R devel - Nov 2018 - Suggestion: Make CRAN source URLs immutable

On Wed, 24 Oct 2018 at 11:40, Kurt Hornik <Kurt.Hornik at wu.ac.at>
wrote:
>
> >>>>> Kurt Wheeler writes:
>
> Try e.g.
>
> https://cran.r-project.org/package=httr&version=1.3.1
> https://cran.r-project.org/package=httr&version=1.3.0
This is a nice feature that I didn't know. I recently proposed
enforcing this scheme in Fedora's packaging guidelines, because in
this way, a SPEC would build correctly even if the package was updated
and the old version was archived (is this guaranteed to continue to
work in future? I assumed so...).

There is an odd thing about this format though, and that is the
absence of a file extension. This is a redirection, yes, but the
spectool can't trust the filename that is sent by the remote server,
and uses only the filename extracted from the URL.

Without extension, RPM doesn't know how to unpack the sources. So we
have to do the following similarly odd trick (note the "#"):

Source0:
https://cran.r-project.org/package=%{packname}&version=%{version}#/%{packname}_%{version}.tar.gz

Did you consider this problem? Is there any alternate immutable URL
*with* extension? If not, is there any potential issue with the trick
above?

Regards,
--
I?aki Ucar

FWIW, you can get the URL and extract the link with extension from there.
Archived packages are always tarballs, so that makes the following possible:

url <- "https://cran.r-project.org/package=httr&version=1.3.0"

library(RCurl)

pkgurl <- gsub(".*(https://cran.+\\.tar.gz).*",
               "\\1",
               getURL(url))

install.packages(pkgurl, type = "source", repos = NULL)

Cheers
Joris


On Fri, Nov 2, 2018 at 5:42 PM I?aki Ucar <iucar at fedoraproject.org>
wrote:
> On Wed, 24 Oct 2018 at 11:40, Kurt Hornik <Kurt.Hornik at wu.ac.at>
wrote:
> >
> > >>>>> Kurt Wheeler writes:
> >
> > Try e.g.
> >
> > https://cran.r-project.org/package=httr&version=1.3.1
> > https://cran.r-project.org/package=httr&version=1.3.0
>
> This is a nice feature that I didn't know. I recently proposed
> enforcing this scheme in Fedora's packaging guidelines, because in
> this way, a SPEC would build correctly even if the package was updated
> and the old version was archived (is this guaranteed to continue to
> work in future? I assumed so...).
>
> There is an odd thing about this format though, and that is the
> absence of a file extension. This is a redirection, yes, but the
> spectool can't trust the filename that is sent by the remote server,
> and uses only the filename extracted from the URL.
>
> Without extension, RPM doesn't know how to unpack the sources. So we
> have to do the following similarly odd trick (note the "#"):
>
> Source0:
>
https://cran.r-project.org/package=%{packname}&version=%{version}#/%{packname}_%{version}.tar.gz
>
> Did you consider this problem? Is there any alternate immutable URL
> *with* extension? If not, is there any potential issue with the trick
> above?
>
> Regards,
> --
> I?aki Ucar
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Joris Meys
Statistical consultant

Department of Data Analysis and Mathematical Modelling
Ghent University
Coupure Links 653, B-9000 Gent (Belgium)
<https://maps.google.com/?q=Coupure+links+653,%C2%A0B-9000+Gent,%C2%A0Belgium&entry=gmail&source=g>

-----------
Biowiskundedagen 2017-2018
http://www.biowiskundedagen.ugent.be/

-------------------------------
Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php

	[[alternative HTML version deleted]]

On Sat, 3 Nov 2018 at 11:54, Joris Meys <jorismeys at gmail.com>
wrote:
>
> FWIW, you can get the URL and extract the link with extension from there.
Archived packages are always tarballs, so that makes the following possible:
>
> url <-
"https://cran.r-project.org/package=httr&version=1.3.0"
>
> library(RCurl)
>
> pkgurl <- gsub(".*(https://cran.+\\.tar.gz).*",
>                "\\1",
>                getURL(url))
>
> install.packages(pkgurl, type = "source", repos = NULL)
The proper way to do this would be to fetch just the headers and
extract the location field:

$ URL="https://cran.r-project.org/package=httr&version=1.3.0"
$ curl -sI $URL | awk '/^Location:/{print $2}'
https://cran.r-project.org/src/contrib/Archive/httr/httr_1.3.0.tar.gz

But that's not the point. I'm talking about RPM packaging, and the
point is, as I said, that the tool that expands and downloads sources
from specfiles simply doesn't do that, because it can't trust an
extension sent from a remote server. We need to provide it explicitly
in the URL.

--
I?aki ?car