Re: [Rd] Intermittent crashes with inset `[<-` command
Fixed in R-devel and R-patched.
Best,
luke
On Wed, 27 Feb 2019, Tierney, Luke wrote:
> Thanks for the report. Should be fixed shortly.
>
> Best,
>
> luke
>
> On Tue, 26 Feb 2019, Brian Montgomery via R-devel wrote:
>
>> The following code crashes after about 300 iterations on my
>> x86_64-w64-mingw32 machine on R 3.5.2 --vanilla.
>> Others have duplicated this (see
>> https://github.com/tidyverse/magrittr/issues/190 if necessary), but I don't
>> know how machine/OS-dependent it may be.
>> If it doesn't crash for you, please try increasing the length of the x
>> vector.
>>
>> Substituting the commented-out line for the one below it works correctly
>> (prints out 1:1000 and ends normally) every time.
>>
>> x <- 1:20
>> y <- rep(letters[1:5], length(x) / 5L)
>> for (i in 1:1000) {
>> # x[y == 'a'] <- x[y == 'b']
>> x <- `[<-`(x, y == 'a', x[y == 'b'])
>> cat(i, '')
>> }
>> cat('\n')
>>
>> The point of using this syntax is to make it work better with pipes, but the
>> errors occur without pipes or magrittr.
>>
>> Thank you for your help!
>>
>> __
>> R-devel@r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>
>
--
Luke Tierney
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa Phone: 319-335-3386
Department of Statistics andFax: 319-335-3017
Actuarial Science
241 Schaeffer Hall email: luke-tier...@uiowa.edu
Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Intermittent crashes with inset `[<-` command
On Tue, 26 Feb 2019, Brian Montgomery via R-devel wrote:
> The following code crashes after about 300 iterations on my
> x86_64-w64-mingw32 machine on R 3.5.2 --vanilla.
> Others have duplicated this (see
> https://github.com/tidyverse/magrittr/issues/190 if necessary), but I don't
> know how machine/OS-dependent it may be.
> If it doesn't crash for you, please try increasing the length of the x vector.
>
> Substituting the commented-out line for the one below it works correctly
> (prints out 1:1000 and ends normally) every time.
>
> x <- 1:20
> y <- rep(letters[1:5], length(x) / 5L)
> for (i in 1:1000) {
> # x[y == 'a'] <- x[y == 'b']
> x <- `[<-`(x, y == 'a', x[y == 'b'])
> cat(i, '')
> }
> cat('\n')
>
> The point of using this syntax is to make it work better with pipes, but the
> errors occur without pipes or magrittr.
Calling replacement functions this way is a Really Bad Idea. Some
assume they are being called properly and will end up mutating data
they should not when called this way.
Best,
luke
>
> Thank you for your help!
>
> __
> R-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
--
Luke Tierney
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa Phone: 319-335-3386
Department of Statistics andFax: 319-335-3017
Actuarial Science
241 Schaeffer Hall email: luke-tier...@uiowa.edu
Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Intermittent crashes with inset `[<-` command
Thanks for the report. Should be fixed shortly.
Best,
luke
On Tue, 26 Feb 2019, Brian Montgomery via R-devel wrote:
> The following code crashes after about 300 iterations on my
> x86_64-w64-mingw32 machine on R 3.5.2 --vanilla.
> Others have duplicated this (see
> https://github.com/tidyverse/magrittr/issues/190 if necessary), but I don't
> know how machine/OS-dependent it may be.
> If it doesn't crash for you, please try increasing the length of the x vector.
>
> Substituting the commented-out line for the one below it works correctly
> (prints out 1:1000 and ends normally) every time.
>
> x <- 1:20
> y <- rep(letters[1:5], length(x) / 5L)
> for (i in 1:1000) {
> # x[y == 'a'] <- x[y == 'b']
> x <- `[<-`(x, y == 'a', x[y == 'b'])
> cat(i, '')
> }
> cat('\n')
>
> The point of using this syntax is to make it work better with pipes, but the
> errors occur without pipes or magrittr.
>
> Thank you for your help!
