Recommended strategy for dealing with broken symlinks?

[jdblischak]

Is there a recommended strategy for finding and gracefully handling broken symlinks? My package functions run path_real() on the project files to resolve any potential symlinks, and I had a user report a bug where a broken symlink broke everything. Is there a better option than always having to check with is_link()/link_path()/file_exists()?

Here is a reprex to demonstrate the issue:

packageVersion("fs")
#> [1] '1.3.1.9000'
library(fs)
d <- file_temp()
dir_create(d)
non_existent <- path(d, "non-existent")
broken_link <- path(d, "broken-link")
link_create(non_existent, broken_link)
file_exists(non_existent)
#> /tmp/RtmpeYDu12/file25f817354d04/non-existent 
#>                                         FALSE
link_exists(broken_link)
#> /tmp/RtmpeYDu12/file25f817354d04/broken-link 
#>                                         TRUE
link_path(broken_link)
#> /tmp/RtmpeYDu12/file25f817354d04/non-existent
file_exists(link_path(broken_link))
#> /tmp/RtmpeYDu12/file25f817354d04/non-existent 
#>                                         FALSE
f <- dir_ls(path = d)
path_real(f)
#> Error: [ENOENT] Failed to realize '/tmp/RtmpeYDu12/file25f817354d04/broken-link': no such file or directory
# Find any broken links
any(!file_exists(link_path(f[is_link(f)])))
#> [1] TRUE

^{Created on 2019-10-29 by the reprex package (v0.3.0.9000)}

Session info

sessioninfo::session_info()
#> ─ Session info ──────────────────────────────────────────────────────────
#>  setting  value                       
#>  version  R version 3.6.1 (2019-07-05)
#>  os       Ubuntu 18.04.3 LTS          
#>  system   x86_64, linux-gnu           
#>  ui       X11                         
#>  language (EN)                        
#>  collate  en_US.UTF-8                 
#>  ctype    en_US.UTF-8                 
#>  tz       America/New_York            
#>  date     2019-10-29                  
#> 
#> ─ Packages ──────────────────────────────────────────────────────────────
#>  package     * version    date       lib source                   
#>  assertthat    0.2.1      2019-03-21 [3] CRAN (R 3.5.3)           
#>  cli           1.1.0      2019-03-19 [3] CRAN (R 3.5.3)           
#>  crayon        1.3.4      2017-09-16 [3] CRAN (R 3.5.0)           
#>  digest        0.6.22     2019-10-21 [3] CRAN (R 3.6.1)           
#>  evaluate      0.14       2019-05-28 [3] CRAN (R 3.6.0)           
#>  fs          * 1.3.1.9000 2019-10-28 [1] Github (r-lib/fs@a292266)
#>  highr         0.8        2019-03-20 [3] CRAN (R 3.5.3)           
#>  htmltools     0.4.0      2019-10-04 [3] CRAN (R 3.6.1)           
#>  knitr         1.25       2019-09-18 [3] CRAN (R 3.6.1)           
#>  magrittr      1.5        2014-11-22 [3] CRAN (R 3.5.0)           
#>  Rcpp          1.0.2      2019-07-25 [3] CRAN (R 3.6.1)           
#>  reprex        0.3.0.9000 2019-07-19 [1] local                    
#>  rlang         0.4.1      2019-10-24 [3] CRAN (R 3.6.1)           
#>  rmarkdown     1.15       2019-08-21 [1] CRAN (R 3.6.1)           
#>  sessioninfo   1.1.1      2018-11-05 [3] CRAN (R 3.5.1)           
#>  stringi       1.4.3      2019-03-12 [3] CRAN (R 3.5.3)           
#>  stringr       1.4.0      2019-02-10 [3] CRAN (R 3.5.2)           
#>  withr         2.1.2      2018-03-15 [3] CRAN (R 3.5.0)           
#>  xfun          0.10       2019-10-01 [3] CRAN (R 3.6.1)           
#>  yaml          2.2.0      2018-07-25 [3] CRAN (R 3.5.1)           
#> 
#> [1] /home/jdb-work/R/x86_64-pc-linux-gnu-library/3.6
#> [2] /usr/local/lib/R/site-library
#> [3] /usr/lib/R/site-library
#> [4] /usr/lib/R/library

[jimhester]

Maybe link_exists() should do the equivalent of file_exists(link_path()), I am sure...

[jdblischak]

Maybe link_exists() should do the equivalent of file_exists(link_path()), I am sure...

They are slightly different, ie is there a link vs does the link point to an existing file.

I was thinking of a change to path_real(). In the above example, it could return the path /tmp/RtmpeYDu12/file25f817354d04/non-existent instead of throwing an error. Maybe this behavior could be controlled by an argument?

[jimhester]

I think path_real() should fail if the file doesn't exist, if the file doesn't exist there is no real path.

[jdblischak]

But what about Issue #144, where you added the ability to resolve the symlink even if the destination file doesn't exist?

Basically, I need a function that will resolve any potential symlinks in a file path, and will not complain if anything is missing. If possible I'd prefer an fs solution since that will be much faster than any recursive R function that I write to handle resolving symlinks.

[jimhester]

Ok I forgot about that, your proposal makes sense then

[jdblischak]

I recently got bit by this again. Would it be possible to add a flag to path_real() that doesn't require the target file to exist? Something like the following behavior:

> fs::path_real("broken-symlink", error = TRUE)
Error: [ENOENT] Failed to realize 'broken-symlink': no such file or directory
> fs::path_real("broken-symlink", error = FALSE)
[1] "broken-symlink"

[jimhester]

I think making path_real() do anything but go to the real file was a mistake. If you try to use path_real() on a file that does not exist it should fail.

[jdblischak]

xref: workflowr/workflowr#204