R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Albrecht Kauffmann-2
Hi all,

after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
R-3.4.1 I never had a problem with these tests. No, the programm stops
at the following line of tests/reg-tests-1d.R:

> ## available.packages() (not) caching in case of errors
> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
    error in assertCondition(expr, "warning",
   .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
   available.packages(repo  ...: wanted warning

The error message is a result of:      ap2 <- available.packages(repos =
"http://foo.bar"),
not of the following condition in the test script:   stopifnot(nrow(ap1)
== 0, identical(ap1, ap2))  .

ap1 <- available.packages(repos = "http://foo.bar") works well, but the
following line stops the program.

After replacing "http://foo.bar" by "http://foo.poi" (every phrase
different from "bar" works well), reg-test-1d.R passed without error
message the interpreter.

Has someone an idea why the original program code stops (only) at line
698:
tools::assertWarning(ap2 <- available.packages(repos =
"http://foo.bar"))?

With many thanks for every hint,
Albrecht

--
  Albrecht Kauffmann
  [hidden email]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Duncan Murdoch-2
On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:

> Hi all,
>
> after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
> R-3.4.1 I never had a problem with these tests. No, the programm stops
> at the following line of tests/reg-tests-1d.R:
>
>> ## available.packages() (not) caching in case of errors
>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
>      error in assertCondition(expr, "warning",
>     .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
>     available.packages(repo  ...: wanted warning
>
> The error message is a result of:      ap2 <- available.packages(repos =
> "http://foo.bar"),
> not of the following condition in the test script:   stopifnot(nrow(ap1)
> == 0, identical(ap1, ap2))  .

What error message did you see?

>
> ap1 <- available.packages(repos = "http://foo.bar") works well, but the
> following line stops the program.
>
> After replacing "http://foo.bar" by "http://foo.poi" (every phrase
> different from "bar" works well), reg-test-1d.R passed without error
> message the interpreter.
>
> Has someone an idea why the original program code stops (only) at line
> 698:
> tools::assertWarning(ap2 <- available.packages(repos =
> "http://foo.bar"))?
>


Perhaps your system is resolving foo.bar to a web address.  If I use a
URL for a web page that doesn't have a repository, I still get a
warning, but the text of the error may explain why you don't.

Duncan Murdoch

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Albrecht Kauffmann-2
Dear Duncan,

Many thanks!

Am Mi, 1. Nov 2017, um 16:17, schrieb Duncan Murdoch:

> On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:
> > Hi all,
> >
> > after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
> > R-3.4.1 I never had a problem with these tests. No, the programm stops
> > at the following line of tests/reg-tests-1d.R:
> >
> >> ## available.packages() (not) caching in case of errors
> >> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> >> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
> >      error in assertCondition(expr, "warning",
> >     .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
> >     available.packages(repo  ...: wanted warning
> >
> > The error message is a result of:      ap2 <- available.packages(repos =
> > "http://foo.bar"),
> > not of the following condition in the test script:   stopifnot(nrow(ap1)
> > == 0, identical(ap1, ap2))  .
>
> What error message did you see?

after: R --vanilla <../R-3.4.2/tests/reg-tests-1d.R

I get

...
> ## available.packages() (not) caching in case of errors
> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
   Error in assertCondition(expr, "warning", .exprString = d.expr) :
     Got simpleError during evaluating of ap2 <- available.packages(repo
      ...: wanted warning
     Calls: <Anonymous> -> assertCondition
     Evaluation stopped


>
> >
> > ap1 <- available.packages(repos = "http://foo.bar") works well, but the
> > following line stops the program.
> >
> > After replacing "http://foo.bar" by "http://foo.poi" (every phrase
> > different from "bar" works well), reg-test-1d.R passed without error
> > message the interpreter.
> >
> > Has someone an idea why the original program code stops (only) at line
> > 698:
> > tools::assertWarning(ap2 <- available.packages(repos =
> > "http://foo.bar"))?
> >
>
>
> Perhaps your system is resolving foo.bar to a web address.  If I use a
> URL for a web page that doesn't have a repository, I still get a
> warning, but the text of the error may explain why you don't.

There was a message, maybe a warning, but I don't understand it. Now I
installed R-3.4.2 also on another PC; there it works fine. Do you have
an idea, what went wrong here?