>
> __
> R-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
--
Luke Tierney
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa Phone: 319-335-3386
Department of Statistics andFax: 319-335-3017
Actuarial Science
241 Schaeffer Hall email: luke-tier...@uiowa.edu
Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Intermittent crashes with inset `[<-` command
Some testing:
Adding `gc()` inside the for loop prevented a crash for 10,000+
iterations, whereas adding `Sys.sleep(.2)` (which takes longer) did
not. I couldn't wrap my head around the `vectorAssign` source code,
but I suspect it is a matter of an intermediate object not being
protected and being gc'ed. Hope that helps someone
Travers
Travers
On Wed, Feb 27, 2019 at 11:48 AM Travers Ching wrote:
>
> On an azure centos VM, I can reproduce this bug which reports either:
>
> *** caught segfault ***
> address 0x7006a, cause 'memory not mapped' (crash)
>
> Or
>
> incompatible types (from builtin to integer) in subassignment type fix
> (no crash)
>
> Like Gabriel, I could not reproduce the bug on a mac laptop. Both R
> versions 3.5.1.
>
> Travers
>
> On Wed, Feb 27, 2019 at 9:08 AM William Dunlap via R-devel
> wrote:
> >
> > Valgrind (without gctorture) reports memory misuse:
> >
> > % R --debugger=valgrind --debugger-args="--leak-check=full --num-callers=18"
> > ...
> > > x <- 1:20
> > > y <- rep(letters[1:5], length(x) / 5L)
> > > for (i in 1:1000) {
> > + # x[y == 'a'] <- x[y == 'b']
> > + x <- `[<-`(x, y == 'a', x[y == 'b'])
> > + cat(i, '')
> > + }
> > 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
> > 29 30 31 32 33 34 35 36 37 ==4711== Invalid read of size 1
> > ==4711==at 0x501A40F: Rf_xlength (Rinlinedfuns.h:542)
> > ==4711==by 0x501A40F: VectorAssign (subassign.c:658)
> > ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> > ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> > ==4711==by 0x4F66398: bcEval (eval.c:6795)
> > ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> > ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> > ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> > ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> > ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> > ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> > ==4711==by 0x40075A: main (Rmain.c:29)
> > ==4711== Address 0x19b3ab90 is 0 bytes inside a block of size 160,048
> > free'd
> > ==4711==at 0x4C2ACBD: free (vg_replace_malloc.c:530)
> > ==4711==by 0x4FAFCB2: ReleaseLargeFreeVectors (memory.c:1055)
> > ==4711==by 0x4FAFCB2: RunGenCollect (memory.c:1825)
> > ==4711==by 0x4FAFCB2: R_gc_internal (memory.c:2998)
> > ==4711==by 0x4FB166F: Rf_allocVector3 (memory.c:2682)
> > ==4711==by 0x4FB2310: Rf_allocVector (Rinlinedfuns.h:577)
> > ==4711==by 0x4FB2310: R_alloc (memory.c:2197)
> > ==4711==by 0x5023F7A: logicalSubscript (subscript.c:575)
> > ==4711==by 0x5026DA3: Rf_makeSubscript (subscript.c:994)
> > ==4711==by 0x501A2F3: VectorAssign (subassign.c:656)
> > ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> > ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> > ==4711==by 0x4F66398: bcEval (eval.c:6795)
> > ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> > ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> > ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> > ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> > ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> > ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> > ==4711==by 0x40075A: main (Rmain.c:29)
> > ==4711== Block was alloc'd at
> > ==4711==at 0x4C29BC3: malloc (vg_replace_malloc.c:299)
> > ==4711==by 0x4FB1B04: Rf_allocVector3 (memory.c:2712)
> > ==4711==by 0x5027574: Rf_allocVector (Rinlinedfuns.h:577)
> > ==4711==by 0x5027574: Rf_ExtractSubset (subset.c:115)
> > ==4711==by 0x502ADCD: VectorSubset (subset.c:198)
> > ==4711==by 0x502ADCD: do_subset_dflt (subset.c:823)
> > ==4711==by 0x502BE90: do_subset (subset.c:661)
> > ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> > ==4711==by 0x4F7BAC3: Rf_evalListKeepMissing (eval.c:2955)
> > ==4711==by 0x50200CB: R_DispatchOrEvalSP (subassign.c:1535)
> > ==4711==by 0x50200CB: do_subassign (subassign.c:1567)
> > ==4711==by 0x4F66398: bcEval (eval.c:6795)
> > ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> > ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> > ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> > ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> > ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> > ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> > ==4711==by 0x40075A: main (Rmain.c:29)
> > ==4711==
> > ==4711== Invalid read of size 8
> > ==4711==at 0x501A856: XLENGTH_EX (Rinlinedfuns.h:189)
> > ==4711==by 0x501A856: Rf_xlength (Rinlinedfuns.h:554)
> > ==4711==by 0x501A856: VectorAssign (subassign.c:658)
> > ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> > ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> > ==4711==by 0x4F66398: bcEval (eval.c:6795)
> > ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> > ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> > ==4711==
Re: [Rd] Intermittent crashes with inset `[<-` command
On an azure centos VM, I can reproduce this bug which reports either:
*** caught segfault ***
address 0x7006a, cause 'memory not mapped' (crash)
Or
incompatible types (from builtin to integer) in subassignment type fix
(no crash)
Like Gabriel, I could not reproduce the bug on a mac laptop. Both R
versions 3.5.1.
Travers
On Wed, Feb 27, 2019 at 9:08 AM William Dunlap via R-devel
wrote:
>
> Valgrind (without gctorture) reports memory misuse:
>
> % R --debugger=valgrind --debugger-args="--leak-check=full --num-callers=18"
> ...
> > x <- 1:20
> > y <- rep(letters[1:5], length(x) / 5L)
> > for (i in 1:1000) {
> + # x[y == 'a'] <- x[y == 'b']
> + x <- `[<-`(x, y == 'a', x[y == 'b'])
> + cat(i, '')
> + }
> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
> 29 30 31 32 33 34 35 36 37 ==4711== Invalid read of size 1
> ==4711==at 0x501A40F: Rf_xlength (Rinlinedfuns.h:542)
> ==4711==by 0x501A40F: VectorAssign (subassign.c:658)
> ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> ==4711==by 0x4F66398: bcEval (eval.c:6795)
> ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> ==4711==by 0x40075A: main (Rmain.c:29)
> ==4711== Address 0x19b3ab90 is 0 bytes inside a block of size 160,048
> free'd
> ==4711==at 0x4C2ACBD: free (vg_replace_malloc.c:530)
> ==4711==by 0x4FAFCB2: ReleaseLargeFreeVectors (memory.c:1055)
> ==4711==by 0x4FAFCB2: RunGenCollect (memory.c:1825)
> ==4711==by 0x4FAFCB2: R_gc_internal (memory.c:2998)
> ==4711==by 0x4FB166F: Rf_allocVector3 (memory.c:2682)
> ==4711==by 0x4FB2310: Rf_allocVector (Rinlinedfuns.h:577)
> ==4711==by 0x4FB2310: R_alloc (memory.c:2197)
> ==4711==by 0x5023F7A: logicalSubscript (subscript.c:575)
> ==4711==by 0x5026DA3: Rf_makeSubscript (subscript.c:994)
> ==4711==by 0x501A2F3: VectorAssign (subassign.c:656)
> ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> ==4711==by 0x4F66398: bcEval (eval.c:6795)
> ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> ==4711==by 0x40075A: main (Rmain.c:29)
> ==4711== Block was alloc'd at
> ==4711==at 0x4C29BC3: malloc (vg_replace_malloc.c:299)
> ==4711==by 0x4FB1B04: Rf_allocVector3 (memory.c:2712)
> ==4711==by 0x5027574: Rf_allocVector (Rinlinedfuns.h:577)
> ==4711==by 0x5027574: Rf_ExtractSubset (subset.c:115)
> ==4711==by 0x502ADCD: VectorSubset (subset.c:198)
> ==4711==by 0x502ADCD: do_subset_dflt (subset.c:823)
> ==4711==by 0x502BE90: do_subset (subset.c:661)
> ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> ==4711==by 0x4F7BAC3: Rf_evalListKeepMissing (eval.c:2955)
> ==4711==by 0x50200CB: R_DispatchOrEvalSP (subassign.c:1535)
> ==4711==by 0x50200CB: do_subassign (subassign.c:1567)
> ==4711==by 0x4F66398: bcEval (eval.c:6795)
> ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> ==4711==by 0x40075A: main (Rmain.c:29)
> ==4711==
> ==4711== Invalid read of size 8
> ==4711==at 0x501A856: XLENGTH_EX (Rinlinedfuns.h:189)
> ==4711==by 0x501A856: Rf_xlength (Rinlinedfuns.h:554)
> ==4711==by 0x501A856: VectorAssign (subassign.c:658)
> ==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
> ==4711==by 0x5020100: do_subassign (subassign.c:1571)
> ==4711==by 0x4F66398: bcEval (eval.c:6795)
> ==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
> ==4711==by 0x4F7DA70: do_for (eval.c:2185)
> ==4711==by 0x4F7741C: Rf_eval (eval.c:691)
> ==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
