Segmentation fault when using data.table::tables() and colorout
eddelbuettel opened this issue · 11 comments
Huge fan of colorout here, but I think I just found a wart.
In non-interactive mode, my ~/.Rprofile does not load it and all is well:
edd@rob:~$ Rscript -e 'library(data.table); D <- data.table(a=1:3, b=LETTERS[1:3]); tables(); cat("Alive and well\n")'
NAME NROW NCOL MB COLS KEY
1: D 3 2 0 a,b
Total: 0MB
Alive and well
edd@rob:~$ In R, however, we die with a segfault once we get to tables():
> library(data.table); D <- data.table(a=1:3, b=LETTERS[1:3]); tables(); cat("Alive and well\n")
data.table 1.14.0 using 6 threads (see ?getDTthreads). Latest news: r-datatable.com
edd@rob:~$ If I set the RSTUDIO variable first to prevent colorout from being load, all is well. Can you take a look?
I can't replicate the issue. I've got "Alive and well" in both the Rscript command line and running R interactively. I'm running R 4.0.4 on Ubuntu 20.10.
Note: I've built R from source and all packages are built from source.
Same -- R 4.0.4 (via the binary at CRAN built from my Debian package sources) under 20.10. Hm.
My ~/.Rprofile.d/colorout.R is "yours" too, but possibly only I these colors. Not that it should matter.
## custom Nord theme thanks to https://github.com/jalvesaq/colorout/issues/27#issuecomment-599207290
if (requireNamespace("colorout", quietly=TRUE) && Sys.getenv("RSTUDIO") != "1") {
## General ----------------------------------------
#message("Hello, coloured world!")
colorout::setOutputColors(
index = '\x1b[38;2;76;86;106m',
normal = '\x1b[38;2;216;222;233m',
number = '\x1b[38;2;236;239;244m',
negnum = '\x1b[38;2;180;142;173m',
zero = '\x1b[38;2;136;192;208m', zero.limit = 0.01,
infinite = '\x1b[38;2;236;239;244m',
string = '\x1b[38;2;235;203;139m',
date = '\x1b[38;2;236;239;244m',
const = '\x1b[38;2;136;192;208m',
true = '\x1b[38;2;163;190;140m',
false = '\x1b[38;2;191;97;106m',
warn = '\x1b[38;2;235;203;139m',
stderror = '\x1b[38;2;191;97;106m', error = '\x1b[38;2;191;97;106m',
verbose = FALSE
)
## Custom patterns --------------------------------
## NOTE Do not copy all. Pick what you use/like.
## _ {data.table} ---------------------------------
colorout::addPattern('[0-9]*:', '\x1b[38;2;143;188;187m') # Row num
colorout::addPattern('---', '\x1b[38;2;76;86;106m') # Row splitter
colorout::addPattern('<[a-z]*>', '\x1b[38;2;143;188;187m') # Col class
## _ `str` ----------------------------------------
## Class
colorout::addPattern(' num ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' int ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' chr ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' Factor ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' Ord.factor ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' logi ', '\x1b[38;2;143;188;187m')
colorout::addPattern('function ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' dbl ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' lgcl ', '\x1b[38;2;143;188;187m')
colorout::addPattern(' cplx ', '\x1b[38;2;143;188;187m')
## Misc
colorout::addPattern('$ ', '\x1b[38;2;76;86;106m')
## _ `str`, {mlr3} --------------------------------
## R6 field name
colorout::addPattern('* [A-z]*:', '\x1b[38;2;235;203;139m')
colorout::addPattern("* [A-z]* [A-z]*:", '\x1b[38;2;235;203;139m')
colorout::addPattern("* [A-z]* [A-z]* [A-z]*:", '\x1b[38;2;235;203;139m')
colorout::addPattern("* [A-z]* [A-z]* [A-z]* [A-z]*:", '\x1b[38;2;235;203;139m')
## So on...
}Is there some tracing or debugging I could or should turn on? I have been using the same setup for a looong time, even checked the repo a few weeks back but it all seems stable.
Could be a fluke, could be me ... but just surprised me in some casual data.table work.
You could try strace. Colorout itself doesn't have any test. I added your code to my .Rprofile, but I still get "Alive and well".
