[Bioc-devel] svn and building packages under R for Windows 2.12.0

[Hervé Pagès]

Hi Gordon,

Make sure you are using the make command from Rtools212. Those commands
that are packaged in Rtools are basically the same as the Cygwin's ones
but sometimes they are slightly different (the R people decide to not
use the latest version and/or to recompile the command with slightly
different settings).

Also it looks like Sophos Anti-Virus is in the list of anti-viruses
known to interfere with Cygwin:

   http://cygwin.com/faq/faq.using.html#faq.using.bloda

Not sure that will help you a lot, sorry :-/

H.

On 10/21/2010 04:33 PM, Gordon K Smyth wrote:
> Hi Mike,
>
> Yes, your suggestion was correct. When I remove the executables the
> error is now introduced on my original copy.
>
> My computer is running Sophos Anti-Virus. Unfortunately, it is corporate
> policy at my institute that anti virus software be active at all times.
> I don't have permission to turn it off. Even if I could turn it off, I
> wouldn't want to do so on an ongoing basis.
>
> Not sure when to go to from here.
>
> Cheers
> Gordon
>
> On Thu, 21 Oct 2010, Mike Smith wrote:
>
>> Hi Gordon,
>>
>> If you remove the compiled files from your local copy (everything in
>> limma/src other than normexp.c) is the error introduced? It appears that
>> gcc isn't being invoked on your local copy as the source file has all
>> ready
>> been compiled, which isn't the case with a fresh bioconductor download.
>>
>> I'm sure you've done some Googling, but there appear to be loads of posts
>> with similar errors that have turned out to be related to various
>> anti-virus
>> / malware software preventing cygwin from running - particularly Symantec
>> software. If you have anything like that on the machine, maybe try
>> turning
>> it off and running again.
>>
>> Sorry that's not a more definitive answer,
>>
>> Mike Smith
>>
>> On Thu, Oct 21, 2010 at 4:28 AM, Gordon K Smyth <smyth at wehi.edu.au>
>> wrote:
>>
>>> Here's a little mystery with R 2.12.0 and Windows.
>>>
>>> I have no trouble building or checking my copy of the limma package
>>> (version 3.7.1) under R for Windows 2.12.0. I do get a cygwin warning
>>> which
>>> is new with R 2.12.0, but the package still builds fine. The output
>>> from R
>>> CMD INSTALL is given below.
>>>
>>> However, after I commit limma to the Bioconductor repository, then
>>> check it
>>> out to a new directory, I cannot build or check the checked out version
>>> using R 2.12.0. The error message is given below.
>>>
>>> I get the same behaviour whether I work with the Bioconductor 2.7 or
>>> Bioconductor 2.8 versions of limma and the corresponding svn
>>> repositories.
>>> In particular, I cannot build the Biodonductor 2.7 Release of limma
>>> 2.6.0
>>> under R 2.12.0 for Windows.
>>>
>>> My version of R tools appears to be up to date, as I re-installed it 7
>>> October 2010.
>>>
>>> I have tried copying non-versioned files from my original version to the
>>> checked out version to ensure they contain exactly the same files.
>>> Doesn't
>>> help.
>>>
>>> If I copy the checked out version to a unix directory, then it builds
>>> and
>>> checks fine using R 2.12.0 for linux.
>>>
>>> I've never seen this sort of error before. All my files have MS-DOS
>>> style
>>> end of lines, except for a file of C source code with Unix style line
>>> ends.
>>> These all seem to be preserved in the original and checked out versions.
>>>
>>> Any ideas?
>>> Gordon
>>>
>>>
>>> COMMAND PROMPT SESSION USING MY ORIGINAL COPY LIMMA:
>>>
>>> R CMD INSTALL --build limma
>>>>
>>> * installing to library 'c:/lib/R/R-2.12.0/library'