Albrecht Kauffmann

>
> Duncan Murdoch

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Duncan Murdoch-2
On 01/11/2017 12:02 PM, Albrecht Kauffmann wrote:

> Dear Duncan,
>
> Many thanks!
>
> Am Mi, 1. Nov 2017, um 16:17, schrieb Duncan Murdoch:
>> On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:
>>> Hi all,
>>>
>>> after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
>>> R-3.4.1 I never had a problem with these tests. No, the programm stops
>>> at the following line of tests/reg-tests-1d.R:
>>>
>>>> ## available.packages() (not) caching in case of errors
>>>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
>>>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
>>>       error in assertCondition(expr, "warning",
>>>      .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
>>>      available.packages(repo  ...: wanted warning
>>>
>>> The error message is a result of:      ap2 <- available.packages(repos =
>>> "http://foo.bar"),
>>> not of the following condition in the test script:   stopifnot(nrow(ap1)
>>> == 0, identical(ap1, ap2))  .
>>
>> What error message did you see?
>
> after: R --vanilla <../R-3.4.2/tests/reg-tests-1d.R

Sorry, I wasn't clear.  I meant to ask what error message you get if you
run the command that's supposed to generate a warning, i.e. run the code

ap1 <- available.packages(repos = "http://foo.bar")
ap2 <- available.packages(repos = "http://foo.bar")

and you'll likely see a warning on the first line, and an error on the
second.  The text of that error may be informative.

Duncan Murdoch

>
> I get
>
> ...
>> ## available.packages() (not) caching in case of errors
>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
>     Error in assertCondition(expr, "warning", .exprString = d.expr) :
>       Got simpleError during evaluating of ap2 <- available.packages(repo
>        ...: wanted warning
>       Calls: <Anonymous> -> assertCondition
>       Evaluation stopped
>
>
>>
>>>
>>> ap1 <- available.packages(repos = "http://foo.bar") works well, but the
>>> following line stops the program.
>>>
>>> After replacing "http://foo.bar" by "http://foo.poi" (every phrase
>>> different from "bar" works well), reg-test-1d.R passed without error
>>> message the interpreter.
>>>
>>> Has someone an idea why the original program code stops (only) at line
>>> 698:
>>> tools::assertWarning(ap2 <- available.packages(repos =
>>> "http://foo.bar"))?
>>>
>>
>>
>> Perhaps your system is resolving foo.bar to a web address.  If I use a
>> URL for a web page that doesn't have a repository, I still get a
>> warning, but the text of the error may explain why you don't.
>
> There was a message, maybe a warning, but I don't understand it. Now I
> installed R-3.4.2 also on another PC; there it works fine. Do you have
> an idea, what went wrong here?
>
> Albrecht Kauffmann
>
>>
>> Duncan Murdoch

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Albrecht Kauffmann-2
Am Mi, 1. Nov 2017, um 17:40, schrieb Duncan Murdoch:

> On 01/11/2017 12:02 PM, Albrecht Kauffmann wrote:
> > Dear Duncan,
> >
> > Many thanks!
> >
> > Am Mi, 1. Nov 2017, um 16:17, schrieb Duncan Murdoch:
> >> On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:
> >>> Hi all,
> >>>
> >>> after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
> >>> R-3.4.1 I never had a problem with these tests. No, the programm stops
> >>> at the following line of tests/reg-tests-1d.R:
> >>>
> >>>> ## available.packages() (not) caching in case of errors
> >>>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> >>>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
> >>>       error in assertCondition(expr, "warning",
> >>>      .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
> >>>      available.packages(repo  ...: wanted warning
> >>>
> >>> The error message is a result of:      ap2 <- available.packages(repos =
> >>> "http://foo.bar"),
> >>> not of the following condition in the test script:   stopifnot(nrow(ap1)
> >>> == 0, identical(ap1, ap2))  .
> >>
> >> What error message did you see?
> >
> > after: R --vanilla <../R-3.4.2/tests/reg-tests-1d.R
>
> Sorry, I wasn't clear.  I meant to ask what error message you get if you
> run the command that's supposed to generate a warning, i.e. run the code
>
Oops! Here are the resulting error messages, which are not the same:
> ap1 <- available.packages(repos = "http://foo.bar")
Warning: Cannot access to the index of the repository
http://foo.bar/src/contrib:
  line beginning with '<!DOCTYPE html PUBLI ...' is incorrect formatted

> ap2 <- available.packages(repos = "http://foo.bar")
Error in readRDS(dest) : unknown input format

On my other PC, the same lines give 2 times the same error message:
"Warning: Cannot access to the index of the repository
http://foo.bar/src/contrib: Cannot open URL
'http://foo.bar/src/contrib/PACKAGES' "

Indeed, the second error message on the present PC is different. What
may be the cause?

Many thanks
Albrecht

>
> and you'll likely see a warning on the first line, and an error on the
> second.  The text of that error may be informative.
>
> Duncan Murdoch
>
> >
> > I get
> >
> > ...
> >> ## available.packages() (not) caching in case of errors
> >> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> >> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
> >     Error in assertCondition(expr, "warning", .exprString = d.expr) :
> >       Got simpleError during evaluating of ap2 <- available.packages(repo
> >        ...: wanted warning
> >       Calls: <Anonymous> -> assertCondition
> >       Evaluation stopped
> >
> >
> >>
> >>>
> >>> ap1 <- available.packages(repos = "http://foo.bar") works well, but the
> >>> following line stops the program.
> >>>
> >>> After replacing "http://foo.bar" by "http://foo.poi" (every phrase
> >>> different from "bar" works well), reg-test-1d.R passed without error
> >>> message the interpreter.
> >>>
> >>> Has someone an idea why the original program code stops (only) at line
> >>> 698:
> >>> tools::assertWarning(ap2 <- available.packages(repos =
> >>> "http://foo.bar"))?
> >>>
> >>
> >>
> >> Perhaps your system is resolving foo.bar to a web address.  If I use a
> >> URL for a web page that doesn't have a repository, I still get a
> >> warning, but the text of the error may explain why you don't.
> >
> > There was a message, maybe a warning, but I don't understand it. Now I
> > installed R-3.4.2 also on another PC; there it works fine. Do you have
> > an idea, what went wrong here?
> >
> > Albrecht Kauffmann
> >
> >>
> >> Duncan Murdoch
>

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Duncan Murdoch-2

On 01/11/2017 1:08 PM, Albrecht Kauffmann wrote:

> Am Mi, 1. Nov 2017, um 17:40, schrieb Duncan Murdoch:
>> On 01/11/2017 12:02 PM, Albrecht Kauffmann wrote:
>>> Dear Duncan,
>>>
>>> Many thanks!
>>>
>>> Am Mi, 1. Nov 2017, um 16:17, schrieb Duncan Murdoch:
>>>> On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:
>>>>> Hi all,
>>>>>
>>>>> after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
>>>>> R-3.4.1 I never had a problem with these tests. No, the programm stops
>>>>> at the following line of tests/reg-tests-1d.R:
>>>>>
>>>>>> ## available.packages() (not) caching in case of errors
>>>>>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
>>>>>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
>>>>>        error in assertCondition(expr, "warning",
>>>>>       .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
>>>>>       available.packages(repo  ...: wanted warning
>>>>>
>>>>> The error message is a result of:      ap2 <- available.packages(repos =
>>>>> "http://foo.bar"),
>>>>> not of the following condition in the test script:   stopifnot(nrow(ap1)
>>>>> == 0, identical(ap1, ap2))  .
>>>>
>>>> What error message did you see?
>>>
>>> after: R --vanilla <../R-3.4.2/tests/reg-tests-1d.R
>>
>> Sorry, I wasn't clear.  I meant to ask what error message you get if you
>> run the command that's supposed to generate a warning, i.e. run the code
>>
> Oops! Here are the resulting error messages, which are not the same:
>> ap1 <- available.packages(repos = "http://foo.bar")
> Warning: Cannot access to the index of the repository
> http://foo.bar/src/contrib:
>    line beginning with '<!DOCTYPE html PUBLI ...' is incorrect formatted
>
>> ap2 <- available.packages(repos = "http://foo.bar")
> Error in readRDS(dest) : unknown input format
>
> On my other PC, the same lines give 2 times the same error message:
> "Warning: Cannot access to the index of the repository
> http://foo.bar/src/contrib: Cannot open URL
> 'http://foo.bar/src/contrib/PACKAGES' "
>
> Indeed, the second error message on the present PC is different. What
> may be the cause?

It looks as though one of your PCs is returning a file in response to
the request, rather than signalling that the URL is not found.

If you go into a browser like Firefox and try to open the URL

http://foo.bar/src/contrib/PACKAGES

what do you see on the two systems?  I get a message from Firefox that
the server is not found.

Some ISPs respond to requests for nonexistent URLs with a message,
basically an ad for something or other.  Maybe that's what you're getting.

In any case, this looks like a bug in R:  the first request failed, but
something got cached, and then the second request got an error reading
the cached value.

Duncan Murdoch

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Albrecht Kauffmann-2
Dear Duncan,

thank you very much, that is it! Indeed, there was activated in my
internet router the "telekom navigation help". When I am (or the system
is) looking for an not existent url, this "help" opens its own website.
After deactivation of telekom navigation help, the normal message
(server not found) comes. I never gave attention to this "help",
although it is bothering sometimes; now it is switched off in my
residence. However, in Germany this problem probably frequently arises,
therefor, it should be considered in the respective code for testing R
after compilation.  