> ==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
> ==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
> ==4711==by 0x40075A: main (Rmain.c:29)
> ==4711== Address 0x19b3abb0 is 32 bytes inside a block of size 160,048
> free'd
> ==4711==at 0x4C2ACBD: free (vg_replace_malloc.c:530)
> ==4711==by 0x4FAFCB2: ReleaseLargeFreeVectors (memory.c:1055)
> ==4711==by 0x4FAFCB2: RunGenCollect (memory.c:1825)
> ==4711==by 0x4FAFCB2: R_gc_internal (memory.c:2998)
> ==4711==by
Re: [Rd] Intermittent crashes with inset `[<-` command
Valgrind (without gctorture) reports memory misuse:
% R --debugger=valgrind --debugger-args="--leak-check=full --num-callers=18"
...
> x <- 1:20
> y <- rep(letters[1:5], length(x) / 5L)
> for (i in 1:1000) {
+ # x[y == 'a'] <- x[y == 'b']
+ x <- `[<-`(x, y == 'a', x[y == 'b'])
+ cat(i, '')
+ }
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
29 30 31 32 33 34 35 36 37 ==4711== Invalid read of size 1
==4711==at 0x501A40F: Rf_xlength (Rinlinedfuns.h:542)
==4711==by 0x501A40F: VectorAssign (subassign.c:658)
==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
==4711==by 0x5020100: do_subassign (subassign.c:1571)
==4711==by 0x4F66398: bcEval (eval.c:6795)
==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
==4711==by 0x4F7DA70: do_for (eval.c:2185)
==4711==by 0x4F7741C: Rf_eval (eval.c:691)
==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
==4711==by 0x40075A: main (Rmain.c:29)
==4711== Address 0x19b3ab90 is 0 bytes inside a block of size 160,048
free'd
==4711==at 0x4C2ACBD: free (vg_replace_malloc.c:530)
==4711==by 0x4FAFCB2: ReleaseLargeFreeVectors (memory.c:1055)
==4711==by 0x4FAFCB2: RunGenCollect (memory.c:1825)
==4711==by 0x4FAFCB2: R_gc_internal (memory.c:2998)
==4711==by 0x4FB166F: Rf_allocVector3 (memory.c:2682)
==4711==by 0x4FB2310: Rf_allocVector (Rinlinedfuns.h:577)
==4711==by 0x4FB2310: R_alloc (memory.c:2197)
==4711==by 0x5023F7A: logicalSubscript (subscript.c:575)
==4711==by 0x5026DA3: Rf_makeSubscript (subscript.c:994)
==4711==by 0x501A2F3: VectorAssign (subassign.c:656)
==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
==4711==by 0x5020100: do_subassign (subassign.c:1571)
==4711==by 0x4F66398: bcEval (eval.c:6795)
==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
==4711==by 0x4F7DA70: do_for (eval.c:2185)
==4711==by 0x4F7741C: Rf_eval (eval.c:691)
==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
==4711==by 0x40075A: main (Rmain.c:29)
==4711== Block was alloc'd at
==4711==at 0x4C29BC3: malloc (vg_replace_malloc.c:299)
==4711==by 0x4FB1B04: Rf_allocVector3 (memory.c:2712)
==4711==by 0x5027574: Rf_allocVector (Rinlinedfuns.h:577)
==4711==by 0x5027574: Rf_ExtractSubset (subset.c:115)
==4711==by 0x502ADCD: VectorSubset (subset.c:198)
==4711==by 0x502ADCD: do_subset_dflt (subset.c:823)
==4711==by 0x502BE90: do_subset (subset.c:661)
==4711==by 0x4F7741C: Rf_eval (eval.c:691)
==4711==by 0x4F7BAC3: Rf_evalListKeepMissing (eval.c:2955)
==4711==by 0x50200CB: R_DispatchOrEvalSP (subassign.c:1535)
==4711==by 0x50200CB: do_subassign (subassign.c:1567)
==4711==by 0x4F66398: bcEval (eval.c:6795)
==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
==4711==by 0x4F7DA70: do_for (eval.c:2185)
==4711==by 0x4F7741C: Rf_eval (eval.c:691)
==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
==4711==by 0x40075A: main (Rmain.c:29)
==4711==
==4711== Invalid read of size 8
==4711==at 0x501A856: XLENGTH_EX (Rinlinedfuns.h:189)
==4711==by 0x501A856: Rf_xlength (Rinlinedfuns.h:554)
==4711==by 0x501A856: VectorAssign (subassign.c:658)
==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
==4711==by 0x5020100: do_subassign (subassign.c:1571)
==4711==by 0x4F66398: bcEval (eval.c:6795)
==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
==4711==by 0x4F7DA70: do_for (eval.c:2185)
==4711==by 0x4F7741C: Rf_eval (eval.c:691)
==4711==by 0x4FA7181: Rf_ReplIteration (main.c:258)
==4711==by 0x4FA7570: R_ReplConsole (main.c:308)
==4711==by 0x4FA760E: run_Rmainloop (main.c:1082)
==4711==by 0x40075A: main (Rmain.c:29)
==4711== Address 0x19b3abb0 is 32 bytes inside a block of size 160,048
free'd
==4711==at 0x4C2ACBD: free (vg_replace_malloc.c:530)
==4711==by 0x4FAFCB2: ReleaseLargeFreeVectors (memory.c:1055)
==4711==by 0x4FAFCB2: RunGenCollect (memory.c:1825)
==4711==by 0x4FAFCB2: R_gc_internal (memory.c:2998)
==4711==by 0x4FB166F: Rf_allocVector3 (memory.c:2682)
==4711==by 0x4FB2310: Rf_allocVector (Rinlinedfuns.h:577)
==4711==by 0x4FB2310: R_alloc (memory.c:2197)
==4711==by 0x5023F7A: logicalSubscript (subscript.c:575)
==4711==by 0x5026DA3: Rf_makeSubscript (subscript.c:994)
==4711==by 0x501A2F3: VectorAssign (subassign.c:656)
==4711==by 0x501CDFE: do_subassign_dflt (subassign.c:1641)
==4711==by 0x5020100: do_subassign (subassign.c:1571)
==4711==by 0x4F66398: bcEval (eval.c:6795)
==4711==by 0x4F7D86D: R_compileAndExecute (eval.c:1407)
==4711==by
Re: [Rd] Intermittent crashes with inset `[<-` command
Hello,
I can also reproduce this, R 3.5.2 on Ubuntu 18.04 LTS.
First run gives error after 148.
After 148:
Error in `[<-`(x, y == "a", x[y == "b"]) :
substituto tem comprimento zero
Execução interrompida
Translation:
replacement has length zero
Execution stopped
Second run gives a different error, I changed the script to start with a
sessionInfo() instruction, everything else is the same as in the OP.
After 180:
*** caught segfault ***
address 0x70002, cause 'memory not mapped'
An irrecoverable exception occurred. R is aborting now ...
Several runs with or without the sessionInfo() always give the two
different errors above.
The sessionInfo() was the following.
sessionInfo()
R version 3.5.2 (2018-12-20)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.1 LTS
Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1
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_3.5.2
Hope this helps,
Rui Barradas
Às 08:46 de 27/02/2019, Serguei Sokol escreveu:
On 26/02/2019 05:18, Brian Montgomery via R-devel wrote:
The following code crashes after about 300 iterations on
my x86_64-w64-mingw32 machine on R 3.5.2 --vanilla.
Others have duplicated this
(see https://github.com/tidyverse/magrittr/issues/190 if necessary),
but I don't know how machine/OS-dependent it may be.
It crashes too on my Mageia6 (RPM based Linux distribution):
184 185 186 187
*** caught segfault ***
address 0x70002, cause 'memory not mapped'
Possible actions:
1: abort (with core dump, if enabled)
2: normal R exit
3: exit R without saving workspace
4: exit R saving workspace
The crash can happen at different moments, sometimes after i=187 like in
the example above, sometimes after i=915. The error is not always
segfault. It can also be
915 Error in `[<-`(x, y == "a", x[y == "b"]) : replacement has length zero
or
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Error in
`[<-`(x, y == "a", x[y == "b"]) :
types (de raw a integer) incompatibles dans l'ajustement
d'affectation de type
(sorry, this crash was in french locale)
Hoping this helps.
Serguei.
> sessionInfo()
R version 3.5.2 (2018-12-20)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Mageia 6
Matrix products: default
BLAS/LAPACK: /home/opt/OpenBLAS/lib/libopenblas_sandybridge-r0.3.3.so
locale:
[1] C
attached base packages:
[1] stats graphics grDevices utils datasets methods base
loaded via a namespace (and not attached):
[1] compiler_3.5.2
If it doesn't crash for you, please try increasing the length of the x
vector.
Substituting the commented-out line for the one below it works
correctly (prints out 1:1000 and ends normally) every time.
x <- 1:20
y <- rep(letters[1:5], length(x) / 5L)
for (i in 1:1000) {
# x[y == 'a'] <- x[y == 'b']
x <- `[<-`(x, y == 'a', x[y == 'b'])
cat(i, '')
}
cat('\n')
The point of using this syntax is to make it work better with pipes,
but the errors occur without pipes or magrittr.
Thank you for your help!
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Intermittent crashes with inset `[<-` command
On Wed, 27 Feb 2019 at 09:51, Serguei Sokol wrote: > > On 26/02/2019 05:18, Brian Montgomery via R-devel wrote: > > The following code crashes after about 300 iterations on my > > x86_64-w64-mingw32 machine on R 3.5.2 --vanilla. > > Others have duplicated this (see > > https://github.com/tidyverse/magrittr/issues/190 if necessary), but I don't > > know how machine/OS-dependent it may be. > It crashes too on my Mageia6 (RPM based Linux distribution): > 184 185 186 187 > *** caught segfault *** > address 0x70002, cause 'memory not mapped' I can reproduce it too. This is the output from valgrind (sessionInfo below): ==3296== Invalid read of size 1 ==3296==at 0x4A2D7F7: UnknownInlinedFun (Rinlinedfuns.h:542) ==3296==by 0x4A2D7F7: VectorAssign (subassign.c:658) ==3296==by 0x4A30540: do_subassign_dflt (subassign.c:1641) ==3296==by 0x4A338F2: do_subassign (subassign.c:1571) ==3296==by 0x49769A1: bcEval (eval.c:6795) ==3296==by 0x498B415: R_compileAndExecute (eval.c:1407) ==3296==by 0x498B985: do_for (eval.c:2185) ==3296==by 0x49848A8: Rf_eval (eval.c:691) ==3296==by 0x49B5131: Rf_ReplIteration (main.c:258) ==3296==by 0x49B5131: Rf_ReplIteration (main.c:198) ==3296==by 0x49B54F0: R_ReplConsole (main.c:308) ==3296==by 0x49B55AF: run_Rmainloop (main.c:1082) ==3296==by 0x1090AE: main (Rmain.c:29) ==3296== Address 0x1dafab90 is 0 bytes inside a block of size 160,048 free'd ==3296==at 0x4839A0C: free (vg_replace_malloc.c:540) ==3296==by 0x49BCA56: ReleaseLargeFreeVectors (memory.c:1055) ==3296==by 0x49BCA56: RunGenCollect (memory.c:1825) ==3296==by 0x49BCA56: R_gc_internal (memory.c:2998) ==3296==by 0x49BCA56: R_gc_internal (memory.c:2964) ==3296==by 0x49BFB2C: Rf_allocVector3 (memory.c:2682) ==3296==by 0x49C09FC: UnknownInlinedFun (Rinlinedfuns.h:577) ==3296==by 0x49C09FC: R_alloc (memory.c:2197) ==3296==by 0x4A377F5: logicalSubscript (subscript.c:575) ==3296==by 0x4A377F5: logicalSubscript (subscript.c:503) ==3296==by 0x4A3A8D3: Rf_makeSubscript (subscript.c:994) ==3296==by 0x4A2D63D: VectorAssign (subassign.c:656) ==3296==by 0x4A30540: do_subassign_dflt (subassign.c:1641) ==3296==by 0x4A338F2: do_subassign (subassign.c:1571) ==3296==by 0x49769A1: bcEval (eval.c:6795) ==3296==by 0x498B415: R_compileAndExecute (eval.c:1407) ==3296==by 0x498B985: do_for (eval.c:2185) ==3296== Block was alloc'd at ==3296==at 0x483880B: malloc (vg_replace_malloc.c:309) ==3296==by 0x49C0031: Rf_allocVector3 (memory.c:2713) ==3296==by 0x4A3B041: UnknownInlinedFun (Rinlinedfuns.h:577) ==3296==by 0x4A3B041: Rf_ExtractSubset (subset.c:115) ==3296==by 0x4A3DA8A: VectorSubset (subset.c:198) ==3296==by 0x4A3DA8A: do_subset_dflt (subset.c:823) ==3296==by 0x4A3FCAA: do_subset (subset.c:661) ==3296==by 0x49848A8: Rf_eval (eval.c:691) ==3296==by 0x4989600: Rf_evalListKeepMissing (eval.c:2955) ==3296==by 0x4A3390B: R_DispatchOrEvalSP (subassign.c:1535) ==3296==by 0x4A3390B: do_subassign (subassign.c:1567) ==3296==by 0x49769A1: bcEval (eval.c:6795) ==3296==by 0x498B415: R_compileAndExecute (eval.c:1407) ==3296==by 0x498B985: do_for (eval.c:2185) ==3296==by 0x49848A8: Rf_eval (eval.c:691) ==3296== ==3296== Invalid read of size 1 ==3296==at 0x4A2E2C0: UnknownInlinedFun (Rinlinedfuns.h:189) ==3296==by 0x4A2E2C0: UnknownInlinedFun (Rinlinedfuns.h:554) ==3296==by 0x4A2E2C0: VectorAssign (subassign.c:658) ==3296==by 0x4A30540: do_subassign_dflt (subassign.c:1641) ==3296==by 0x4A338F2: do_subassign (subassign.c:1571) ==3296==by 0x49769A1: bcEval (eval.c:6795) ==3296==by 0x498B415: R_compileAndExecute (eval.c:1407) ==3296==by 0x498B985: do_for (eval.c:2185) ==3296==by 0x49848A8: Rf_eval (eval.c:691) ==3296==by 0x49B5131: Rf_ReplIteration (main.c:258) ==3296==by 0x49B5131: Rf_ReplIteration (main.c:198) ==3296==by 0x49B54F0: R_ReplConsole (main.c:308) ==3296==by 0x49B55AF: run_Rmainloop (main.c:1082) ==3296==by 0x1090AE: main (Rmain.c:29) ==3296== Address 0x1dafab90 is 0 bytes inside a block of size 160,048 free'd ==3296==at 0x4839A0C: free (vg_replace_malloc.c:540) ==3296==by 0x49BCA56: ReleaseLargeFreeVectors (memory.c:1055) ==3296==by 0x49BCA56: RunGenCollect (memory.c:1825) ==3296==by 0x49BCA56: R_gc_internal (memory.c:2998) ==3296==by 0x49BCA56: R_gc_internal (memory.c:2964) ==3296==by 0x49BFB2C: Rf_allocVector3 (memory.c:2682) ==3296==by 0x49C09FC: UnknownInlinedFun (Rinlinedfuns.h:577) ==3296==by 0x49C09FC: R_alloc (memory.c:2197) ==3296==by 0x4A377F5: logicalSubscript (subscript.c:575) ==3296==by 0x4A377F5: logicalSubscript (subscript.c:503) ==3296==by 0x4A3A8D3: Rf_makeSubscript (subscript.c:994) ==3296==by 0x4A2D63D: VectorAssign (subassign.c:656) ==3296==by 0x4A30540: do_subassign_dflt (subassign.c:1641) ==3296==by 0x4A338F2: do_subassign
Re: [Rd] Intermittent crashes with inset `[<-` command
On 26/02/2019 05:18, Brian Montgomery via R-devel wrote:
The following code crashes after about 300 iterations on my x86_64-w64-mingw32
machine on R 3.5.2 --vanilla.
