Hi,

I can confirm that most of those "strange R CMD build/check errors"
we observe on Windows are actually a consequence of the "temp
Rscript file collision" I reported yesterday here:

  https://stat.ethz.ch/pipermail/r-devel/2010-September/058648.html

I just applied the following patch to the R we use for the Bioconductor
Windows builds:

Index: src/gnuwin32/system.c
===================================================================
--- src/gnuwin32/system.c       (revision 53067)
+++ src/gnuwin32/system.c       (working copy)
@@ -821,6 +821,8 @@

 #include <sys/stat.h>
 #include <time.h>
+#include <sys/types.h>
+#include <unistd.h>
 static int isDir(char *path)
 {
     struct stat sb;
@@ -1093,7 +1095,7 @@
                }
            }
            srand( (unsigned) time(NULL) );
-           sprintf(ifile, "%s/Rscript%x%x", tm, rand(), rand());
+           sprintf(ifile, "%s/Rscript%dx%x", tm, getpid(), rand());
            ifp = fopen(ifile, "w+b");
if(!ifp) R_Suicide(_("creation of tmpfile failed -- set TMPDIR suitably?"));
        }

and the "strange R CMD build/check errors" are now all gone, except
those of the type:

  R\bin\R.exe CMD build baySeq
  * checking for file 'baySeq/DESCRIPTION' ... OK
  * preparing 'baySeq':
  * checking DESCRIPTION meta-information ... OK
  * installing the package to re-build vignettes
  * creating vignettes ...Warning in file(con, "r") :
cannot open file 'E:\biocbld\bbs-2.7-bioc\tmpdir\RtmpL9YTVu\xshell45ae180c': Permission denied
  Error in file(con, "r") : cannot open the connection
  Execution halted

I didn't initially report failures of this particular type because I
thought they might be yet another manifestation of the same bug but
apparently they are not (looks like it could be a temp file collision
again, but a different temp file now), so I'm going to report them in
a separate post.

Can the patch above be applied? Thanks!

H.


On 09/11/2010 11:10 PM, Hervé Pagès wrote:
Hi,

This is a follow up to:

https://stat.ethz.ch/pipermail/r-devel/2010-July/057921.html

The Bioconductor daily builds have been reporting a lot of strange
things lately on Windows using R-2.12. This started 2 or 3 months
ago and things are not getting better with recent R-2.12.
Here is a sample from today's build results. We use Windows Server
2003 R2 for the 32-bit builds, Windows Server 2008 R2 Enterprise
for the 64-bit builds, and R-2.12 (2010-08-26 r52817) on both machines:


 >>> On 64-bit Windows:

R\bin\R.exe CMD build ACME
--> produces no output at all and returns code 0.


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings affypdnn_1.23.0.tar.gz
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/affypdnn.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* using option '--no-vignettes'
* checking for file 'affypdnn/DESCRIPTION' ... OK
* this is package 'affypdnn' version '1.23.0'
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'affypdnn' can be installed ...Warning in
file(con, "r") :
cannot open file
'E:/biocbld/bbs-2.7-bioc/meat/affypdnn.Rcheck/00install.out': No such
file or directory
Error in file(con, "r") : cannot open the connection
Execution halted


 >>> On 32-bit Windows:

R\bin\R.exe CMD build ArrayTools
* checking for file 'ArrayTools/DESCRIPTION' ... OK
* preparing 'ArrayTools':
* checking DESCRIPTION meta-information ... OK
* installing the package to re-build vignettes
-----------------------------------
Warning: unknown option '-l'
* checking for file
'E:\biocbld\bbs-2.7-bioc\tmpdir\Rtmpg0MxFa\Rinst69801ce4/DESCRIPTION'
... NO
-----------------------------------
ERROR: Installation failed
Removing installation dir


 >>> On 64-bit Windows:

R\bin\R.exe CMD build baySeq
* checking for file 'baySeq/DESCRIPTION' ... OK
* preparing 'baySeq':
* checking DESCRIPTION meta-information ... OK
* installing the package to re-build vignettes
* creating vignettes ... ERROR
Error: ERROR: no packages specified


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings BioSeqClass_1.7.0.tar.gz
Warning: unknown option '--no-vignettes'
Warning: unknown option '--timings'
Converting Rd files to LaTeX ...
BioSeqClass_1.7.0.tar.gz
Warning in readLines(f) :
incomplete final line found on 'BioSeqClass_1.7.0.tar.gz'
Warning in parse_Rd("BioSeqClass_1.7.0.tar.gz", encoding = "unknown",
fragment = FALSE, :
BioSeqClass_1.7.0.tar.gz:2: unexpected UNKNOWN '\Â'
Warning: BioSeqClass_1.7.0.tar.gz:1: All text must be in a section
Warning: BioSeqClass_1.7.0.tar.gz:2: All text must be in a section
Warning: BioSeqClass_1.7.0.tar.gz:2: All text must be in a section
Warning: BioSeqClass_1.7.0.tar.gz:3: All text must be in a section
Warning: BioSeqClass_1.7.0.tar.gz:4: All text must be in a section
Error : BioSeqClass_1.7.0.tar.gz: Sections \title, and \name must exist
and be unique in Rd files


 >>> On 64-bit Windows:

R\bin\R.exe CMD build BufferedMatrix
--> produces no output at all and returns code 0.


