peter dalgaard
2015-May-11 08:31 UTC
[Rd] Wrongly checked MD5 checksums in R 3.2.0's windows binary
On 11 May 2015, at 09:53 , Martin Maechler <maechler at lynne.stat.math.ethz.ch> wrote:>>>>>> Tal Galili <tal.galili at gmail.com> >>>>>> on Sun, 10 May 2015 23:13:09 +0300 writes: > >> Dear R-devel members, Several R user recently reported >> <https://github.com/talgalili/installr/issues/30> (while >> using the installr >> <http://cran.r-project.org/web/packages/installr>package) >> that when running MD5 checksums on a recent R installation >> (R 3.2.0), they get that the files ?bin/R.exe?, >> ?bin/Rscript.exe? have the wrong MD5 checksums on Windows. > >> I intend to remove the test for these files in the next >> version of installr, but I thought this might interest >> people to know. > > Well, one could argue that this is a bug in your package. >...especially since this seems to be about executable files _after_ install, not the installer binaries. I don't think we provide the md5sums for those, do we? -- Peter Dalgaard, Professor, Center for Statistics, Copenhagen Business School Solbjerg Plads 3, 2000 Frederiksberg, Denmark Phone: (+45)38153501 Office: A 4.23 Email: pd.mes at cbs.dk Priv: PDalgd at gmail.com
Duncan Murdoch
2015-May-11 09:06 UTC
[Rd] Wrongly checked MD5 checksums in R 3.2.0's windows binary
On 11/05/2015 4:31 AM, peter dalgaard wrote:> > On 11 May 2015, at 09:53 , Martin Maechler <maechler at lynne.stat.math.ethz.ch> wrote: > >>>>>>> Tal Galili <tal.galili at gmail.com> >>>>>>> on Sun, 10 May 2015 23:13:09 +0300 writes: >> >>> Dear R-devel members, Several R user recently reported >>> <https://github.com/talgalili/installr/issues/30> (while >>> using the installr >>> <http://cran.r-project.org/web/packages/installr>package) >>> that when running MD5 checksums on a recent R installation >>> (R 3.2.0), they get that the files ?bin/R.exe?, >>> ?bin/Rscript.exe? have the wrong MD5 checksums on Windows. >> >>> I intend to remove the test for these files in the next >>> version of installr, but I thought this might interest >>> people to know. >> >> Well, one could argue that this is a bug in your package. >> > > ...especially since this seems to be about executable files _after_ install, not the installer binaries. I don't think we provide the md5sums for those, do we? >Yes, they're in the MD5 file in the top level directory after installing. The issue with files in the bin directory is that they depend on whether it's a 32 bit or 64 bit install, but the MD5 sums are computed only for the 32 bit files. Those two files should be identical to each other, and to one of i386/Rfe.exe or x64/Rfe.exe. They need to be special-cased. Duncan Murdoch
Tal Galili
2015-May-11 12:35 UTC
[Rd] Wrongly checked MD5 checksums in R 3.2.0's windows binary
Thank you Duncan, Peter and Martin for the responses. Just to mention that the code is based on tools::md5sum, and the issue can be reproduced (in Windows) using: if(!require(installr)) install.packages("installr") installr::checkMD5sums2(dir=R.home()) With regards, Tal [[alternative HTML version deleted]]
Reasonably Related Threads
- Wrongly checked MD5 checksums in R 3.2.0's windows binary
- Wrongly checked MD5 checksums in R 3.2.0's windows binary
- Wrongly checked MD5 checksums in R 3.2.0's windows binary
- Wrongly checked MD5 checksums in R 3.2.0's windows binary
- Wrongly checked MD5 checksums in R 3.2.0's windows binary