>>> * installing *source* package 'limma' ...
>>> ** libs
>>> cygwin warning:
>>> MS-DOS style path detected: c:/lib/R/R-2.12.0/etc/i386/Makeconf
>>> Preferred POSIX equivalent is:
>>> /cygdrive/c/lib/R/R-2.12.0/etc/i386/Makeconf
>>> CYGWIN environment variable option "nodosfilewarning" turns off this
>>> warning.
>>> Consult the user's guide for more details about POSIX paths:
>>> http://cygwin.com/cygwin-ug-net/using.html#using-pathnames
>>> make: Nothing to be done for `all'.
>>> installing to c:/lib/R/R-2.12.0/library/limma/libs/i386
>>> ** R
>>> ** inst
>>> ** preparing package for lazy loading
>>> ** help
>>> *** installing help indices
>>> ** building package indices ...
>>> ** testing if installed package can be loaded
>>> * MD5 sums
>>> packaged installation of 'limma' as limma_3.7.1.zip
>>>
>>> * DONE (limma)
>>>
>>>
>>>
>>> COMMAND PROMPT SESSION AFTER CHECKING LIMMA OUT FROM BIOCONDUCTOR 2.8:
>>>
>>> R CMD INSTALL --build limma
>>>>
>>> * installing to library 'c:/lib/R/R-2.12.0/library'
>>> * installing *source* package 'limma' ...
>>> ** libs
>>> cygwin warning:
>>> MS-DOS style path detected: c:/lib/R/R-2.12.0/etc/i386/Makeconf
>>> Preferred POSIX equivalent is:
>>> /cygdrive/c/lib/R/R-2.12.0/etc/i386/Makeconf
>>> CYGWIN environment variable option "nodosfilewarning" turns off this
>>> warning.
>>> Consult the user's guide for more details about POSIX paths:
>>> http://cygwin.com/cygwin-ug-net/using.html#using-pathnames
>>> gcc -I"c:/lib/R/R-2.12.0/include" -O3 -Wall -std=gnu99 -c
>>> normexp.c -o
>>> normexp.o
>>> 142 [main] make 4420 c:\lib\Rtools\bin\make.exe: *** fatal error -
>>> couldn't
>>> allocate heap, Win32 error 487, base 0x480000, top 0x520000,
>>> reserve_size
>>> 651264
>>> , allocsize 655360, page_const 4096
>>> Stack trace:
>>> Frame Function Args
>>> 0022E4C0 6102749B (0022E4C0, 00000000, 00000000, 00000000)
>>> 0022E7B0 6102749B (61177B80, 00008000, 00000000, 61179977)
>>> 0022F7E0 61004AFB (611AF64C, 00480000, 00520000, 0009F000)
>>> 0022F810 6106BEB3 (7FFEFFFF, 00000003, 00000002, 7C90DA0A)
>>> 0022F900 610BDF9B (0000078C, 02000000, 61163664, 61163660)
>>> 0022F930 61006090 (00000000, 00000000, 00000000, 00000000)
>>> 0022F9F0 6106D0E9 (61000000, 00000001, 0022FD30, 6106CE10)
>>> 0022FA10 7C90118A (6106CE10, 61000000, 00000001, 0022FD30)
>>> 0022FB18 7C91C4FA (0022FD30, 7FFDE000, 7FFDF000, 00000000)
>>> 0022FC94 7C9211B4 (0022FD30, 7C900000, 0022FCE0, 7C90E900)
>>> 0022FD1C 7C9210AF (0022FD30, 7C900000, 00000000, 00010017)
>>> End of stack trace
>>> 79 [main] make 4780 fork: child -1 - died waiting for longjmp before
>>> initia
>>> lization, retry 0, exit code 0x100, errno 11
>>> make: vfork: Resource temporarily unavailable
>>> ERROR: compilation failed for package 'limma'
>>> * removing 'c:/lib/R/R-2.12.0/library/limma'
>>> * restoring previous 'c:/lib/R/R-2.12.0/library/limma'
>>>
>>>
>>> PATH INFORMATION FOR MY COMMAND SESSION:
>>>
>>>
>>> PATH=c:\lib\Rtools\bin;c:\lib\Rtools\perl\bin;c:\lib\Rtools\MinGW\bin;c:\lib\bin
>>>
>>> ;c:\lib\R\R-2.12.0\bin;c:\progra~1\HTML Help
>>> workshop\;C:\WINDOWS\system32;C:\WI
>>> NDOWS;C:\WINDOWS\System32\Wbem;C:\lib\MiKTeX 2.6\miktex\bin;C:\lib\Inno
>>> Setup 5;
>>>
>>> ______________________________________________________________________
>>> The information in this email is confidential and intend...{{dropped:4}}
>>>
>>> _______________________________________________
>>> Bioc-devel at stat.math.ethz.ch mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>
>>
>
> ______________________________________________________________________
> The information in this email is confidential and inte...{{dropped:21}}