or plot to the default graphics device.
We need a workaround for these, e.g. an alternate graphics device, or a headless X server might work.

I don't know what happens to these on Windows.

Cleanup more

E.g.

  • .tar.gz source packages
  • vign_test directories
  • all other files, except for 00check.log and 00install.txt, maybe?

Rewrite to check in parallel

Features we want to have

  • docker & localbuilder support (localbuilder needs to mature a lot before this....)
  • persistent state in liteq
  • multiple workers via processx / sys
  • Nice summarization, run with old and new package, compare.
  • sysreqs support
  • time limit for a check

@krlmlr ideas welcome!

Works for RSQLite?

I'm having trouble revdep-checking the current dev version of RSQLite with devtools. I'd like to check both CRAN and Bioconductor packages.

Do you think it's worthwhile to try revdepcheck on this project already?

Someway to access partial results during checking

i.e. I see:

✔ ARTool 0.10.4                          ── E: 1     | W: 0     | N: 0                        
✔ alfred 0.1.1                           ── E: 0     | W: 0     | N: 0    
✔ atlantistools 0.4.2                    ── E: 1     | W: 0     | N: 0
✔ bkmr 0.2.0                             ── E: 1     | W: 0     | N: 0                        

So I suspect something is really wrong, and I'd look to start looking at the results while the checks the continue.

Markdown reports

To replace existing devtools reports:

  • README.md should give platform and package versions of dependencies (for both new and old). Then print revdep summaries

  • broken.md should contain the full differences for all packages that broke

Mid run install?

I saw this:

✔ neuropsychology 0.5.0                  ── E: 0     | W: 0     | N: 1                               
✔ NFP 0.99.2                             ── E: 1     | W: 0     | N: 0                               
[146/219] 03:50:24 | ETA:  2h | (4) nonmemica [I], nzelect [D], openEBGM [I], outreg [I]trying URL 'https://cran.rstudio.com/src/contrib/nzelect_0.3.3.tar.gz'
Content type 'application/x-gzip' length 448342 bytes (437 KB)
==================================================
downloaded 437 KB

✔                                                                                                    
✔ openEBGM 0.1.0                         ── E: 0     | W: 0     | N: 0                               

NIce report comparing old and new versions

✖ package version             ── a-b+c E ✔ | a-b+c W ✔ | a-b+c N ✔
✖ package version             ── a-b+c E ✔ | a-b+c W ✔ | a-b+c N ✔
✖ package version             ── a-b+c E ✔ | a-b+c W ✔ | a-b+c N ✔
  • a: number of events present with both versions
  • b: number of events present only with the old version
  • c: number of events present only with the new version

E.g.

✖ remotes 1.0.0               ── 0     E ✔ | 1  +2 W ✔ | 1-1+1 N ✔
✖ spareserver 1.1.2           ── 1-1   E ✔ | 0  +1 W ✔ | 2-1+1 N ✔

Test on Linux

In particular, work out how the locally built binary packages should be cached. This is a crancache issue as well.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.