module WhatLineIsTheError implicit none interface module function thing(it) implicit none integer :: thing integer, intent(in) :: it end function end interface end module WhatLineIsTheError submodule(WhatLineIsTheError) subm1_WhatLineIsTheError implicit none contains module function thing(it) implicit none integer :: thing integer, intent(in) :: it end function end submodule subm1_WhatLineIsTheError
Compiling with Intel(R) Visual Fortran Compiler 22.214.171.124 [IA-32]... warning #6178: The return value of this FUNCTION has not been defined. [THING]
It is but a minor problem. The error has no line association to the source so clicking it takes you to line 1. There have been other errors in SUBMODULE of this type that have been fixed, I suspect there are possibly a number of similar errors of the same class.
BTW, The syntax highlighter for in Fortran windows posts misses some keywords e.g. the word SUBMODULE is post #1. I suspect this maybe some third party feature that you have no control over, but if not.....
Yours pedantically, Andrew
The incorrect VS editor focus when clicking on the warning was reproduced and reported to Development under the internal tracking id noted. The lack of SUBMODULE keyword coloring was reported to the IDZ Forum developers. Thank you.
(Internal tracking id: CMPLRS-43270)
Kevin D (Intel) wrote:
.. The lack of SUBMODULE keyword coloring was reported to the IDZ Forum developers. ..
Fyi on a similar request made a couple of years ago that is pending: https://software.intel.com/en-us/forums/intel-visual-fortran-compiler-for-windows/topic/542008
My word time flies, 2+ years. I remember that post, that must be where I (correctly) had the idea in my head that the syntax was a third party tool. Thanks Kevin, maybe you should log that as a defect. OK is might not get fixed in a hurry but it is a defect and at least it is on a do list that hopefully gets continuous review......
From what I'm finding Steve took action on that earlier report FortranFan.
Andrew, it was filed as defect earlier w/our Forum developers who are reviewing the file. I had included the suggestion of continual review by Fortran developers also so hopefully that can be done. Thank you.
A year ago or so I provided to the forum developers a revised syntax coloring script that should be more up to date. I don't recall if they ever deployed it.
As a bizarre twist on the above, please let me share this message:
------ Build started: Project: common_v10.0, Configuration: Debug|x64 ------ Compiling with Intel(R) Visual Fortran Compiler 126.96.36.1992 [Intel(R) 64]... STUDENT_T_VARIATE.f90 C:\Users\ces1nk\AppData\Local\Temp\783213.i90: warning #6178: The return value of this FUNCTION has not been defined. [ASA^NIBETA] Build log written to "file://C:\Users\ces1nk\Dropbox\Shuttlebackup\Dept\Research\Programs\Common\v10.0\common_v10.0\common_v10.0\x64\Debug\BuildLog.htm" common_v10.0 - 0 error(s), 1 warning(s)
The problem in the above is [ASA^NIBETA]
There is a module called ASA and the module that was compiled to generate the above message (STUDENT_T_VARIATE) does USE ASA. But I have no idea what NIBETA is. A simple search for the characters (not whole word) NIBETA yields nothing independent of case. And I am pretty sure that all the functions in module ASA are assigned return values.
STUDENT_T_VARIATE does not have any submodules, not does ASA
Everything seems to run OK
Over several hundred other modules I have a few other similar examples of Warning 6178, none has been traced back to a 'real' problem. Each warning points to a different missing function, each in a different module.