runit-help Mailing List for R Unit Test Framework

Suppose one test in a suite has an error.  Normally, RUnit records this
outcome and moves on to the next test.

Alerted to the problem, I want to investigate.  I'd like to repeat the test,
but this time have R enter the debugger
to check variables etc., as it does when it hits stop() outside of RUnit.
Does RUnit have a provision for this?

If not, I think it could be implemented by passing a new argument to disable
"try".

/Christian

Hi Christian,

Thank you for reporting this issue.

Christian Brechbühler wrote:
> RUnit 0.4.17 comes with tests/runitInspect.r,v 1.3.  (It defines
> .TearDown, the inspection target foo, and the two test cases
> testRUnit.inspect and testRUnit.getTrackInfo.)
> 
> The Problem:  Both test cases get an error.  Here is the complete output:
> 
>     RUnit test cases for 'RUnit:inspect' functions
> 
> 
> 
>     Executing test function testRUnit.getTrackInfo  ... Error in
> get(x, envir, mode, inherits) : variable "foo" was not found
>      done successfully.
> 
> 
> 
>     Executing test function testRUnit.inspect  ... Error in get(x,
> envir, mode, inherits) : variable "foo" was not found
>     In addition: Warning message:
>     remove: variable "track" was not found
>      done successfully.
> 
>     Number of test functions: 2
>     Number of errors: 0
>     Number of failures: 0
>     Warning message:
>     remove: variable "track" was not found
> 
> 
> 
> Apparently the error gets caught but not reported; as the total shows
> zero errors.
> I tried under two releases of R:
 >
>     R 2.4.1 (Patched), 2007-03-25, svn.rev 40917, x86_64-unknown-linux-gnu
>     R 2.6.0 (Patched), 2007-11-01, svn.rev 43319, x86_64-unknown-linux-gnu
>
Nothing to do with R version but with RUnit version. With 0.4.16 the 
sandbox was introduced but the inspect test case not updated accordingly.

> My explanation: The function ".sourceTestFile" sources the test file
> in the environment "sandbox", which is passed as argument "envir" to
> ".executeTestCase".  That function "get"s the test function from
> "envir", but executes it in its own environment in the following line.
> 
>     timing <- try(system.time(func()))
> 
> It fails inside "inspect" with the message
> 
>     Error in get(fname) : variable "foo" was not found
> 
> BTW, checkTrue( exists("foo")) in testRUnit.getTrackInfo passes.
> 
> I wonder how it worked previously, because it must have when the test
> was created.
> Things I tried:
> * In testRUnit.inspect,   get("foo", get("envir", sys.frame(-1)))
> will sucessfully find the right function (as confirmed by calling
>   it without inspect).  But inspect then cannot resolve it.
> * renamed "foo" to make sure it's unique.  (I had other functions
> named "foo" lying around.)   No difference.
> 
> Do the tests pass for others?  All other test cases in the other test
> files pass, just these two   Could my installation be messed up?
> 
Be assured the failure is caused by a RUnit problem not your setup.

The issue is that inside inspect get() looks in the wrong place *iff*
executed inside the sandbox, that is in the context of a unit test case.
So in interactive mode the same code, copied from the test case to the R 
prompt works w/o issues.
The failure does not trigger an error report as it is itself 
encapsulated inside a try statement outside the tracking mechanism.

I have a fix but will take more time to test it.

Having said that please note that the inspect mechanism is not heavily 
used and tested as it is rather impractical as only one function call at 
a time can be tracked. This is far from being a code coverage tool.

Regards

   Matthias

> Thanks for any insights.
> /Christian
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Runit-help mailing list
> Run...@li...
> https://lists.sourceforge.net/lists/listinfo/runit-help
> 


-- 
Matthias Burger                     Project Manager/ Biostatistician
Epigenomics AG    Kleine Praesidentenstr. 1    10178 Berlin, Germany
phone:+49-30-24345-371                          fax:+49-30-24345-555
http://www.epigenomics.com           mat...@ep...
--
Epigenomics AG Berlin           Amtsgericht Charlottenburg HRB 75861
Vorstand:   Geert Nygaard (CEO/Vorsitzender),  Dr. Kurt Berlin (CSO)
               Oliver Schacht PhD (CFO),  Christian Piepenbrock (COO)
Aufsichtsrat:   Prof. Dr. Dr. hc. Rolf Krebs (Chairman/Vorsitzender)

On Nov 24, 2007 9:46 PM, Christian Brechb=FChler <bre...@gm...>
wrote:

>   In addition: Warning message:
>    remove: variable "track" was not found
>     done successfully.


I forgot to say, that problem is simple to fix, e.g.,

 .tearDown <- function() {
>    if (exists("track", envir=3D.GlobalEnv)) {
> -    rm(track, envir=3D.GlobalEnv)
> +    rm(track)
>    }
>  }
>

or with inherits=3DTRUE.

The problem remains how to make the test case work, so that it can exercise
inspect.
/Christian

RUnit 0.4.17 comes with tests/runitInspect.r,v 1.3.  (It defines
.TearDown, the inspection target foo, and the two test cases
testRUnit.inspect and testRUnit.getTrackInfo.)

The Problem:  Both test cases get an error.  Here is the complete output:

    RUnit test cases for 'RUnit:inspect' functions



    Executing test function testRUnit.getTrackInfo  ... Error in
get(x, envir, mode, inherits) : variable "foo" was not found
     done successfully.



    Executing test function testRUnit.inspect  ... Error in get(x,
envir, mode, inherits) : variable "foo" was not found
    In addition: Warning message:
    remove: variable "track" was not found
     done successfully.

    Number of test functions: 2
    Number of errors: 0
    Number of failures: 0
    Warning message:
    remove: variable "track" was not found



Apparently the error gets caught but not reported; as the total shows
zero errors.
I tried under two releases of R:

    R 2.4.1 (Patched), 2007-03-25, svn.rev 40917, x86_64-unknown-linux-gnu
    R 2.6.0 (Patched), 2007-11-01, svn.rev 43319, x86_64-unknown-linux-gnu

My explanation: The function ".sourceTestFile" sources the test file
in the environment "sandbox", which is passed as argument "envir" to
".executeTestCase".  That function "get"s the test function from
"envir", but executes it in its own environment in the following line.

    timing <- try(system.time(func()))

It fails inside "inspect" with the message

    Error in get(fname) : variable "foo" was not found

BTW, checkTrue( exists("foo")) in testRUnit.getTrackInfo passes.

I wonder how it worked previously, because it must have when the test
was created.
Things I tried:
* In testRUnit.inspect,   get("foo", get("envir", sys.frame(-1)))
will sucessfully find the right function (as confirmed by calling
  it without inspect).  But inspect then cannot resolve it.
* renamed "foo" to make sure it's unique.  (I had other functions
named "foo" lying around.)   No difference.

Do the tests pass for others?  All other test cases in the other test
files pass, just these two   Could my installation be messed up?

Thanks for any insights.
/Christian

Hi Jim,

see
?DEACTIVATED

For RUnit test case and test function are synonymous concepts.
There is no mechanism to deactivate individual check* function calls and 
report those as deactivated.

In your example checkEquals(c2f(0), 10) is evaluated, fails and thus 
interrupts the test function execution and signals a FAILURE to the test 
runner. DEACTIVATED() has no role here.

Intention of DEACTIVATED is to remind any developer of test functions 
that exist but are not executed due to some known issue to be eventually 
resolved.

test.c2f <- function() {
    checkEquals(c2f(0), 32)
    DEACTIVATED("next check needs fixing")
    checkEquals(c2f(0), 10)
    checkEquals(c2f(10), 50)
    checkException(c2f("xx"))
  }

The alternative praxis is to comment out the test function code for now 
and later forget about it altogether.

It might have its real benefit only for larger test suites (several 
packages) and several developers where the info that some code was 
tested but currently is expected to fail is useful.


Regards,

   Matthias


Jim...@bd... wrote:
> Hi all,
> 
> I've just discovered RUnit, and I'm trying it on examples to make sure I
> understand how to use it.  I'm excited to find it and would like to promote
> it among the R programmers in my company.
> 
> I find that I'm not getting results I expect using DEACTIVATED.  Is this in
> fact a bug, or am I doing something wrong?  Thanks in advance.
> 
> Here's an example, based on the example in the vignette.  I have an R
> function
> 
> c2f <- function(Centigrade) {
>   9/5 * Centigrade + 32
> }
> 
> and a test function
> 
> test.c2f <- function() {
>   checkEquals(c2f(0), 32)
>   DEACTIVATED(checkEquals(c2f(0), 10))
>   checkEquals(c2f(10), 50)
>   checkException(c2f("xx"))
> }
> 
> in file "runitc2f.r".  Basically, I've just added a test case that is
> expected to fail, and wrapped it in "DEACTIVATED", to see if the test case
> is actually turned off.
> 
> In this case, running
> 
> printTextProtocol(runTestFile("runitc2f.r"))
> 
> yields
> 
> Number of test functions: 1
> Number of errors: 0
> Number of failures: 1
> 
> and the 1 failure is of course the case that is "DEACTIVATED".
> 
> If I change that line to
> 
> DEACTIVATED(checkEquals(c2f(0), 32))
> 
> so that it should pass, should it be executed, I get
> 
> Number of test functions: 0
> Number of deactivated test functions: 1
> Number of errors: 0
> Number of failures: 0
> 
> I would have expected the number of test functions to be 1 in both cases,
> and 1 case to be deactivated, whether the line in question would have
> failed or not.  Is my expectation off-base?  If so, what is the intended
> purpose of "DEACTIVATED," and could anyone clarify how I should use it?
> 
> Thanks,
> 
> Jim Garrett
> 
> -----------------------------------------
> *******************************************************************
> IMPORTANT MESSAGE FOR RECIPIENTS IN THE
> U.S.A.:
> This message may constitute an advertisement of
> a BD group's products or services or a
> solicitation of interest in them. If this is
> such a message and you would like to opt out of
> receiving future advertisements or
> solicitations from this BD group, please
> forward this e-mail to opt...@bd....
> *******************************************************************
> This message (which includes any attachments)
> is intended only for the designated
> recipient(s).  It may contain confidential or
> proprietary information and may be subject to
> the attorney-client
> privilege or other confidentiality protections.
>  If you are not a designated recipient, you may
> not review, use, copy or distribute this
> message. If you received this in error, please
> notify the sender by reply e-mail and delete
> this message. Thank you.
> *******************************************************************
> Corporate Headquarters Mailing Address: BD
> (Becton, Dickinson and Company) 1 Becton Drive
> Franklin Lakes, NJ 07417 U.S.A.
> *******************************************************************
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Runit-help mailing list
> Run...@li...
> https://lists.sourceforge.net/lists/listinfo/runit-help
> 


-- 
Matthias Burger                     Project Manager/ Biostatistician
Epigenomics AG    Kleine Praesidentenstr. 1    10178 Berlin, Germany
phone:+49-30-24345-371                          fax:+49-30-24345-555
http://www.epigenomics.com           mat...@ep...
--
Epigenomics AG Berlin           Amtsgericht Charlottenburg HRB 75861
Vorstand:   Geert Nygaard (CEO/Vorsitzender),  Dr. Kurt Berlin (CSO)
               Oliver Schacht PhD (CFO),  Christian Piepenbrock (COO)
Aufsichtsrat:   Prof. Dr. Dr. hc. Rolf Krebs (Chairman/Vorsitzender)

Hi all,

I've just discovered RUnit, and I'm trying it on examples to make sure I
understand how to use it.  I'm excited to find it and would like to promote
it among the R programmers in my company.

I find that I'm not getting results I expect using DEACTIVATED.  Is this in
fact a bug, or am I doing something wrong?  Thanks in advance.

