cygwin and tar -- still?

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

cygwin and tar -- still?

Jeff Hamann
I've upgraded laptops and have put a new version of R on the new machine as
well as the tools for building packages. I know R has required a
"non-broken" version of tar for some time now and was hoping to not have to
install any cygwin tools. Since building packages requires the use of MinGW
and MSYS, why can't we simply use the tar that's part of msys?

After looking at the following results:

E:\conifers>rcmd build --binary rconifers
* checking for file 'rconifers/DESCRIPTION' ... OK
* preparing 'rconifers':
* checking DESCRIPTION meta-information ... OK
* cleaning src
* checking whether 'INDEX' is up-to-date ... NO
* use '--force' to overwrite the existing 'INDEX'
* removing junk files
tar: /cygdrive/E/conifers/rconifers_0.0-5.tar: Cannot open: No such file or
directory
tar: Error is not recoverable: exiting now
tar: /cygdrive/E/conifers/rconifers_0.0-5.tar: Cannot open: No such file or
directory
tar: Error is not recoverable: exiting now
Error: cannot open file 'rconifers/DESCRIPTION' for reading

E:\conifers>

I still don't understand why the cygwin tool is required. I'm just a mere
simpleton and haven't spent the time required to become king-fu master with
this particular minutiae. Could the script that builds the package be fixed?
Is it broken? If it's not broken, what's with the cygdrive stuff? why does
the entire path need to be there?

Is the tar that's in the tools.zip from
http://www.murdoch-sutherland.com/Rtools/ any different than the one with
MSYS?

Jeff.

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Reply | Threaded
Open this post in threaded view
|

Re: cygwin and tar -- still?

Uwe Ligges
Jeff D. Hamann wrote:
> I've upgraded laptops and have put a new version of R on the new machine as
> well as the tools for building packages. I know R has required a
> "non-broken" version of tar for some time now and was hoping to not have to
> install any cygwin tools. Since building packages requires the use of MinGW
> and MSYS, why can't we simply use the tar that's part of msys?

No, "MSYS" is not required nor supported.


> After looking at the following results:
>
> E:\conifers>rcmd build --binary rconifers
> * checking for file 'rconifers/DESCRIPTION' ... OK
> * preparing 'rconifers':
> * checking DESCRIPTION meta-information ... OK
> * cleaning src
> * checking whether 'INDEX' is up-to-date ... NO
> * use '--force' to overwrite the existing 'INDEX'
> * removing junk files
> tar: /cygdrive/E/conifers/rconifers_0.0-5.tar: Cannot open: No such file or
> directory
> tar: Error is not recoverable: exiting now
> tar: /cygdrive/E/conifers/rconifers_0.0-5.tar: Cannot open: No such file or
> directory
> tar: Error is not recoverable: exiting now
> Error: cannot open file 'rconifers/DESCRIPTION' for reading
>
> E:\conifers>
>
> I still don't understand why the cygwin tool is required. I'm just a mere
> simpleton and haven't spent the time required to become king-fu master with
> this particular minutiae. Could the script that builds the package be fixed?
> Is it broken? If it's not broken, what's with the cygdrive stuff? why does
> the entire path need to be there?
>
> Is the tar that's in the tools.zip from
> http://www.murdoch-sutherland.com/Rtools/ any different than the one with
> MSYS?

Yes, it is different and it works.

Simply don't use MSYS, but use the recommended tools.


Uwe Ligges



>
> Jeff.
>
> ______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel