#6 .f90, ifort, gcc

closed
nobody
None
7
2013-05-06
2013-03-31
No

Dear all,

I observed that [ pfunit-Bugs-3571838 ] is still present in the actual git repository e157d3b939558.

Additionally I was not able to use pFUnit with ifort (IFORT) 13.1.0 20130121 plus gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3 out of the box with the stated commit.
Compilation of the _wrap.F90 modules was not possible due to trailing whitespaces in the output of CONCATERNAT(obj,_private). Activating HAS_CONCATERNATE_OPERATOR in fortran_tokens.h did the trick.

Also, the string "method" in skeleton.F90 was not replaced by the correct string, this was fixed by replacing it with TOSTRING(method) and including "refelction.h" with #define STRINGIFY_OPERATOR.

As all these features are alredy there, is it desired behaviour that they shall be activated by the user? If so, should this be documented somehow or incorporated in the Makefile? Took me quite a while to figure it all out...

Best regards and thanks a lot for this nice piece of software, Chris

Discussion

  • Christoph Piefke

    I am sorry, I overlooked the f2kunit branch. Are bugs still fixed in masterbranch?
    Since I am just starting to use pFUnit for my projects, would recommend using f2kunit yet?

    Regards, Chris

     
  • Tom Clune

    Tom Clune - 2013-03-31
    • priority: 5 --> 7
     
  • Tom Clune

    Tom Clune - 2013-03-31

    Chris,

    I'm very sorry that you are experiencing this issue. The issues you are encountering are not part of the core framework, and are not well covered by our regression tests. This is not an excuse as the items you list are critical for the framework to be usable.

    There was some instability in CPP due to recent upgrades of OS X, and it is possible that in fixing those, I've broken the preprocessing facility for other OS's. (I work mostly under OS X.)

    I will attempt to replicate the issue on our linux cluster later today. It should be very easy to fix, but not the kind of thing I expect users to have to deal with.

    I'd also like to let you know that within 1-2 weeks, I hope to role out pfunit 2.0_beta. In addition to a generally superior design owing to OO language features, there are a number of improvements in usability.

    Cheers,

    - Tom

     
  • Tom Clune

    Tom Clune - 2013-03-31
    • status: open --> pending
     
  • Tom Clune

    Tom Clune - 2013-03-31

    Chris,

    I just submitted a patch that I believe fixes the issues that you have identified. They were all closely related and came about due to variations in how to concatenate tokens using CPP. (Wish CPP wore more standard than one would think.)

    Please try a fresh pull, build, and install and see if this fixes the issue for you.

    Thank you for bringing this to my attention, and please write back immediately with more details if you are still seeing problems.

    - Tom

     
  • Tom Clune

    Tom Clune - 2013-03-31

    Chris,

    Just saw your earlier response (reading things out of order).

    The f2kunit branch is still not ready for primetime. The core architecture is there, but lots of little things need to be done before it will be useful for most projects. E.g. I still have not pulled in support for multidimensional arrays.
    Patches are to go on the trunk for some time.

    If you've not done much yet with pfunit and can wait 1-2 weeks, I think you'll be happier with 2.0. The interfaces are not completely backward compatible, so users will need to adapt their tests to some degree.

    Cheers,

    - Tom

     
  • Christoph Piefke

    Tom,

    thanks a lot for your effort!
    I am really looking forward to pfunit-2.0-beta then, and I will wait. I am sure, TDD with pFUnit wil be a big gain for my projects!

    Unfortunately, the newest commit still does not work out of the box.

    1. I had to add a rule for .F90 files to my Makefile, since usually I use .f90 as the extention for my projects, but the automatic driver file is called '$(PFUNIT)/include/driver.F90'. Also, the wrap-modules come with an .F90 extension.

    2. Exchanging the string \"method\" in the skeleton files is not executed, I recive output like:

    Failure in top::testfancyThing::method -

    where 'method' should be replaced by the name of the SUT, hence 'showFancyInfo' in this mock-case.

    Best regards, Chris

     
  • Christoph Piefke

    • status: pending --> open
     
  • Christoph Piefke

    Hi Tom,

    I get an pFUnit update-notification from SF via mail almost every day, but when I look at the git repo, there are no changes available. Am I doing something wrong?

    Best regards, Chris

     
  • Tom Clune

    Tom Clune - 2013-05-06

    Hello Chris,

    I've just discovered another trove of messages/bugs/requests on source forge that I never saw before - including this one. Very sorry that I had not responded.

    Hopefully you are moving on with the f2kunit branch and don't care about this issue any longer?

    Cheers,

    - Tom

    PS Am composing the release announcement just now. Complex numbers are in and I'd appreciate feedback on the assertions.

     
  • Tom Clune

    Tom Clune - 2013-05-06
    • status: open --> closed
     
  • Tom Clune

    Tom Clune - 2013-05-06

    fixed (but branch is now obsolete)

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks