Re: [Plplot-devel] Space in full pathname issues

[Arjen M. <Arj...@de...>]

Hi Alan,



I will be using this message to answer the other two as well ;).



See below.



Regards,



Arjen



> -----Original Message-----
> From: Alan W. Irwin [mailto:ir...@be...]
> Sent: Monday, May 08, 2017 7:18 PM
> To: Arjen Markus
> Cc: Chris Marshall; PLplot development list
> Subject: RE: [Plplot-devel] Space in full pathname issues
>
> On 2017-05-08 08:03-0000 Arjen Markus wrote:
>
> > Hi Alan, Chris,
> >
> >
> >
> > I tried the same thing with MinGW-w64/MSYS2. This succeeded in so far, that:
> >
> > -        CMake produced a full report and makefiles
>
> Assuming the form I suggested on Cygwin
>
> cmake  .... "-CMAKE_INSTALL_PREFIX=d:/plplot-src/install with spaces"
>
> does work there, please check that same form (used by the comprehensive test
> script) works on this platform as well.
>
I have thought of that form too - it might do the trick indeed. I just hadn't expected the phenomenon I got - CMake in an infinite loop.

> >
> > -        make succeeded
> >
> > -        make install succeeded as well
>
> Congratulations.  This is all very encouraging.
>
> > -        making the examples worked, except for Java - the path to the Java
> compiler was not quoted
>
> Interesting.  According to <http://repo.msys2.org/mingw/x86_64/>,
> MinGW-w64/MSYS2 does not provide Java.  The conclusion that this is a non-
> platform version of Java also seems to be confirmed by the space in the Java
> compiler pathname since I assume this particular test was for an ordinary "non-
> space" installation of this platform.  If it is true this is a non-platform Java, I am
> pretty sure you are going to have ABI issues when attempting to use it on MinGW-
> w64/MSYS2 so you should take some measure to exclude Java there.  Brute force
> would be -DENABLE_java=OFF (with a big note in your submission script), but
> perhaps a better overall choice would be to adjust your PATH on this platform so
> CMake does not find non-platform software like this.
>
> That said, note if this version of Java is ABI compatible with the MSVC compilers,
> then it should be expected to work for your MSVC tests, and the case (4) issue you
> have found with our build system (Java compiler not quoted) will need to be
> addressed, and I look forward to your commit to that effect.

Sure, I think CMake found the "bare Windows" version of Java. I expect it will work, once the spaces issue is solved.

>
> And by the way, I thought you had solved some Tcl build system issue?
> If so, you should commit it before you lose it.
>
In theory I have, I just haven't had the opportunity yet to test it :). I didn't want to commit the changes beforehand.

>
> > -        running a C example didn't work, as the drivers are not installed
> > The latter is an omission in the installation procedure - I never noticed that before
> but I seldom use this full build procedure.
>
> I am virtually positive from your and Greg's previous comprehensive test successes
> on this platform that "make install" does install the drivers.  However, if you look at
> the comprehensive test script, special measures had to be used to put the installed
> driver location on the PATH.  But I hasten to add this PATH change should only be
> done when the install tree is being tested, and otherwise (e.g., when testing the
> build tree) the installation location for the drivers should be removed from the PATH
> (as you will see the comprehensive test script implements).
>
Argh, sorry for the noise - they are indeed installed, but under a different directory than I thought. Okay, that sorts out that issue.

>From your last message:

Hi Arjen:



It sounds like you are making good progress on this platform as well.

For all three Windows platforms I look forward to your commits solving the straightforward issues that you find and I assume the problem above is of that type.



However, if this or some other issue is tricky for you to solve, feel free to consult in detail with me using the normal report tarball method (which typically is required to give me all the details that I need to help you).

Well, there is indeed a conundrum - the diff executable. I have a solution in mind (using a temporary environment variable) to transfer the context from the shell script to CMake (this involves effectively going from one operating environment - MinGW-w64/MSYS2 - to another - "bare Windows") and I expect it to work, but I need to test it. And it may be there is a more elegant and robust way. Will keep you informed.

Regards,

Arjen





DISCLAIMER: This message is intended exclusively for the addressee(s) and may contain confidential and privileged information. If you are not the intended recipient please notify the sender immediately and destroy this message. Unauthorized use, disclosure or copying of this message is strictly prohibited. The foundation 'Stichting Deltares', which has its seat at Delft, The Netherlands, Commercial Registration Number 41146461, is not liable in any way whatsoever for consequences and/or damages resulting from the improper, incomplete and untimely dispatch, receipt and/or content of this e-mail.