Skip to content

R-devel error

5 messages · Pearce, Robert, Martin Maechler, Göran Broström +2 more

Original
Pearce, Robert
# 
I am resubmitting this bug report but with additional information. I am running this with windows 10:  w64-mingw32 with R Under development (unstable) (2017-12-04 r73829). I build 'httk' from the command prompt using 'R CMD build httk' after installing the required packages. Then when the vignettes are being created, it crashes. Today I installed the latest versions of Rtools, MikTex and R-devel and still got the same error. We have also replicated this on another machine. I am able to successfully build the package in this way with R 3.4.3.

Robert

From: Pearce, Robert
Sent: Monday, November 20, 2017 11:32 AM
To: 'r-devel at r-project.org' <r-devel at r-project.org>
Subject: R-devel error

Hello,

During the last week I've been getting an error with r-devel when building the source package for 'httk'. Everything works fine until: 'creating vignettes...' . Then I get a pop-up stating: 'R for Windows front-end has stopped working... close the program'. I have installed the latest version of R, and this error does not occur.

Robert Pearce
Martin Maechler
# 
> I am resubmitting this bug report but with additional information. I am running this with windows 10:  w64-mingw32 with R Under development (unstable) (2017-12-04 r73829). 

There is no such message (as you cite below) in the R-devel archives:
      https://stat.ethz.ch/pipermail/r-devel/2017-November/thread.html

Hence it was never posted for some reason
(You may have *sent* it, but it never arrived "safely")


    > I build 'httk' from the command prompt using 'R CMD build httk' after installing the required packages. Then when the vignettes are being created, it crashes. Today I installed the latest versions of Rtools, MikTex and R-devel and still got the same error. We have also replicated this on another machine. I am able to successfully build the package in this way with R 3.4.3.
To make this reproducible, indeed, I also get 
problem, possibly the same as you on my Linux machine (Fedora 26):

1. Get  httk_1.7.tar.gz from https://cran.r-project.org/package=httk

2. unpack (for me:)
	tar xf httk_1.7.tar.gz
3.
	R-devel CMD build httk   

I get the following.
Notably after having waited several minutes at the prompt
(line  of output):

 '* creating vignettes ' 

  ------------------------------------------------------------------------
  [nfslocal at lynne tmp]$ R-devel CMD build httk
  * checking for file ?httk/DESCRIPTION? ... OK
  * preparing ?httk?:
  * checking DESCRIPTION meta-information ... OK
  * cleaning src
  * installing the package to process help pages
  * saving partial Rd database
  * creating vignettes ...sh: line 1: 16259 Segmentation fault      (core dumped) '/u/maechler/R/D/r-devel/inst/bin/Rscript' --vanilla --default-packages= -e "tools::buildVignettes(dir = '.', tangle = TRUE)" > '/tmp/RtmpV1grVw/xshell3e793c7d5258' 2>&1
   ERROR

   *** caught segfault ***
  address 0x1, cause 'memory not mapped'

  Traceback:
   1: .Call(C_stri_locate_all_regex, str, pattern, omit_no_match, opts_regex)
   2: stri_locate_all_regex(string, pattern, omit_no_match = TRUE,     opts_regex = opts)
   3: stringr::str_locate_all(input, inline.code)
   4: parse_inline(g, patterns)
   5: FUN(X[[i]], ...)
   6: lapply(groups, function(g) {    block = grepl(chunk.begin, g[1])    if (!set.preamble && !parent_mode()) {        return(if (block) "" else g)    }    if (block) {        n = length(g)        if (n >= 2 && grepl(chunk.end, g[n]))             g = g[-n]        g = strip_block(g, patterns$chunk.code)        params.src = if (group_pattern(chunk.begin)) {            stringr::str_trim(gsub(chunk.begin, "\\1", g[1]))        }        else ""        parse_block(g[-1], g[1], params.src)    }    else parse_inline(g, patterns)})
   7: split_file(lines = text)
   8: process_file(text, output)
   9: knitr::knit(knit_input, knit_output, envir = envir, quiet = quiet,     encoding = encoding)
  10: rmarkdown::render(file, encoding = encoding, quiet = quiet, envir = globalenv())
  11: vweave_rmarkdown(...)
  12: engine$weave(file, quiet = quiet, encoding = enc)
  13: doTryCatch(return(expr), name, parentenv, handler)
  14: tryCatchOne(expr, names, parentenv, handlers[[1L]])
  15: tryCatchList(expr, classes, parentenv, handlers)
  16: tryCatch({    engine$weave(file, quiet = quiet, encoding = enc)    setwd(startdir)    find_vignette_product(name, by = "weave", engine = engine)}, error = function(e) {    stop(gettextf("processing vignette '%s' failed with diagnostics:\n%s",         file, conditionMessage(e)), domain = NA, call. = FALSE)})
  17: tools::buildVignettes(dir = ".", tangle = TRUE)
  An irrecoverable exception occurred. R is aborting now ...
  [nfslocal at lynne tmp]$ 
  ------------------------------------------------------------------------


I've seen this (segfault, triggered from stringr stuff) before,
occasionally and always thought it was a consequence of a messup
(different R versions and package installations in libraries I
use on the same computer....),  but now I have convinced myself
that this is not the case: I can easily reproduce the above using

       tools::buildVignettes(dir = "<mypath_to>/httk")

==> So I'd suggest there's a pretty low-level bug somewhere.

- Either in 'stringr' or 'stringi' in their C code using
  things they should not (because not in R's API),
  which now shows in R-devel where many "inner workings" have been
  modified (key "ALTREP")

- or in R-devel .. in these new "inner workings".


I don't have time to dig further at the moment.

Thank you, Robert, for your report!

Martin Maechler,
ETH Zurich


    > From: Pearce, Robert
    > Sent: Monday, November 20, 2017 11:32 AM
    > To: 'r-devel at r-project.org' <r-devel at r-project.org>
    > Subject: R-devel error


    > Hello,

    > During the last week I've been getting an error with r-devel when building the source package for 'httk'. Everything works fine until: 'creating vignettes...' . Then I get a pop-up stating: 'R for Windows front-end has stopped working... close the program'. I have installed the latest version of R, and this error does not occur.

    > Robert Pearce
Göran Broström
# 
Interesting, yesterday I had exactly this problem, but today I solved it 
(see the R-pkg-devel list) by _installing_ R-devel and

 > update.packages(checkBuilt = TRUE)

I also tried it on Martin's Fedora 26 example below. Worked fine (had to 
install some packages...)

On ubuntu 16.04.

G?ran Brostr?m
On 2017-12-06 11:29, Martin Maechler wrote:
Gabriel Becker
# 
Robert,

To expand a bit on Goran's point, is it possible that you have an old
library with packages installed that your R-devel is hitting. The R headers
changes relatively recently, so any packages with compiled code that were
built long enough ago (I don't have the exact date off the top of my head)
need to be rebuilt before they can be safely used.

If this is happening in a state where both the package in question AND all
it's dependencies are freshly built under the exact version of R-devel
you're trying to check httk with, then the above isn't the problem and
something else must be going on.

Best,
~G

On Wed, Dec 6, 2017 at 8:14 AM, G?ran Brostr?m <goran.brostrom at umu.se>
wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Luke Tierney
    

    

      
      #
    

  


  

      
library() in R-devel checks that packages are built for the current
ABI and signals an error if they are not. A segfault is likely a
pre-existing bug unmasked by the changed memory layout. Several of
these have been resolved by recent package updates. I also cannot
reproduce the problem on current R-devel with fairly current packages
from CRAN.

Best,

luke

        
On Wed, 6 Dec 2017, Gabriel Becker wrote: