Re: Trouble installing rk.anova again...

[meik michalke]

Am Sonntag, 30. Juli 2017, 09:37:09 CEST schrieb Aaron Batty:
> I've scheduled a straight hour for software installation on Tuesday, because 
> I'm anticipating a bloodbath.

made my day :-D :: m.

-- 
  dipl. psych. meik michalke
  institut f"ur experimentelle psychologie
  abt. f"ur diagnostik und differentielle psychologie
  heinrich-heine-universit"at d-40204 d"usseldorf

signature.asc
Description: This is a digitally signed message part.

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

Strike that last message. It was just a side effect of having one million
different versions of RKWard on this machine. I blew everything out and
started again, and it was fine.

So the only problem appears to be the relative path on the shortcut, which,
as I said before, is easy to fix, so if you can't or don't want to fiddle
with it, no worries.

Sorry for the annual panic!


Aaron

On Sun, Jul 30, 2017 at 9:46 AM, Aaron Batty  wrote:

> Well, crap.
>
> I was about to just call this done for now, but I tried running an ANOVA
> with the new 0.65, just to be sure, and it has trouble installing ez as
> well, but this one is new.
>
> It said it couldn't move the installation of plyr from its temporary
> location, and I get this in the output:
>
> Warning in library(package, lib.loc = lib.loc, character.only = TRUE, 
> logical.return = TRUE,  :
>   there is no package called 'ez'
> Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = 
> vI[[j]]) :
>   there is no package called 'plyr'
> Error in eval(expr, envir, enclos) : could not find function "ezANOVA"
>
> So something is still off...
>
>
> Aaron
>
>
>
> On Sun, Jul 30, 2017 at 9:37 AM, Aaron Batty 
> wrote:
>
>> Thomas:
>>
>> K, that new one works fine with the .ini change, but the shortcut
>> creation still isn't working (on Win7). Weird. The old 0.65 release worked
>> fine...
>>
>> If you're out of time for this little project supporting an old release
>> that you're trying to kick to the curb anyway, then we can just go ahead
>> with this, and I'll have the students make their own shortcuts. I've
>> scheduled a straight hour for software installation on Tuesday, because I'm
>> anticipating a bloodbath.
>>
>> I realized last night as I was falling asleep that I really do need it to
>> run in its own little directory because at least one guy is going to be
>> using a university laptop and he might need to run this off of a USB stick.
>>
>> ...Because he only has an iPad.
>>
>> As a graduate student in linguistics.
>>
>> —A field in which 86% of published literature is quantitative.
>>
>> So I'm a bit nervous about this group...
>>
>>
>> Aaron
>>
>> On Sun, Jul 30, 2017 at 3:51 AM, Thomas Friedrichsmeier <
>> thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:
>>
>>> On 2017-07-29 14:22, Aaron Batty wrote:
>>>
 PS: The new 0.65b pulls packages down just fine.

 On Sat, Jul 29, 2017 at 9:17 PM, Aaron Batty 
 wrote:

 Thomas:
>
> Thanks. That indeed did do it.
>

>>> Ok, good.
>>>
>>> I think the shortcut problem is similar. Here's its target:
>
> C:\RKWard\RKWard\KDE\bin\rkward.exe
>
> Starting in:
>
> C:\RKWard\RKWard\KDE\bin
>
> --Whereas the old 0.65 shortcut is aware of where it is.
>

>>> Yeah, the root problem is that Window does not support the exotic
>>> concept of relative paths in shortcuts. Instead it appears to stack a bunch
>>> of hacks on shortcuts to make sure they will work even after moving them -
>>> if all goes well. To make matters as complicated as possible, these hacks
>>> differ between different versions of Windows (
>>> https://en.wikipedia.org/wiki/Shortcut_(computing)#Microsoft_Windows).
>>>
>>> I'm not sure. The problem _may_ have been that I had failed to update
>>> the shortcut after updating rkward.exe. An updated bundle will be available
>>> for download in an hour or so (slooow connection). I _hope_ that will fix
>>> both problems.
>>>
>>> Regards
>>> Thomas
>>>
>>
>>
>

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

K, that new one works fine with the .ini change, but the shortcut creation
still isn't working (on Win7). Weird. The old 0.65 release worked fine...

If you're out of time for this little project supporting an old release
that you're trying to kick to the curb anyway, then we can just go ahead
with this, and I'll have the students make their own shortcuts. I've
scheduled a straight hour for software installation on Tuesday, because I'm
anticipating a bloodbath.

I realized last night as I was falling asleep that I really do need it to
run in its own little directory because at least one guy is going to be
using a university laptop and he might need to run this off of a USB stick.

...Because he only has an iPad.

As a graduate student in linguistics.

—A field in which 86% of published literature is quantitative.

So I'm a bit nervous about this group...


Aaron

On Sun, Jul 30, 2017 at 3:51 AM, Thomas Friedrichsmeier <
thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:

> On 2017-07-29 14:22, Aaron Batty wrote:
>
>> PS: The new 0.65b pulls packages down just fine.
>>
>> On Sat, Jul 29, 2017 at 9:17 PM, Aaron Batty 
>> wrote:
>>
>> Thomas:
>>>
>>> Thanks. That indeed did do it.
>>>
>>
> Ok, good.
>
> I think the shortcut problem is similar. Here's its target:
>>>
>>> C:\RKWard\RKWard\KDE\bin\rkward.exe
>>>
>>> Starting in:
>>>
>>> C:\RKWard\RKWard\KDE\bin
>>>
>>> --Whereas the old 0.65 shortcut is aware of where it is.
>>>
>>
> Yeah, the root problem is that Window does not support the exotic concept
> of relative paths in shortcuts. Instead it appears to stack a bunch of
> hacks on shortcuts to make sure they will work even after moving them - if
> all goes well. To make matters as complicated as possible, these hacks
> differ between different versions of Windows (
> https://en.wikipedia.org/wiki/Shortcut_(computing)#Microsoft_Windows).
>
> I'm not sure. The problem _may_ have been that I had failed to update the
> shortcut after updating rkward.exe. An updated bundle will be available for
> download in an hour or so (slooow connection). I _hope_ that will fix both
> problems.
>
> Regards
> Thomas
>

Re: Trouble installing rk.anova again...

[Thomas Friedrichsmeier]

On 2017-07-29 14:22, Aaron Batty wrote:

PS: The new 0.65b pulls packages down just fine.

On Sat, Jul 29, 2017 at 9:17 PM, Aaron Batty 
wrote:


Thomas:

Thanks. That indeed did do it.


Ok, good.


I think the shortcut problem is similar. Here's its target:

C:\RKWard\RKWard\KDE\bin\rkward.exe

Starting in:

C:\RKWard\RKWard\KDE\bin

--Whereas the old 0.65 shortcut is aware of where it is.


Yeah, the root problem is that Window does not support the exotic 
concept of relative paths in shortcuts. Instead it appears to stack a 
bunch of hacks on shortcuts to make sure they will work even after 
moving them - if all goes well. To make matters as complicated as 
possible, these hacks differ between different versions of Windows 
(https://en.wikipedia.org/wiki/Shortcut_(computing)#Microsoft_Windows).


I'm not sure. The problem _may_ have been that I had failed to update 
the shortcut after updating rkward.exe. An updated bundle will be 
available for download in an hour or so (slooow connection). I _hope_ 
that will fix both problems.


Regards
Thomas

Re: Trouble installing rk.anova again...

[Aaron Batty]

PS: The new 0.65b pulls packages down just fine.

On Sat, Jul 29, 2017 at 9:17 PM, Aaron Batty  wrote:

> Thomas:
>
> Thanks. That indeed did do it.
>
> I think the shortcut problem is similar. Here's its target:
>
> C:\RKWard\RKWard\KDE\bin\rkward.exe
>
> Starting in:
>
> C:\RKWard\RKWard\KDE\bin
>
> --Whereas the old 0.65 shortcut is aware of where it is.
>
> Yoroshiku!
>
>
> Aaron
>
> On Sat, Jul 29, 2017 at 9:06 PM, Thomas Friedrichsmeier <
> thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:
>
>> Hi,
>>
>> On 2017-07-29 13:46, Aaron Batty wrote:
>>
>>> Thomas:
>>>
>>> Sorry to bother, but there's something wrong with that 0.65b you just
>>> made. I've downloaded it twice, but there seem to be some problems
>>> with the paths in some of the files.
>>>
>>> The shortcut to rkward.exe in the root level of the directory can't
>>> find the rkward.exe, even thought the path it displays is correct. If
>>> you try to just start the real .exe, it throws up a command line
>>> window and eventually has some other path error. Could you take a
>>> look?
>>>
>>
>> hm, I'm unsure about the problem with the shortcut to rkward.exe not
>> working. I know that problem has come up, before, but I've never quite
>> understood it. The other problem is quite simply that I forgot the step to
>> make the installation moveable. Try editing RKWard/KDE/bin/rkward.ini. It's
>> contents should be:
>>
>> R executable=..\\..\\R\\bin\\i386\\R.exe
>>
>> Does that do the trick, or any other problems remaining?
>>
>> I'll try to look into the link issue tonight, and upload a fixed bundle,
>> then.
>>
>> Regards
>> Thomas
>>
>
>

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

Thanks. That indeed did do it.

I think the shortcut problem is similar. Here's its target:

C:\RKWard\RKWard\KDE\bin\rkward.exe

Starting in:

C:\RKWard\RKWard\KDE\bin

--Whereas the old 0.65 shortcut is aware of where it is.

Yoroshiku!


Aaron

On Sat, Jul 29, 2017 at 9:06 PM, Thomas Friedrichsmeier <
thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:

> Hi,
>
> On 2017-07-29 13:46, Aaron Batty wrote:
>
>> Thomas:
>>
>> Sorry to bother, but there's something wrong with that 0.65b you just
>> made. I've downloaded it twice, but there seem to be some problems
>> with the paths in some of the files.
>>
>> The shortcut to rkward.exe in the root level of the directory can't
>> find the rkward.exe, even thought the path it displays is correct. If
>> you try to just start the real .exe, it throws up a command line
>> window and eventually has some other path error. Could you take a
>> look?
>>
>
> hm, I'm unsure about the problem with the shortcut to rkward.exe not
> working. I know that problem has come up, before, but I've never quite
> understood it. The other problem is quite simply that I forgot the step to
> make the installation moveable. Try editing RKWard/KDE/bin/rkward.ini. It's
> contents should be:
>
> R executable=..\\..\\R\\bin\\i386\\R.exe
>
> Does that do the trick, or any other problems remaining?
>
> I'll try to look into the link issue tonight, and upload a fixed bundle,
> then.
>
> Regards
> Thomas
>

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

Sorry to bother, but there's something wrong with that 0.65b you just made.
I've downloaded it twice, but there seem to be some problems with the paths
in some of the files.

The shortcut to rkward.exe in the root level of the directory can't find
the rkward.exe, even thought the path it displays is correct. If you try to
just start the real .exe, it throws up a command line window and eventually
has some other path error. Could you take a look?


Aaron

On Sat, Jul 29, 2017 at 8:17 PM, Aaron Batty  wrote:

> Thomas:
>
> Thanks! I'm downloading the "new" 0.65b now.
>
> Just finished testing out the 0.70 on Windows. It works great, and
> installs to \Program Files\ like people are expecting. I always had some
> people flummoxed by running an application from a folder...
>
> If the 0.65b works fine, though, I'll stick with that. The new GUI is much
> improved, to be sure, but I'm getting the feeling I'm not going to be
> working with the most technologically-savvy people next week, and I fear
> that a very slightly updated UI will be enough to send them over the edge.
>
> But it's wonderful to have options. You guys are just too good to me.
>
>
> Aaron
>
> On Sat, Jul 29, 2017 at 6:27 PM, Thomas Friedrichsmeier <
> thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:
>
>> Hi,
>>
>> On 2017-07-29 10:46, Aaron Batty wrote:
>>
>>> K, so I set the Windows version to build packages from source, removed
>>> ez (which was indeed installed), and reinstalled it. I got this error:
>>>
>>
>> you won't be able to compile most packages from source on Windows out of
>> the box.
>>
>> Anyway, a new bundle with RKWard 0.6.5 and R 3.3.3 will appear at
>> https://files.kde.org/rkward/testing/bundles/Windows/ in a few minutes.
>> Try that.
>>
>> Interestingly, this one will download packages from http://files.kde.org
>> just fine. Not sure whether that's due to R 3.3.3, Windows, or fixes on
>> files.kde.org.
>>
>> Regards
>> Thomas
>>
>
>

Re: Trouble installing rk.anova again...

[Thomas Friedrichsmeier]

Hi,

On 2017-07-29 10:46, Aaron Batty wrote:

K, so I set the Windows version to build packages from source, removed
ez (which was indeed installed), and reinstalled it. I got this error:


you won't be able to compile most packages from source on Windows out of 
the box.


Anyway, a new bundle with RKWard 0.6.5 and R 3.3.3 will appear at 
https://files.kde.org/rkward/testing/bundles/Windows/ in a few minutes. 
Try that.


Interestingly, this one will download packages from http://files.kde.org 
just fine. Not sure whether that's due to R 3.3.3, Windows, or fixes on 
files.kde.org.


Regards
Thomas

Re: Trouble installing rk.anova again...

[Aaron Batty]

Meik:

The error came up with 0.6.5 for Windows, which uses R 3.2.3.

I think that it installed the 3.2.5 version of ez (because it seemed to
work the first time through), but something about the version differences
caused the problem, and now it says it can't find it at all, which is weird.

K, so I set the Windows version to build packages from source, removed ez
(which was indeed installed), and reinstalled it. I got this error:

Warning messages:

1: running command '"C:/Users/abatty/Desktop/RKWard/R/bin/i386/R" CMD
INSTALL -l "C:\Users\abatty\Desktop\RKWard\R\library"
C:\Users\abatty\AppData\Local\Temp\RtmpiO8s0l/downloaded_packages/Matrix_1.2-10.tar.gz'
had status 1

2: In install.packages(c("ez"), lib =
"C:/Users/abatty/Desktop/RKWard/R/library") :

installation of package 'Matrix' had non-zero exit status

3: running command '"C:/Users/abatty/Desktop/RKWard/R/bin/i386/R" CMD
INSTALL -l "C:\Users\abatty\Desktop\RKWard\R\library"
C:\Users\abatty\AppData\Local\Temp\RtmpiO8s0l/downloaded_packages/ez_4.4-0.tar.gz'
had status 1

4: In install.packages(c("ez"), lib =
"C:/Users/abatty/Desktop/RKWard/R/library") :

installation of package 'ez' had non-zero exit status

Hmm.


Aaron


On Sat, Jul 29, 2017 at 4:53 PM, meik michalke <
meik.micha...@uni-duesseldorf.de> wrote:

> hi aaron,
>
> Am Samstag, 29. Juli 2017, 16:26:04 CEST schrieb Aaron Batty:
> > Warning: package 'ez' was built under R version 3.2.5
>
> FWIW, i took your mail as a reason to rebuild all plugins using R 3.4, so
> there are now also 3.4 packages available for windows and macOS as well. [i
> also pushed rkwarddev 0.08-2, but you wouldn't need that]
>
> the thing is that windows and macOS install precompiled binary packages by
> default. CRAN seems to still host 3.3 packages:
>  https://cran.r-project.org/web/packages/ez/index.html
>  (see the r-oldrel links)
> but no longer 3.2. the error message above however suggests that you're not
> running R 3.2, but are using an older installation of ez, right? if you're
> on
> R 3.3, it might help to just reinstall ez. although, like thomas said,
> RKWard
> was build for R 3.2 in the windows installer.
>
> i wonder if it helps to try and install source packages instead, like on
> linux. that would build the package locally for the installed R version. i
> have no idea whether this needs additional software setup on windows to
> work,
> but you could simply try it -- check "install packages from source code" in
> the R packages section of RKWard's configuration; simply undo it if it
> doesn't
> help.
>
>
> viele grüße :: m.eik
>
> --
>   dipl. psych. meik michalke
>   institut f"ur experimentelle psychologie
>   abt. f"ur diagnostik und differentielle psychologie
>   heinrich-heine-universit"at d-40204 d"usseldorf

Re: Trouble installing rk.anova again...

[meik michalke]

hi aaron,

Am Samstag, 29. Juli 2017, 16:26:04 CEST schrieb Aaron Batty:
> Warning: package 'ez' was built under R version 3.2.5

FWIW, i took your mail as a reason to rebuild all plugins using R 3.4, so 
there are now also 3.4 packages available for windows and macOS as well. [i 
also pushed rkwarddev 0.08-2, but you wouldn't need that]

the thing is that windows and macOS install precompiled binary packages by 
default. CRAN seems to still host 3.3 packages:
 https://cran.r-project.org/web/packages/ez/index.html
 (see the r-oldrel links)
but no longer 3.2. the error message above however suggests that you're not 
running R 3.2, but are using an older installation of ez, right? if you're on 
R 3.3, it might help to just reinstall ez. although, like thomas said, RKWard 
was build for R 3.2 in the windows installer.

i wonder if it helps to try and install source packages instead, like on 
linux. that would build the package locally for the installed R version. i 
have no idea whether this needs additional software setup on windows to work, 
but you could simply try it -- check "install packages from source code" in 
the R packages section of RKWard's configuration; simply undo it if it doesn't 
help.


viele grüße :: m.eik

-- 
  dipl. psych. meik michalke
  institut f"ur experimentelle psychologie
  abt. f"ur diagnostik und differentielle psychologie
  heinrich-heine-universit"at d-40204 d"usseldorf

signature.asc
Description: This is a digitally signed message part.

Re: Trouble installing rk.anova again...

[Thomas Friedrichsmeier]

On 2017-07-29 09:26, Aaron Batty wrote:

Ugh, sorry for the delay. Nothing was set up on this thing because I
usually boot Ubuntu on it. There is nothing as thoroughly obnoxious as
an enterprise-managed Windows machine.

Anyway, it looks like rk.anova is installed by default, but the
problem comes up when trying to install ez:

Warning: package 'ez' was built under R version 3.2.5
Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()),
versionCheck = vI[[j]]) :
  there is no package called 'Matrix'
Error in eval(expr, envir, enclos) : could not find function "ezANOVA"
When I tried to just let it run, it didn't say it couldn't find ez,
but when I try running it now, it says it can't. Also, there was some
error about not being able to uninstall Matrix first...

I put in the https repository, but that didn't change anything (of
course--that's just for the RK packages, right?).


Yes. Doesn't help the issue that CRAN no longer has packages for 3.2.x


What can we do? I'm expecting mostly Windows users among these
students...


Like I said, the options are
a) Use the 0.7.0 preview on Windows. This "should" work, but you'll have 
some minor UI differences between that and 0.6.5b on Mac.

b) Roll your own bundle with R 3.3.3 as outlined in my previous mail.
c) Wait for me to do this. In fact I'm just getting started to, but I 
cannot promise when I will get around to finish this.


[...]


So sorry to be down to the wire on this. I have been watching the
dev list and site and since there hasn't been a new version since
last fall, I didn't think I needed to check everything again. I'm
lucky that I had to make a new final for this course and wanted to
generate the model answers, which led me to the problem...


Yeah, totally annoying when things just break behind (y)our back(s). 
Don't worry, we'll find a workable solution before Monday.


(And staying on top will become considerably easier, once we finally 
have a KF5-based 0.7.0, and can work with a supported toolchain, again; 
we're mostly just missing some build scripts on Mac, and some finishing 
touches, there).


Regards
Thomas

Re: Trouble installing rk.anova again...

[Aaron Batty]

Ugh, sorry for the delay. Nothing was set up on this thing because I
usually boot Ubuntu on it. There is nothing as thoroughly obnoxious as an
enterprise-managed Windows machine.

Anyway, it looks like rk.anova is installed by default, but the problem
comes up when trying to install ez:

Warning: package 'ez' was built under R version 3.2.5
Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()),
versionCheck = vI[[j]]) :
  there is no package called 'Matrix'
Error in eval(expr, envir, enclos) : could not find function "ezANOVA"

When I tried to just let it run, it didn't say it couldn't find ez, but
when I try running it now, it says it can't. Also, there was some error
about not being able to uninstall Matrix first...

I put in the https repository, but that didn't change anything (of
course--that's just for the RK packages, right?).

What can we do? I'm expecting mostly Windows users among these students...


Aaron




On Sat, Jul 29, 2017 at 9:03 AM, Aaron Batty  wrote:

> Thomas:
>
> Ugh. I think that this time the students will mostly be on Windows... It's
> a... cheaper school. And they aren't web development majors, so they aren't
> swayed by UNIXiness...
>
> My course starts on Tuesday, 8/1, but that's Japan time, so Monday, 7/31
> in Europe.
>
> I have to go in to work today and get my Windows laptop anyway, so I'll
> try the install right away and let you know.
>
> So sorry to be down to the wire on this. I have been watching the dev list
> and site and since there hasn't been a new version since last fall, I
> didn't think I needed to check everything again. I'm lucky that I had to
> make a new final for this course and wanted to generate the model answers,
> which led me to the problem...
>
> Anyway, I'll reply again in a few hours with a report from the Windows
> version...
>
>
> Aaron
>
> On Sat, Jul 29, 2017 at 5:41 AM, Thomas Friedrichsmeier <
> thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:
>
>> Hi Aaron,
>>
>> On 2017-07-28 04:29, Aaron Batty wrote:
>>
>>> Thomas:
>>>
>>> Thanks! The https URL works. I'll just have the students do that on
>>> the first day and all will be fine.
>>>
>>
>> an ugly afterthought: You do have some Windows users among your students,
>> right? You may run into trouble, here, as the Windows bundle still has R
>> 3.2.3, and AFAIK, CRAN only has packages for R3.3+, today. (Stefan has been
>> bugging me about providing an update, but...)
>>
>> When does your course start? If it's really soon, you'll probably have to
>> go with the frameworks preview for Windows (see
>> https://rkward.kde.org/Windows). It's reported to work rather well, but
>> of course there will be _some_ differences. If that is not too late, I
>> _might_ get around to create an updated bundle for 0.6.5b, tomorrow. Or you
>> could roll one yourself, by unpacking the bundle, updating the R inside of
>> it (better to 3.3, only), and then running RKWard's "custom installer" on
>> top of that.
>>
>> Regards
>> Thomas
>>
>
>

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

Ugh. I think that this time the students will mostly be on Windows... It's
a... cheaper school. And they aren't web development majors, so they aren't
swayed by UNIXiness...

My course starts on Tuesday, 8/1, but that's Japan time, so Monday, 7/31 in
Europe.

I have to go in to work today and get my Windows laptop anyway, so I'll try
the install right away and let you know.

So sorry to be down to the wire on this. I have been watching the dev list
and site and since there hasn't been a new version since last fall, I
didn't think I needed to check everything again. I'm lucky that I had to
make a new final for this course and wanted to generate the model answers,
which led me to the problem...

Anyway, I'll reply again in a few hours with a report from the Windows
version...


Aaron

On Sat, Jul 29, 2017 at 5:41 AM, Thomas Friedrichsmeier <
thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:

> Hi Aaron,
>
> On 2017-07-28 04:29, Aaron Batty wrote:
>
>> Thomas:
>>
>> Thanks! The https URL works. I'll just have the students do that on
>> the first day and all will be fine.
>>
>
> an ugly afterthought: You do have some Windows users among your students,
> right? You may run into trouble, here, as the Windows bundle still has R
> 3.2.3, and AFAIK, CRAN only has packages for R3.3+, today. (Stefan has been
> bugging me about providing an update, but...)
>
> When does your course start? If it's really soon, you'll probably have to
> go with the frameworks preview for Windows (see
> https://rkward.kde.org/Windows). It's reported to work rather well, but
> of course there will be _some_ differences. If that is not too late, I
> _might_ get around to create an updated bundle for 0.6.5b, tomorrow. Or you
> could roll one yourself, by unpacking the bundle, updating the R inside of
> it (better to 3.3, only), and then running RKWard's "custom installer" on
> top of that.
>
> Regards
> Thomas
>

Re: Trouble installing rk.anova again...

[Aaron Batty]

Thomas:

Thanks! The https URL works. I'll just have the students do that on the
first day and all will be fine.


Aaron

On Fri, Jul 28, 2017 at 4:20 AM, Thomas Friedrichsmeier <
thomas.friedrichsme...@ruhr-uni-bochum.de> wrote:

> Hi Aaron,
>
> On Thu, 27 Jul 2017 21:17:25 +0900
> Aaron Batty  wrote:
>
>> Warning: unable to access index for repository
>> http://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/3.3:
>>   cannot open URL '
>> http://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/3.3/PACKAGES'
>>
>
> brr. Something broken with the mirror, again... Here, I can work around
> this by changing the repository from http://files.kde.org/rkward/R/ to
> https://files.kde.org/rkward/R/ (Settings->Configure RKWard->R
> packages).
>
> Your second best option (good enough to avoid the panic, at least)
> should be to download
> https://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/
> 3.3/rk.ANOVA_0.01-22.tgz
> in your browser (which will follow the redirect just fine), and install
> from the command line via R CMD INSTALL.
>
> Regards
> Thomas
>

Re: Trouble installing rk.anova again...

[Thomas Friedrichsmeier]

Hi Aaron,

On Thu, 27 Jul 2017 21:17:25 +0900
Aaron Batty  wrote:

Warning: unable to access index for repository
http://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/3.3:
  cannot open URL '
http://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/3.3/PACKAGES'


brr. Something broken with the mirror, again... Here, I can work around
this by changing the repository from http://files.kde.org/rkward/R/ to
https://files.kde.org/rkward/R/ (Settings->Configure RKWard->R
packages).

Your second best option (good enough to avoid the panic, at least)
should be to download
https://files.kde.org/rkward/R/bin/macosx/mavericks/contrib/3.3/rk.ANOVA_0.01-22.tgz
in your browser (which will follow the redirect just fine), and install
from the command line via R CMD INSTALL.

Regards
Thomas