strace is hard around all of R. But my dotfiles tickle on interactive() which is off in Rscript or R -q -e but I can override it for littler and its r. It ends in a lot of munmap() . Here are the last two screenfuls:
openat(AT_FDCWD, "/usr/lib/R/library/tools/R/sysdata.rdx", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=212, ...}) = 0
read(3, "\37\213\10\0\0\0\0\0\0\3\213\340b```f`\1BfV \223\201\317\321/\3303>\302"..., 4096) = 212
close(3) = 0
stat("/usr/lib/R/library/tools/R/sysdata.rdx", {st_mode=S_IFREG|0644, st_size=212, ...}) = 0
openat(AT_FDCWD, "/usr/lib/R/library/tools/R/sysdata.rdx", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=212, ...}) = 0
read(3, "\37\213\10\0\0\0\0\0\0\3\213\340b```f`\1BfV \223\201\317\321/\3303>\302"..., 16384) = 212
read(3, "", 12288) = 0
lseek(3, 0, SEEK_CUR) = 212
read(3, "", 16384) = 0
close(3) = 0
stat("/usr/lib/R/library/tools/data/Rdata.rdb", 0x7fff9204cba0) = -1 ENOENT (No such file or directory)
brk(0x55bc2ca93000) = 0x55bc2ca93000
openat(AT_FDCWD, "/usr/lib/R/library/tools/R/tools.rdb", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=1913572, ...}) = 0
fstat(3, {st_mode=S_IFREG|0644, st_size=1913572, ...}) = 0
lseek(3, 1912832, SEEK_SET) = 1912832
read(3, "\227\203\256\262\"\254%c\230\313R\217\210n\255\4\3253\347\346[\225r\347UK~~\253\230<\205"..., 740) = 740
brk(0x55bc2cc67000) = 0x55bc2cc67000
lseek(3, 0, SEEK_SET) = 0
read(3, "\0\0\2\203x\234\235R\261n\3020\20\275\304qQ\20T\f\355\2\02300\21\25\21U\355\310\310"..., 1912832) = 1912832
read(3, "\227\203\256\262\"\254%c\230\313R\217\210n\255\4\3253\347\346[\225r\347UK~~\253\230<\205"..., 4096) = 740
close(3) = 0
brk(0x55bc2cc89000) = 0x55bc2cc89000
brk(0x55bc2ccab000) = 0x55bc2ccab000
brk(0x55bc2cccc000) = 0x55bc2cccc000
brk(0x55bc2cced000) = 0x55bc2cced000
brk(0x55bc2cd0e000) = 0x55bc2cd0e000
brk(0x55bc2cd2f000) = 0x55bc2cd2f000
brk(0x55bc2cd50000) = 0x55bc2cd50000
brk(0x55bc2cd71000) = 0x55bc2cd71000
brk(0x55bc2cd92000) = 0x55bc2cd92000
brk(0x55bc2cdb3000) = 0x55bc2cdb3000
stat("/usr/lib/R/library/tools/libs/tools.so", {st_mode=S_IFREG|0644, st_size=97344, ...}) = 0
lstat("/usr", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
lstat("/usr/lib", {st_mode=S_IFDIR|0755, st_size=12288, ...}) = 0
lstat("/usr/lib/R", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
lstat("/usr/lib/R/library", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
lstat("/usr/lib/R/library/tools", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
lstat("/usr/lib/R/library/tools/libs", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
openat(AT_FDCWD, "/usr/lib/R/library/tools/libs/tools.so", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\200N\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0644, st_size=97344, ...}) = 0
mmap(NULL, 100776, PROT_READ, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f3f65c5a000
mprotect(0x7f3f65c5e000, 77824, PROT_NONE) = 0
mmap(0x7f3f65c5e000, 57344, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x4000) = 0x7f3f65c5e000
mmap(0x7f3f65c6c000, 16384, PROT_READ, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x12000) = 0x7f3f65c6c000
mmap(0x7f3f65c71000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x16000) = 0x7f3f65c71000
close(3) = 0
mprotect(0x7f3f65c71000, 4096, PROT_READ) = 0
brk(0x55bc2cdd4000) = 0x55bc2cdd4000
brk(0x55bc2cdf5000) = 0x55bc2cdf5000
brk(0x55bc2ce17000) = 0x55bc2ce17000
getpid() = 2152238
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
getpid() = 2152238
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
futex(0x7f3fc8318460, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318410, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc83184e0, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318490, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318560, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc83185e0, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318590, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318660, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc83186e0, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318760, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc83187e0, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318790, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318860, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318810, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc83188e0, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318890, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318960, FUTEX_WAKE_PRIVATE, 1) = 1
futex(0x7f3fc8318910, FUTEX_WAKE_PRIVATE, 1) = 1
munmap(0x7f3fc3784000, 8392704) = 0
munmap(0x7f3fbaf83000, 8392704) = 0
munmap(0x7f3fb2782000, 8392704) = 0
munmap(0x7f3fa9f81000, 8392704) = 0
munmap(0x7f3fa1780000, 8392704) = 0
munmap(0x7f3f98f7f000, 8392704) = 0
munmap(0x7f3f9077e000, 8392704) = 0
munmap(0x7f3fbb784000, 134217728) = 0
munmap(0x7f3fb2f83000, 134217728) = 0
munmap(0x7f3faa782000, 134217728) = 0
munmap(0x7f3fa1f81000, 134217728) = 0
munmap(0x7f3f99780000, 134217728) = 0
munmap(0x7f3f90f7f000, 134217728) = 0
munmap(0x7f3f8877e000, 134217728) = 0
munmap(0x7f3f7ff7d000, 134217728) = 0
munmap(0x7f3f7777c000, 134217728) = 0
munmap(0x7f3f6ddb5000, 134217728) = 0
munmap(0x7f3f65cb5000, 134217728) = 0
exit_group(0) = ?
+++ exited with 0 +++
Let me try on another (similar) machine just in case.
Yes. let's close. The minimal example didn't replicate on a smaller laptop (same R, Ubuntu, ... ) after I setup a similar minimal ~/.R/profile.d and its colourout.R. Rebuilding data.table doesn't do anything for me. Odd. But likely something else...
Thanks for the feedback. Please, reopen the issue if necessary.
Thanks for lending an ear, and the super-speedy replies. I will try bring the "small machine" setup closer to my main one, maybe it replicates there. Else files under Heisenbug...
Thanks as always for colorout.
More completeness, I found the offending line, and it has (of course) nothing to do with colorour. Sorry for making noise earlier.
Thanks for the feedback! I'm glad to know that colorout is still healthy.