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings daMA_1.21.0.tar.gz
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/daMA.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* using option '--no-vignettes'
* checking for file 'daMA/DESCRIPTION' ... OK
* this is package 'daMA' version '1.21.0'
* checking package name space information ... OK
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'daMA' can be installed ...Warning in
file(con, "r") :
cannot open file
'E:/biocbld/bbs-2.7-bioc/meat/daMA.Rcheck/00install.out': No such file
or directory
Error in file(con, "r") : cannot open the connection
Execution halted


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings ddCt_1.3.2.tar.gz
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/ddCt.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* using option '--no-vignettes'
* checking for file 'ddCt/DESCRIPTION' ... OK
* this is package 'ddCt' version '1.3.2'
* checking package name space information ... OK
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'ddCt' can be installed ...Warning in
file(con, "r") :
cannot open file
'E:/biocbld/bbs-2.7-bioc/meat/ddCt.Rcheck/00install.out': No such file
or directory
Error in file(con, "r") : cannot open the connection
Execution halted


 >>> On 64-bit Windows:

R\bin\R.exe CMD build GEOmetadb
* checking for file 'GEOmetadb/DESCRIPTION' ... OK
* preparing 'GEOmetadb':
* checking DESCRIPTION meta-information ... OK
* installing the package to re-build vignettes
* creating vignettes ... ERROR
Error: ERROR: no packages specified


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings IRanges_1.7.33.tar.gz
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/IRanges.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* using option '--no-vignettes'
* checking for file 'IRanges/DESCRIPTION' ... OK
* this is package 'IRanges' version '1.7.33'
* checking package name space information ... OK
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'IRanges' can be installed ... ERROR
Installation failed.
See 'E:/biocbld/bbs-2.7-bioc/meat/IRanges.Rcheck/00install.out' for
details.

Content of IRanges.Rcheck/00install.out:

* install options '--no-html --no-multiarch'

Warning: unknown option '-l'
Warning: unknown option '--no-html'
Warning: unknown option '--no-multiarch'
Error in .pkg2tex(files, outfile, encoding = encoding, append = append, :
this package does not have either a 'latex' or a (source) 'man' directory


 >>> On 32-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings keggorthology_2.1.1.tar.gz
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/keggorthology.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* using option '--no-vignettes'
* checking for file 'keggorthology/DESCRIPTION' ... OK
* this is package 'keggorthology' version '2.1.1'
* checking package name space information ... OK
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'keggorthology' can be installed ... ERROR
Installation failed.
See 'E:/biocbld/bbs-2.7-bioc/meat/keggorthology.Rcheck/00install.out'
for details.

Content of keggorthology.Rcheck/00install.out:

* install options '--no-html --no-multiarch'

Warning: unknown option '--no-html'
* using log directory 'E:/biocbld/bbs-2.7-bioc/meat/KEGGOR˜1.Rcheck'
* using R version 2.12.0 Under development (unstable) (2010-08-26 r52817)
* using platform: i386-pc-mingw32 (32-bit)
* using session charset: ISO8859-1
* checking for file 'KEGGOR˜1/DESCRIPTION' ... OK
* this is package 'keggorthology' version '2.1.1'
* checking package name space information ... OK
* checking package dependencies ... OK
* checking if this is a source package ... OK
* checking whether package 'keggorthology' can be installed ... OK
* checking package directory ... OK
* checking for portable file names ... OK
* checking DESCRIPTION meta-information ... OK
* checking top-level files ... OK
* checking index information ... OK
* checking package subdirectories ... OK
* checking R files for non-ASCII characters ... OK
* checking R files for syntax errors ... OK
* checking whether the package can be loaded ... ERROR
Error in library(KEGGOR ˜ 1) : 'package' must be of length 1
Execution halted

It looks like this package has a loading problem: see the messages for
details.


 >>> On 64-bit Windows:

R\bin\R.exe CMD check --no-vignettes --timings OLIN_1.27.0.tar.gz
Warning: unknown option '--no-vignettes'
Warning: unknown option '--timings'
Converting Rd files to LaTeX ...
OLIN_1.27.0.tar.gz
Warning in readLines(f) :
incomplete final line found on 'OLIN_1.27.0.tar.gz'
Warning in parse_Rd("OLIN_1.27.0.tar.gz", encoding = "unknown", fragment
= FALSE, :
OLIN_1.27.0.tar.gz:2: unexpected '}'
Warning in parse_Rd("OLIN_1.27.0.tar.gz", encoding = "unknown", fragment
= FALSE, :
OLIN_1.27.0.tar.gz:4: unexpected '}'
Warning in parse_Rd("OLIN_1.27.0.tar.gz", encoding = "unknown", fragment
= FALSE, :
OLIN_1.27.0.tar.gz:7: unexpected '{'
Warning in parse_Rd("OLIN_1.27.0.tar.gz", encoding = "unknown", fragment
= FALSE, :
OLIN_1.27.0.tar.gz:8: unexpected '}'
Warning: OLIN_1.27.0.tar.gz:1: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:2: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:2: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:2: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:2: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:3: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:4: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:4: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:4: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:5: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:6: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:7: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:7: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:7: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:8: All text must be in a section
Warning: OLIN_1.27.0.tar.gz:8: All text must be in a section
Error : OLIN_1.27.0.tar.gz: Sections \title, and \name must exist and be
unique in Rd files


 >>> and so on...


AFAICT those problems were never seen before (i.e. with R < 2.12).
They show up randomly everyday for a small number of packages
(between 10 and 20 out of 400). The set of victims changes everyday
and any package seems to be a potential victim (I've not been able
to observe any obvious pattern so far).
Does anyone have any idea what could make 'R CMD build' and
'R CMD check' so confused/unreliable on Windows?

Thanks,
H.



--
Hervé Pagès

Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M2-B876
P.O. Box 19024
Seattle, WA 98109-1024

E-mail: hpa...@fhcrc.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319

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

Reply via email to