• Solving common problems
• Why is the auto-opening of R so slow?
• When R auto-opens, the console window of the Fortran app is difficult to close and R closes automatically. Can this be changed?
• How do I get RFortran to recognize the new version of R I have installed?
• When I use Rget with allocatable string arrays, there is a length mismatch
• I get a compiler error when using the // operator from MUtilsLib
• I cannot see the variables inside R, but there is no error
• When Rgui.exe is automatically loaded the system freezes
• The pdf/postscript files I created don't work
• Why can't I see the Rgui console when R starts? I can only see the graphics windows (even when =.TRUE.)?
• What does my error code mean?
• R automatically closed
• Error when checking out RFortran source code: "repository.svn: Error parsing svn:externals property on - mutilslib-read-only"
• I don't have Visual Studio 2008 and/or IVF 11 - how can I use the solution/project files in older version of Visual Studio and/or IVF?
• Why do I get so many warnings when I compile RFortran?
• Who do I get a type mismatch when transferring a R variable to Fortran?

Solving common problems

If RFortran does not work it, please undertake the following before submitting an issue:

1. Check rcom was installed properly Follows the steps given in manual installation and ensure they have all been performed correctly.
2. Check log file, rfortran.log RFortran creates rfortran.log which includes errors, warning and diagnostic debug information. The default location is the directory of the Fortran executable using RFortran. Open rfortran.log to find information for diagnosing your problems.
3. Check environment variable, RFORTRAN_PATH During installation an environmental variable RFORTRAN_PATH is set to the install directory of RFortran. You can check its value by using the batch file 'RFortran\setup\Check_status_envvar_RFortran_Path.bat'
4. Check settings file, RFortran.set RFortran reads its settings from the file RFortran\settings\RFortran.set. Check that all settings are correct. See RFortran Settings for further information
5. Restart Visual Studio If you installed RFortran with Visual Studio open you will need to close Visual Studio and reopen it again. RFortran sets and uses a windows environment variable RFORTRAN_PATH (which should contain the path to RFortran). Visual Studio only reads environment variables from the system during startup and needs to be restarted after RFortran installation.
6. Check issue list Check the issues to see if others have had your problem
7. Submit an issue list Submit a detailed description of your problem as a new issue - be sure to include a detailed description of your problem, including a copy of the output from Visual Studio build (if there were build errors) and/or copy of the RFortran.log file (if there were run-time errors)

Why is the auto-opening of R so slow?

When a Fortran app with RFortran runs, it looks for Rgui.exe and if it isn't running, RFortran auto-opens a new Rgui.exe. To ensure Rgui.exe is able to fully start before RFortran sends requests to R, a deliberate delay is implemented. The length of this delay can be set in two ways:

• Globally: setting wait4Rdef parameter in RFortran_globalvars.f90

• Run-time: setting wait4RIn argument in Rinit see RFortran_Rinit_Rput_Rget.f90

When R auto-opens, the console window of the Fortran app is difficult to close and R closes automatically. Can this be changed?

This has been a problem since September 2008, and whilst the cause is unknown is likely due to changes in the rcom server which is outside of the control of the RFortran development team. If R is started prior to running your console application, this is no longer a problem.

How do I get RFortran to recognize the new version of R I have installed?

Simply re-run the installer and don't re-install the source code. You will need to update the parameter RGuiPathDef in RFortran_globalvars.f90 to the path for the latest version of the Rgui.exe

When I use Rget with allocatable string arrays, there is a length mismatch

Please see  issue 9 .

I get a compiler error when using the // operator from MUtilsLib

Please see  issue 13 .

I cannot see the variables inside R, but there is no error

Please see  issue 14 .

When Rgui.exe is automatically loaded the system freezes

Please see  issue 15 .

The pdf/postscript files I created don't work

Please see  issue 18 .

Why can't I see the Rgui console when R starts? I can only see the graphics windows (even when RguiVisible=.TRUE.)?

Please see  issue 19 .

What does my error code mean?

A list of error codes and possible rememdies is provided at Error Codes

R automatically closed

Please see  issue 25 .

Error when checking out RFortran source code: "repository.svn: Error parsing svn:externals property on -r21 http://mutilslib.../ mutilslib-read-only"

The source code for mutilslib is stored in seperate repository http://code.google.com/p/mutilslib/ and is automatically checked out when RFortran is checked out.

Often (but not always) this is caused by problems with proxy servers having difficulty accessing the multilslib repository. Please check that the proxy server settings in your SVN client are correct (for Tortoise SVN, this is under Settings|Network).

I don't have Visual Studio 2008 and/or IVF 11 - how can I use the solution/project files in older version of Visual Studio and/or IVF?

(with thanks to Dmitri Kavetski)

I. Microsoft Visual Studio 2005

In order to downdate to a VS2005 .sln file, change the version number and release date as follows: (after closing the solution file! - dont try to edit an open .sln file!)

From:

Microsoft Visual Studio Solution File, Format Version 10.00
# Visual Studio 2008

To:

Microsoft Visual Studio Solution File, Format Version 9.00
# Visual Studio 2005

No other changes are necessary.

Alternatively, if you are scared of text-editing files u r not suppposed to, you can just create a new 'native' MS Visual Studio 2005 solution from scratch, but this requires re-adding projects, which is inconvenient and just plain time-consuming.

II. Intel Visual Fortran v10.X

Intel Fortran has incremented the vernum of its .vproj files and also added some new options. If you have IVF10 and IVF11 coexisting, then these diffences are handled automatically.

However, if you do not have IVF11 installed, then you will get an error when opening an IVF11 .vproj with IVF10. Your newer projects will not be loaded.

In order to downdate to an IVF10 .vproj file, in the second line of the file change the version number and release date as follows (after closing the entire solution file! - dont try to edit an open .vproj file!)

From: IVF >11.0

Version="11.0"

To: IVF <10.1.029

Version="9.10"

If your project is a static library you will also need to remove the following:

ProjectType="typeStaticLibrary"

This change will allow loading and (seemingly) correct execution of Fortran projects.

However, you will still get an error message that the "Project file is corrupt". Presumably this happens when some newer options (added in IVF11) remain in a version 9.10 vproj file.

Fortunately, the error is effectively a warning, because the unsupported options are simply ignored.

To really get rid of warning oscillate the project settings and save!

Alternatively, if you are scared of text-editing files u r not suppposed to, you can just create a new 'native' IVF10 project from scratch, but this requires re-add files and options, which is error prone and just plain inconvenient.

Why do I get so many warnings when I compile RFortran?

These warnings occur because RFortran is compiled with F2003 standards checking and it uses a few minor non-standard IVF/CVF extensions (e.g. the LOC function and variables in the format statement) which generate warnings. These are of no concern.

As at v3.0.0 the number of warnings were: (with IVF 11.1)

• TEST_RFortran_IVF: 439 warning(s)
• RFORTRAN_IVF_LIB: 462 warning(s)

• RFORTRAN_CVF_LIB: 895 warning(s)

Who do I get a type mismatch when transferring a R variable to Fortran?

This maybe because R creates a "double" by default, e.g. "k=1" typed at the R prompt creates a double, see also  issue 86