All the best,
Albrecht
--
  Albrecht Kauffmann
  [hidden email]

Am Mi, 1. Nov 2017, um 19:24, schrieb Duncan Murdoch:

>
> On 01/11/2017 1:08 PM, Albrecht Kauffmann wrote:
> > Am Mi, 1. Nov 2017, um 17:40, schrieb Duncan Murdoch:
> >> On 01/11/2017 12:02 PM, Albrecht Kauffmann wrote:
> >>> Dear Duncan,
> >>>
> >>> Many thanks!
> >>>
> >>> Am Mi, 1. Nov 2017, um 16:17, schrieb Duncan Murdoch:
> >>>> On 01/11/2017 8:11 AM, Albrecht Kauffmann wrote:
> >>>>> Hi all,
> >>>>>
> >>>>> after compiling R-3.4.2 on opensuse leap 42.3, make check failed. Until
> >>>>> R-3.4.1 I never had a problem with these tests. No, the programm stops
> >>>>> at the following line of tests/reg-tests-1d.R:
> >>>>>
> >>>>>> ## available.packages() (not) caching in case of errors
> >>>>>> tools::assertWarning(ap1 <- available.packages(repos = "http://foo.bar"))
> >>>>>> tools::assertWarning(ap2 <- available.packages(repos = "http://foo.bar"))
> >>>>>        error in assertCondition(expr, "warning",
> >>>>>       .exprString = d.expr) :  Got simpleError evaluating of ap2 <-
> >>>>>       available.packages(repo  ...: wanted warning
> >>>>>
> >>>>> The error message is a result of:      ap2 <- available.packages(repos =
> >>>>> "http://foo.bar"),
> >>>>> not of the following condition in the test script:   stopifnot(nrow(ap1)
> >>>>> == 0, identical(ap1, ap2))  .
> >>>>
> >>>> What error message did you see?
> >>>
> >>> after: R --vanilla <../R-3.4.2/tests/reg-tests-1d.R
> >>
> >> Sorry, I wasn't clear.  I meant to ask what error message you get if you
> >> run the command that's supposed to generate a warning, i.e. run the code
> >>
> > Oops! Here are the resulting error messages, which are not the same:
> >> ap1 <- available.packages(repos = "http://foo.bar")
> > Warning: Cannot access to the index of the repository
> > http://foo.bar/src/contrib:
> >    line beginning with '<!DOCTYPE html PUBLI ...' is incorrect formatted
> >
> >> ap2 <- available.packages(repos = "http://foo.bar")
> > Error in readRDS(dest) : unknown input format
> >
> > On my other PC, the same lines give 2 times the same error message:
> > "Warning: Cannot access to the index of the repository
> > http://foo.bar/src/contrib: Cannot open URL
> > 'http://foo.bar/src/contrib/PACKAGES' "
> >
> > Indeed, the second error message on the present PC is different. What
> > may be the cause?
>
> It looks as though one of your PCs is returning a file in response to
> the request, rather than signalling that the URL is not found.
>
> If you go into a browser like Firefox and try to open the URL
>
> http://foo.bar/src/contrib/PACKAGES
>
> what do you see on the two systems?  I get a message from Firefox that
> the server is not found.
>
> Some ISPs respond to requests for nonexistent URLs with a message,
> basically an ad for something or other.  Maybe that's what you're
> getting.
>
> In any case, this looks like a bug in R:  the first request failed, but
> something got cached, and then the second request got an error reading
> the cached value.
>
> Duncan Murdoch

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: R-3.4.2: make check stops at line 698 of tests/reg-tests-1d.R

Jeff Newmiller
This idea of hijacking failed connections has been tried before and it breaks many things beyond R software and most ISPs have given up on doing this. If you, who have this "feature" at hand can figure out a reliable way to detect this then it might be "considered". However, it seems unlikely that a general solution can be found after all this time, so the better option is probably for you to disable this "feature".
--
Sent from my phone. Please excuse my brevity.

On November 1, 2017 1:52:16 PM PDT, Albrecht Kauffmann <[hidden email]> wrote:

>Dear Duncan,
>
>thank you very much, that is it! Indeed, there was activated in my
>internet router the "telekom navigation help". When I am (or the system
>is) looking for an not existent url, this "help" opens its own website.
>After deactivation of telekom navigation help, the normal message
>(server not found) comes. I never gave attention to this "help",
>although it is bothering sometimes; now it is switched off in my
>residence. However, in Germany this problem probably frequently arises,
>therefor, it should be considered in the respective code for testing R
>after compilation.  
>
>All the best,
>Albrecht

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.