Here's an example, based on the example in the vignette.  I have an R
function

c2f <- function(Centigrade) {
  9/5 * Centigrade + 32
}

and a test function

test.c2f <- function() {
  checkEquals(c2f(0), 32)
  DEACTIVATED(checkEquals(c2f(0), 10))
  checkEquals(c2f(10), 50)
  checkException(c2f("xx"))
}

in file "runitc2f.r".  Basically, I've just added a test case that is
expected to fail, and wrapped it in "DEACTIVATED", to see if the test case
is actually turned off.

In this case, running

printTextProtocol(runTestFile("runitc2f.r"))

yields

Number of test functions: 1
Number of errors: 0
Number of failures: 1

and the 1 failure is of course the case that is "DEACTIVATED".

If I change that line to

DEACTIVATED(checkEquals(c2f(0), 32))

so that it should pass, should it be executed, I get

Number of test functions: 0
Number of deactivated test functions: 1
Number of errors: 0
Number of failures: 0

I would have expected the number of test functions to be 1 in both cases,
and 1 case to be deactivated, whether the line in question would have
failed or not.  Is my expectation off-base?  If so, what is the intended
purpose of "DEACTIVATED," and could anyone clarify how I should use it?

Thanks,

Jim Garrett

-----------------------------------------
*******************************************************************
IMPORTANT MESSAGE FOR RECIPIENTS IN THE
U.S.A.:
This message may constitute an advertisement of
a BD group's products or services or a
solicitation of interest in them. If this is
such a message and you would like to opt out of
receiving future advertisements or
solicitations from this BD group, please
forward this e-mail to opt...@bd....
*******************************************************************
This message (which includes any attachments)
is intended only for the designated
recipient(s).  It may contain confidential or
proprietary information and may be subject to
the attorney-client
privilege or other confidentiality protections.
 If you are not a designated recipient, you may
not review, use, copy or distribute this
message. If you received this in error, please
notify the sender by reply e-mail and delete
this message. Thank you.
*******************************************************************
Corporate Headquarters Mailing Address: BD
(Becton, Dickinson and Company) 1 Becton Drive
Franklin Lakes, NJ 07417 U.S.A.
*******************************************************************

--- Tony Plate <tp...@ac... wrote:

Gregor, the unit test stuff on the
R Wiki works the same under Windows 

> XP as it does under Linux.

> 

>
However, I've found a problem with it: if you run the check supplying a 

> library location, as in

> 

> R CMD check -l "...some library location..."
myPackage

> 

> then tests/doUnitTests.R is not able to find the tests. 
I think using 

> system.file() to find out where the library is installed
is going to be 

> more robust.



Great! Could you please fix the script
in R Wiki?



Thanks, Gregor

Gregor, the unit test stuff on the R Wiki works the same under Windows 
XP as it does under Linux.

However, I've found a problem with it: if you run the check supplying a 
library location, as in

R CMD check -l "...some library location..." myPackage

then tests/doUnitTests.R is not able to find the tests.  I think using 
system.file() to find out where the library is installed is going to be 
more robust.

Here's my current version of tests/doUnitTests.R:

if(require("RUnit", quietly=TRUE)) { ## unit tests will not be done if 
RUnit is not available

   ## --- Setup ---
   wd <- getwd()
   pkg <- basename(sub(pattern="tests$", replacement="", wd))
   pkg <- sub(pattern="\.Rcheck$", replacement="", pkg)
   library(package=pkg, character.only=TRUE)
   if(Sys.getenv("RCMDCHECK") == "FALSE") {
     ## Path for standalone i.e. not by R CMD check testing
     path <- file.path(wd, "..", "inst", "unitTests")
   } else {
     path <- system.file(package=pkg, "unitTests")
     if (path=="")
         stop("could not find 'unitTests' directory in installed ", pkg, 
" (", system.file(package=pkg), ")")
   }
   cat("Running unit tests in ", path, "\n", sep="")
   cat("Using RUnit from ", system.file(package="RUnit"), "\n", sep="")

   ## --- Load the name space to allow testing of private functions ---
   ## (not for "unit tests", but can be useful for internal tests)
   # if (is.element(pkg, loadedNamespaces()))
   #     attach(loadNamespace(pkg), name=paste("namespace", pkg, 
sep=":"), pos=3)

   ## --- Testing ---

   ## Define tests
   testSuite <- defineTestSuite(name=paste(pkg, "unit testing"), dirs=path)

   ## Run
   tests <- runTestSuite(testSuite)

   ## Print results
   printTextProtocol(tests, showDetails=FALSE)
   printTextProtocol(tests, fileName=file.path(path, 
"reportSummary.txt"), showDetails=FALSE)
   printTextProtocol(tests, fileName=file.path(path, 
"reportDetailed.txt"), showDetails=TRUE)

   ## Print HTML version to a file
   # printHTMLProtocol(tests, fileName=file.path(path, "report.html"))

   ## Return stop() if there are any failures i.e. FALSE to unit test.
   ## This will cause R CMD check to return error and stop
   if(getErrors(tests)$nFail > 0) {
     stop("one of unit tests failed")
   }
} else {
   warning("cannot run tests -- package RUnit is not available")
}


Gregor Gorjanc wrote:
> Hi,
> 
> Tony Plate wrote:
>> Hi Gregor, OK, that answers one of the questions I sent.  If I get the
>> unit tests as described in the R Wiki working under Windows XP, I'll put
>> a note in the Wiki.
> 
> That will be great!
> 
>> The other question I had was whether you knew if RUnit supports
>> transcript-style tests? (i.e., textual matching of R output in
>> transcripts)  I thought I saw something about this on one of the R
>> mailing lists recently, but I can't find it anywhere when I look now.
> 
> No, as far as I know. I might be wrong here. However regular test
> provided by R have textual matching or R output. You can of course mixe
> these two aproaches.
> 
> Regards, Gregor
> 

Matthias Burger wrote:
...
>> I might be wrong here. However regular test
>> provided by R have textual matching or R output. You can of course mixe
>> these two aproaches.
> Given infrequent complaints from some members of R core I'd suggest not
> to do so, at least not to use the 'tests' folder for unit tests but
> rather 'unitTests'  or 'UnitTests' (e.g. used by some BioConductor
> packages) and to put them under the inst/ folder to provide the end user
> with a means to test on his/her platform as well.

When I said mix the approaches I meant use tests folder for textual
matching and something else for unit tests. I do not know the reason for
 complaints by R core. Anyway, I do follow the same proposal i.e. using
inst/unitTests and R Wiki has a page describing how to "trick" the R CMD
check to also perform unitTests.

Gregor

Gregor Gorjanc wrote:
> Hi,
> 
> Tony Plate wrote:
>> Hi Gregor, OK, that answers one of the questions I sent.  If I get the
>> unit tests as described in the R Wiki working under Windows XP, I'll put
>> a note in the Wiki.
> 
> That will be great!
> 
>> The other question I had was whether you knew if RUnit supports
>> transcript-style tests? (i.e., textual matching of R output in
>> transcripts)  I thought I saw something about this on one of the R
>> mailing lists recently, but I can't find it anywhere when I look now.
> 
> No, as far as I know. 
You're correct.

>I might be wrong here. However regular test
> provided by R have textual matching or R output. You can of course mixe
> these two aproaches.
Given infrequent complaints from some members of R core I'd suggest not
to do so, at least not to use the 'tests' folder for unit tests but
rather 'unitTests'  or 'UnitTests' (e.g. used by some BioConductor
packages) and to put them under the inst/ folder to provide the end user
with a means to test on his/her platform as well.


Best,

  Matthias

> Regards, Gregor
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Runit-help mailing list
> Run...@li...
> https://lists.sourceforge.net/lists/listinfo/runit-help
> 


-- 
Matthias Burger                     Project Manager/ Biostatistician
Epigenomics AG    Kleine Praesidentenstr. 1    10178 Berlin, Germany
phone:+49-30-24345-371                          fax:+49-30-24345-555
http://www.epigenomics.com           mat...@ep...
--
Epigenomics AG Berlin           Amtsgericht Charlottenburg HRB 75861
Vorstand:   Geert Nygaard (CEO/Vorsitzender),  Dr. Kurt Berlin (CSO)
              Oliver Schacht PhD (CFO),  Christian Piepenbrock (COO)
Aufsichtsrat:   Prof. Dr. Dr. hc. Rolf Krebs (Chairman/Vorsitzender)

Hi,

Tony Plate wrote:
> Hi Gregor, OK, that answers one of the questions I sent.  If I get the
> unit tests as described in the R Wiki working under Windows XP, I'll put
> a note in the Wiki.

That will be great!

> The other question I had was whether you knew if RUnit supports
> transcript-style tests? (i.e., textual matching of R output in
> transcripts)  I thought I saw something about this on one of the R
> mailing lists recently, but I can't find it anywhere when I look now.

No, as far as I know. I might be wrong here. However regular test
provided by R have textual matching or R output. You can of course mixe
these two aproaches.

Regards, Gregor

Hi Iago,

by definition test functions cannot have arguments, i.e. the test case
runner will obtain the test function from the local environment and call
it with no arguments.

Your alternative would not work as foo has to be a new function for each
class in order to have the corresponding 'myclass' value set correctly.
(appart from the error in the function where I believe you intended
foo <- function(object) {
     checkTrue( is(class(object), 'myclass'))
}

but

foo <- function(object, className) {
     checkTrue( is(class(object), className))
}

would work in

test.myclass <- function() {
    foo(new('myclass'), 'myclass')
}

For this to work foo has to be defined in the same environment as
test.myclass that is in the same test case code file. (Each test case
code file is sourced to a new environment and evaluated therein so foo
from a previous file is no longer known).

Regards,

  Matthias



Iago Mosqueira wrote:
> Hello,
> 
> I would like to re-use some of the test functions in an RUnit suite by 
> calling them with different inputs. Can this be done? What is the best 
> way of doing so?
> 
> If I define
> 
> test.foo <- function(object) {
>     checkTrue(class(object), 'myclass')
> }
> 
> How can this be called my runTestSuite for various object(s)?
> 
> Should I instead do
> 
> foo <- function(object) {
>     checkTrue(class(object), 'myclass')
> }
> 
> test.myclass <- function() {
>     foo(new('myclass'))
> }
> 
> and so on?
> 
> Many thanks,
> 
> 
> Iago
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Runit-help mailing list
> Run...@li...
> https://lists.sourceforge.net/lists/listinfo/runit-help
> 


-- 
Matthias Burger                     Project Manager/ Biostatistician
Epigenomics AG    Kleine Praesidentenstr. 1    10178 Berlin, Germany
phone:+49-30-24345-371                          fax:+49-30-24345-555
http://www.epigenomics.com           mat...@ep...
--
Epigenomics AG Berlin           Amtsgericht Charlottenburg HRB 75861
Vorstand:   Geert Nygaard (CEO/Vorsitzender),  Dr. Kurt Berlin (CSO)
              Oliver Schacht PhD (CFO),  Christian Piepenbrock (COO)
Aufsichtsrat:   Prof. Dr. Dr. hc. Rolf Krebs (Chairman/Vorsitzender)

Hello,

I would like to re-use some of the test functions in an RUnit suite by 
calling them with different inputs. Can this be done? What is the best 
way of doing so?

If I define

test.foo <- function(object) {
    checkTrue(class(object), 'myclass')
}

How can this be called my runTestSuite for various object(s)?

Should I instead do

foo <- function(object) {
    checkTrue(class(object), 'myclass')
}

test.myclass <- function() {
    foo(new('myclass'))
}

and so on?

