Re: [Rd] unit testing for R packages?
If you haven't done so already, you might want to read the RUnit Vignette: http://cran.r-project.org/web/packages/RUnit/vignettes/RUnit.pdf. On a somewhat related note: is anyone doing continuous integration with their R packages? I'm thinking about trying to use something like CruiseControl (which I used before for Java development), and was wondering if anyone had experience with how much work will be involved... On Mon, Oct 5, 2009 at 3:01 PM, Blair Christian blair.christ...@gmail.comwrote: Hi All, I'm interested in putting some unit tests into an R package I'm building. I have seen assorted things such as Runit library, svUnit library, packages with 'tests' directories, etc I grep'd unit test through the writing R extensions manual but didn't find anything. Are there any suggestions out there? Currently I have several (a lot?) classes/methods that I keep tinkering with, and I'd like to run a script frequently to check that I don't cause any unforeseen problems. Right now I'm writing the classes/methods, but will be packaging it up soon. Thx, Blair __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel [[alternative HTML version deleted]] __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
If you haven't done so already, you might want to read the RUnit Vignette: http://cran.r-project.org/web/packages/RUnit/vignettes/RUnit.pdf. On a somewhat related note: is anyone doing continuous integration with their R packages? I'm thinking about trying to use something like CruiseControl (which I used before for Java development), and was wondering if anyone had experience with how much work will be involved... It's not quite continuous integration, but I've been working on some code for autotesting - i.e. any time your code or tests change, rerun the appropriate part of the test suite (as much as I can guess at that) Hadley -- http://had.co.nz/ __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On a related note: I recently started to set up CruiseControl ( http://cruisecontrol.sourceforge.net/) to do continuous integration with my R packages (along with my RUnit test cases). Has anyone had any luck with this before? I use it for other software and it provides a very nice interface... On Tue, Oct 6, 2009 at 10:33 AM, hadley wickham h.wick...@gmail.com wrote: If you haven't done so already, you might want to read the RUnit Vignette: http://cran.r-project.org/web/packages/RUnit/vignettes/RUnit.pdf. On a somewhat related note: is anyone doing continuous integration with their R packages? I'm thinking about trying to use something like CruiseControl (which I used before for Java development), and was wondering if anyone had experience with how much work will be involved... It's not quite continuous integration, but I've been working on some code for autotesting - i.e. any time your code or tests change, rerun the appropriate part of the test suite (as much as I can guess at that) Hadley -- http://had.co.nz/ [[alternative HTML version deleted]] __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On 10/5/2009 3:01 PM, Blair Christian wrote: Hi All, I'm interested in putting some unit tests into an R package I'm building. I have seen assorted things such as Runit library, svUnit library, packages with 'tests' directories, etc I grep'd unit test through the writing R extensions manual but didn't find anything. Are there any suggestions out there? Nothing very specific, but what there is is described in that manual in the discussion of the tests subdirectory. Duncan Murdoch Currently I have several (a lot?) classes/methods that I keep tinkering with, and I'd like to run a script frequently to check that I don't cause any unforeseen problems. Right now I'm writing the classes/methods, but will be packaging it up soon. Thx, Blair __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
Hi, On Mon, Oct 5, 2009 at 12:01 PM, Blair Christian blair.christ...@gmail.com wrote: I'm interested in putting some unit tests into an R package I'm building. I have seen assorted things such as Runit library, svUnit library, packages with 'tests' directories, etc I grep'd unit test through the writing R extensions manual but didn't find anything. Are there any suggestions out there? Currently I have several (a lot?) classes/methods that I keep tinkering with, and I'd like to run a script frequently to check that I don't cause any unforeseen problems. I've had good experiences using RUnit. To date, I've mostly used RUnit by putting tests in inst/unitTests and creating a Makefile there to run the tests. You should also be able to use RUnit in a more interactive fashion inside an interactive R session in which you are doing development. The vignette in svUnit has an interesting approach for integrating unit testing into R CMD check via examples in an Rd file within the package. + seth -- Seth Falcon | @sfalcon | http://userprimary.net/user __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
Seth I should look at RUnit more carefully sometime, but having had my own system in place from much earlier, I've never made the move. Basically I just use files in the tests/ directory which are run by a makefile in the package directory above. Roughly, this just means that I can run the tests quickly before building and checking the package. Is there a (very short) explanation of the main things that RUnit gives beyond this? Thanks, Paul -Original Message- From: r-devel-boun...@r-project.org [mailto:r-devel-boun...@r- project.org] On Behalf Of Seth Falcon Sent: October 5, 2009 3:48 PM To: Blair Christian Cc: r-devel@r-project.org Subject: Re: [Rd] unit testing for R packages? Hi, On Mon, Oct 5, 2009 at 12:01 PM, Blair Christian blair.christ...@gmail.com wrote: I'm interested in putting some unit tests into an R package I'm building. I have seen assorted things such as Runit library, svUnit library, packages with 'tests' directories, etc I grep'd unit test through the writing R extensions manual but didn't find anything. Are there any suggestions out there? Currently I have several (a lot?) classes/methods that I keep tinkering with, and I'd like to run a script frequently to check that I don't cause any unforeseen problems. I've had good experiences using RUnit. To date, I've mostly used RUnit by putting tests in inst/unitTests and creating a Makefile there to run the tests. You should also be able to use RUnit in a more interactive fashion inside an interactive R session in which you are doing development. The vignette in svUnit has an interesting approach for integrating unit testing into R CMD check via examples in an Rd file within the package. + seth -- Seth Falcon | @sfalcon | http://userprimary.net/user __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel La version française suit le texte anglais. This email may contain privileged and/or confidential information, and the Bank of Canada does not waive any related rights. Any distribution, use, or copying of this email or the information it contains by other than the intended recipient is unauthorized. If you received this email in error please delete it immediately from your system and notify the sender promptly by email that you have done so. Le présent courriel peut contenir de l'information privilégiée ou confidentielle. La Banque du Canada ne renonce pas aux droits qui s'y rapportent. Toute diffusion, utilisation ou copie de ce courriel ou des renseignements qu'il contient par une personne autre que le ou les destinataires désignés est interdite. Si vous recevez ce courriel par erreur, veuillez le supprimer immédiatement et envoyer sans délai à l'expéditeur un message électronique pour l'aviser que vous avez éliminé de votre ordinateur toute copie du courriel reçu. __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On 05/10/2009 4:39 PM, spencerg wrote:
I put unit test in the examples, using \dontshow to hide stopifnot.
Many help pages I've written contain code like the following:
A - functionDocumentedHere()
B - manuallyComputedAnswer
\dontshow{stopifnot(}
all.equal(A, B)
\dontshow{)}
This will fail in a future release of R, because those aren't valid
expressions within \dontshow{}, which expects R code. You can achieve
the same effect using the clearer
all.equal(A,B)
\dontshow{ stopifnot(isTrue(.Last.value)) }
Duncan Murdoch
I think it helps the documentation to include an example comparing
a special case computed using a function with a manual computation.
However, stopifnot contributes nothing to user understanding, so I
hide it. One could also use \dontshow to hide entire examples that
check trivial details you think would not interest users.
Spencer
Seth Falcon wrote:
Hi,
On Mon, Oct 5, 2009 at 12:01 PM, Blair Christian
blair.christ...@gmail.com wrote:
I'm interested in putting some unit tests into an R package I'm
building. I have seen assorted things such as Runit library, svUnit
library, packages
with 'tests' directories, etc
I grep'd unit test through the writing R extensions manual but didn't find
anything. Are there any suggestions out there? Currently I have
several (a lot?) classes/methods that I keep tinkering with, and I'd
like to run a script frequently to check that I don't cause any
unforeseen problems.
I've had good experiences using RUnit. To date, I've mostly used
RUnit by putting tests in inst/unitTests and creating a Makefile there
to run the tests. You should also be able to use RUnit in a more
interactive fashion inside an interactive R session in which you are
doing development.
The vignette in svUnit has an interesting approach for integrating
unit testing into R CMD check via examples in an Rd file within the
package.
+ seth
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On 05/10/2009 6:06 PM, spencerg wrote:
Hi, Duncan:
Thanks for the warning. Can you give me a hint of which release
might require this? In particular, will it be R 2.10.0, coming quite
soon?
The alpha version of 2.10.0 lets this pass, and I can't see changing
that now. But 2.11.0 (due next spring) may well enforce the rules that
have been documented but unenforced since before the release of 2.9.0.
Duncan Murdoch
Thanks,
Spencer
Duncan Murdoch wrote:
On 05/10/2009 4:39 PM, spencerg wrote:
I put unit test in the examples, using \dontshow to hide
stopifnot. Many help pages I've written contain code like the
following:
A - functionDocumentedHere()
B - manuallyComputedAnswer
\dontshow{stopifnot(}
all.equal(A, B)
\dontshow{)}
This will fail in a future release of R, because those aren't valid
expressions within \dontshow{}, which expects R code. You can achieve
the same effect using the clearer
all.equal(A,B)
\dontshow{ stopifnot(isTrue(.Last.value)) }
Duncan Murdoch
I think it helps the documentation to include an example
comparing a special case computed using a function with a manual
computation. However, stopifnot contributes nothing to user
understanding, so I hide it. One could also use \dontshow to hide
entire examples that check trivial details you think would not
interest users.
Spencer
Seth Falcon wrote:
Hi,
On Mon, Oct 5, 2009 at 12:01 PM, Blair Christian
blair.christ...@gmail.com wrote:
I'm interested in putting some unit tests into an R package I'm
building. I have seen assorted things such as Runit library, svUnit
library, packages
with 'tests' directories, etc
I grep'd unit test through the writing R extensions manual but
didn't find
anything. Are there any suggestions out there? Currently I have
several (a lot?) classes/methods that I keep tinkering with, and I'd
like to run a script frequently to check that I don't cause any
unforeseen problems.
I've had good experiences using RUnit. To date, I've mostly used
RUnit by putting tests in inst/unitTests and creating a Makefile there
to run the tests. You should also be able to use RUnit in a more
interactive fashion inside an interactive R session in which you are
doing development.
The vignette in svUnit has an interesting approach for integrating
unit testing into R CMD check via examples in an Rd file within the
package.
+ seth
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
Hi, I am the writer of svUnit. My initial goal was to build functions on top of RUnit which is older, and certainly deserves all the credit for core test unit functions. Unfortunately, the way RUnit is working internally did not allowed me to build the extensions I needed. Now, as Seth already told, there are several functions in svUnit that ease interactive use of the tests in an R session. What he did not told is that RUnit has some nice code coverage functions that svUnit does not have. Thus, not easy to decide... The good news is that both RUnit and svUnit are almost completely compatible with test unit code, that is, the same code put in inst/unitTests should run with both packages. So, you could write your test units, try both approaches in practice, and finally decide after a little bit of experiment which one you prefer (at least, read the vignette of each packages). In case you use Komodo Edit as R code editor, there is a nice GUI interactive panel to svUnit available (see http://www.sciviews.org/SciViews-K). Best, Philippe Grosjean Seth Falcon wrote: Hi, On Mon, Oct 5, 2009 at 12:01 PM, Blair Christian blair.christ...@gmail.com wrote: I'm interested in putting some unit tests into an R package I'm building. I have seen assorted things such as Runit library, svUnit library, packages with 'tests' directories, etc I grep'd unit test through the writing R extensions manual but didn't find anything. Are there any suggestions out there? Currently I have several (a lot?) classes/methods that I keep tinkering with, and I'd like to run a script frequently to check that I don't cause any unforeseen problems. I've had good experiences using RUnit. To date, I've mostly used RUnit by putting tests in inst/unitTests and creating a Makefile there to run the tests. You should also be able to use RUnit in a more interactive fashion inside an interactive R session in which you are doing development. The vignette in svUnit has an interesting approach for integrating unit testing into R CMD check via examples in an Rd file within the package. + seth __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On Mon, Oct 5, 2009 at 5:44 PM, hadley wickham h.wick...@gmail.com wrote: Now, as Seth already told, there are several functions in svUnit that ease interactive use of the tests in an R session. What he did not told is that RUnit has some nice code coverage functions that svUnit does not have. What are those? I've skimmed through the documentation and couldn't see anything? The R wiki has information about both RUnit and svUnit. I'm not sure that it details all of the differences. I have successfully used the methodology of the gdata package described there. http://wiki.r-project.org/rwiki/doku.php?id=developers:runit Peter Hadley -- http://had.co.nz/ __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] unit testing for R packages?
On Mon, Oct 5, 2009 at 8:48 PM, Peter Cowan cowan...@gmail.com wrote: On Mon, Oct 5, 2009 at 5:44 PM, hadley wickham h.wick...@gmail.com wrote: Now, as Seth already told, there are several functions in svUnit that ease interactive use of the tests in an R session. What he did not told is that RUnit has some nice code coverage functions that svUnit does not have. What are those? I've skimmed through the documentation and couldn't see anything? The R wiki has information about both RUnit and svUnit. I'm not sure that it details all of the differences. I have successfully used the methodology of the gdata package described there. Sorry, I should have been more explicit. I was interested particularly in the code coverage functions. Hadley -- http://had.co.nz/ __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