Others have duplicated this (see
https://github.com/tidyverse/magrittr/issues/190 if necessary), but I don't
know how machine/OS-dependent it may be.
It crashes too on my Mageia6 (RPM based Linux distribution):
184 185 186 187
*** caught segfault ***
address 0x70002, cause 'memory not mapped'
Possible actions:
1: abort (with core dump, if enabled)
2: normal R exit
3: exit R without saving workspace
4: exit R saving workspace
The crash can happen at different moments, sometimes after i=187 like in
the example above, sometimes after i=915. The error is not always
segfault. It can also be
915 Error in `[<-`(x, y == "a", x[y == "b"]) : replacement has length zero
or
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Error in
`[<-`(x, y == "a", x[y == "b"]) :
types (de raw a integer) incompatibles dans l'ajustement
d'affectation de type
(sorry, this crash was in french locale)
Hoping this helps.
Serguei.
> sessionInfo()
R version 3.5.2 (2018-12-20)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Mageia 6
Matrix products: default
BLAS/LAPACK: /home/opt/OpenBLAS/lib/libopenblas_sandybridge-r0.3.3.so
locale:
[1] C
attached base packages:
[1] stats graphics grDevices utils datasets methods base
loaded via a namespace (and not attached):
[1] compiler_3.5.2
If it doesn't crash for you, please try increasing the length of the x vector.
Substituting the commented-out line for the one below it works correctly
(prints out 1:1000 and ends normally) every time.
x <- 1:20
y <- rep(letters[1:5], length(x) / 5L)
for (i in 1:1000) {
# x[y == 'a'] <- x[y == 'b']
x <- `[<-`(x, y == 'a', x[y == 'b'])
cat(i, '')
}
cat('\n')
The point of using this syntax is to make it work better with pipes, but the
errors occur without pipes or magrittr.
Thank you for your help!
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] Intermittent crashes with inset `[<-` command
Hi Brian,
I don't have a windows machine, but on a modern macbook pro I'm not able to
get this to crash in 3.5.1 or a very recent built-from-source R-devel. I
increased the length of x by 2 orders of magnitude but that just made the
loop take a lot longer to successfully run in both R versions i tested.
So seems like there is some OS/machine dependency at work here (?).
Best,
~G
On Wed, Feb 27, 2019 at 12:01 AM Brian Montgomery via R-devel <
r-devel@r-project.org> wrote:
> The following code crashes after about 300 iterations on
> my x86_64-w64-mingw32 machine on R 3.5.2 --vanilla.
> Others have duplicated this (see
> https://github.com/tidyverse/magrittr/issues/190 if necessary), but I
> don't know how machine/OS-dependent it may be.
> If it doesn't crash for you, please try increasing the length of the x
> vector.
>
> Substituting the commented-out line for the one below it works correctly
> (prints out 1:1000 and ends normally) every time.
>
> x <- 1:20
> y <- rep(letters[1:5], length(x) / 5L)
> for (i in 1:1000) {
> # x[y == 'a'] <- x[y == 'b']
> x <- `[<-`(x, y == 'a', x[y == 'b'])
> cat(i, '')
> }
> cat('\n')
>
> The point of using this syntax is to make it work better with pipes, but
> the errors occur without pipes or magrittr.
>
> Thank you for your help!
>
> __
> R-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
[[alternative HTML version deleted]]
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