Many thanks,


Iago

Hi,

VIAvvGRA $3. 35
VALvvIUM $1. 25
CIAvvLIS $3. 75
XAvvNAX
SOvvMA

FOR LESS!

http://vedrx.+com

Remove "+" in the above link



slightly. He set off up the gap between the Gryffindor and Hufflepuff
tables. It felt like an immensely long walk; the top table didnt seem
to be getting any nearer at all, and he could feel hundreds and hundreds

Hi,

Viuagra - 3.35
Vaulium - 1.25
Ciualis - 3.75
Amabien - 2.90
Souma   - 1.15

http://www.zodr!x.com

Important: Remove "!" in the above link

Hi,

Vinagra - 3. 35
Vanlium - 1. 25
Cinalis - 3. 75
Amnbien - 2. 90

http://lodrx!com

Important, Replace "!" with "."

--

blissfully blank. All that mattered in the world was that he kept
watching the veela, because if they stopped dancing, terrible things
would happen.

Hi,

Purchase your medication from our site because it costs less.

http://www.vodrx*.com - Remove "*" to make the link working!

--

Bit embarrassing,  Mr. Diggory said grimly, looking down at Winkys
unconscious form. Barty Crouchs house-elf... I mean to say... 
Come off it, Amos,  said Mr. Weasley quietly, you dont seriously

Hi,
Pharmacy directly from the manufacturer, economize over 50 %

http://www.lodrx*.com - Remove "*" to make the link working!

--

It was over. Harry thought dazedly, as Madam Pomfrey began herding the
champions and hostages back to the castle to get into dry clothes ...it
was over, he had got through ...he didnt have to worry about anything

Good day,

Viazzgra  $1, 80
Ciazzlis  $3, 00
Levizztra $3, 35

http://www.printeryml.*com ( Important ! Remove "*" )

--
Maybe Id better stay here. Headmaster... 
You will take Harry back to school, Hagrid,  Dumbledore repeated
firmly. Take him right up to Gryffindor Tower. And Harry I want you to

Hi,

VIxAGxRA $3. 35
CIxALxIS $3. 75
VAxLIxUM $1. 30
AMxBIxEN $2. 90
SOxMA    $1. 15

and many other

http://www.rx555*com ( Do not forget to replace "*" with "." )

--
perhaps, that made her give them so much homework at the end of the class.
A detailed analysis of the way the planetary movements in the coming
month will affect you, with reference to your personal chart,  she

Hi,

When you need it now VIAxxGRA ( from $3, 35 )
When you need it later CIAxxLIS ( from $3, 35 )
 
http://www.fod*rx.com
( remove "*" from the domain name )

----
was only half listening. So tired every bone in his body was aching, he
wanted nothing more than to sit here, undisturbed, for hours and hours,
until he fell asleep and didnt have to think or feel anymore.

AMB_BlEN $ 2, 90
VAL_LlUM $ 1, 25
VlA_AGRA $ 3, 30
ClA_ALlS $ 3, 75
XAN_NAX $ 1, 50
=20
http://www.vaserungenionyngandsa.com
=20
=20
=20
unconcernedly, and slid sideways through it... and as they did so,
platform nine and three-quarters materialized in front of them.
The Hogwarts Express, a gleaming scarlet steam engine, was already

Hi,

VrAGRA for less

http://www.berionkdegunahsein.com
=20
  _____ =20


What do we do? Dreadnought asked unhappily. His companion was just
clinked among the coins I stopped. Opened my attach) case and dragged
and waved a smoke-stained flamethrower in the direction of the door. I

Hi,

ClAvLIS

AMBvlEN

VALvlUM

VlAvGRA

Economize 60% http://www.berinkdasepodes.com
=20
  _____ =20


do await you!
swinging, right side up now, while the blood seeped out of my florid
box and started building a second tower from the blocks.

Hi,

ClArLIS

VALrlUM

VlArGRA

AMBrlEN

Economize 50% http://www.haishikionlderunhqer.com
=20
  _____ =20


awards to be made-then the crime had to occur. In that brief period we
pressure continuous. And I am surrounded by morons-stulteguloj,
fading instantly when I thought about my own problems.