Re: [Rd] Crash/bug when calling match on latin1 strings

[Martin Maechler]

> Rui Barradas 
> on Mon, 11 Oct 2021 07:41:51 +0100 writes:

> Hello,

> R 4.1.1 on Ubuntu 20.04.

> I can reproduce this error but not ~90% of the time, only the 1st time I 
> run the script.
> If I run other (terminal) commands before rerunning the R script it 
> sometimes segfaults again but once again very far from 90% of the time.


> rui@rui:~/tmp$ R -q -f rhelp.R
>> sessionInfo()
> R version 4.1.1 (2021-08-10)
> Platform: x86_64-pc-linux-gnu (64-bit)
> Running under: Ubuntu 20.04.3 LTS

> Matrix products: default
> BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0
> LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.9.0

> locale:
> [1] LC_CTYPE=pt_PT.UTF-8   LC_NUMERIC=C
> [3] LC_TIME=pt_PT.UTF-8LC_COLLATE=pt_PT.UTF-8
> [5] LC_MONETARY=pt_PT.UTF-8LC_MESSAGES=pt_PT.UTF-8
> [7] LC_PAPER=pt_PT.UTF-8   LC_NAME=C
> [9] LC_ADDRESS=C   LC_TELEPHONE=C
> [11] LC_MEASUREMENT=pt_PT.UTF-8 LC_IDENTIFICATION=C

> attached base packages:
> [1] stats graphics  grDevices utils datasets  methods   base

> loaded via a namespace (and not attached):
> [1] compiler_4.1.1
>> 
>> # A bunch of words in UTF8; replace *'s
>> words <- readLines("h://pastebin.c**/raw/MFCQfhpY", encoding = 
> "UTF-8")
>> words2 <- iconv(words, "utf-8", "latin1")
>> gctorture(TRUE)
>> y <- match(words2, words2)

> *** caught segfault ***
> address 0x10, cause 'memory not mapped'
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation
> *** recursive gc invocation

> Traceback:
> 1: match(words2, words2)
> An irrecoverable exception occurred. R is aborting now ...
> Falta de segmentação (núcleo despejado)



> This last line is Portuguese for

> Segmentation fault (core dumped)

> Hope this helps,

Yes, it does, thank you!

I can confirm the problem:  Only in R 4.1.0 and newer, and
including current "R-patched" and "R-devel" versions.

I've now turned this into a formal R bug report on R's bugzilla,
and (slightly) extended your (Travers') example into self
contained (no internet access) R script.

Bugzilla PR#18211 :" match() memory corruption "

  https://bugs.r-project.org/show_bug.cgi?id=18211

  with attachment 2929
  --> https://bugs.r-project.org/attachment.cgi?id=2929&action=edit

==> please if possible follow up on bugzilla

Thanks again to you both!
Martin Maechler


> Rui Barradas

> Às 06:05 de 11/10/21, Travers Ching escreveu:
>> Here's a brief example:
>> 
>> # A bunch of words in UTF8; replace *'s
>> words <- readLines("h://pastebin.c**/raw/MFCQfhpY", encoding = 
"UTF-8")
>> words2 <- iconv(words, "utf-8", "latin1")
>> gctorture(TRUE)
>> y <- match(words2, words2)
>> 
>> 
>> I searched bugzilla but didn't see anything. Apologies if this is already
>> reported.
>> 
>> The bug appears in both R-devel and the release, but doesn't seem to 
affect
>> R 4.0.5.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Rd] int overflow writing long vectors to socketConnection

[Zafer Barutcuoglu]

Hi,

Writing >=2GB to a socketConnection (e.g. via writeBin) does not work 
correctly, because of this int typecast in modules/internet/sockconn.c:
> static size_t sock_write(const void *ptr, size_t size, size_t nitems,
>Rconnection con)
> {
> Rsockconn this = (Rsockconn)con->private;
> ssize_t n = R_SockWrite(this->fd, ptr, (int)(size * nitems),
>   this->timeout)/((ssize_t)size);
> return n > 0 ? n : 0;
> }
which seems uncalled for, given:
> ssize_t R_SockWrite(int sockp, const void *buf, size_t len, int timeout)


Is there a rationale for it, or should it be fixed?

Best,
--
Zafer


[[alternative HTML version deleted]]

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel