This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Nir, > To: address@hidden > From: address@hidden > Subject: udunits problems > Organization: UCAR/Unidata > Keywords: 200403042037.i24KburV008355 The above message contained the following: > I am writing you again - after we managed to compile the libraries > on the Intel Compiler, we are now having problems with the time > variables. When we try to run CDFreader to read NCEP data we got the > error message (below the corresponding subroutine). We tried to > fix it by moving the include commands of the subroutine to the main > program (rather than the subroutne gridread where they originally > are), all date-time variables got the value 1 (exept the hour that got > 0). The include-file "/home/sapir/udunits-1.12.1/include/udunits.inc" defines the types of UTMAKE, UTDEC, and UTTIME -- so it's not necessary (and apparently an error) to define them within the "gridread" subroutine. If you remove the following lines > integer UTMAKE !!!!this is line 271 > integer UTDEC,uttime from the "gridread" subroutine, then the file should compile without the following errors: > fortcom: Error: CDFreaderII.F, line 271: This name has already been assigned a > data type. [UTMAKE] > integer UTMAKE > ..............^ > fortcom: Error: CDFreaderII.F, line 272: This name has already been assigned a > data type. [UTDEC] > integer UTDEC,uttime > ..............^ > fortcom: Error: CDFreaderII.F, line 272: This name has already been assigned a > data type. [UTTIME] > integer UTDEC,uttime > ....................^ > compilation aborted for CDFreaderII.F (code 1) > The same code is working perfectly other machines that are using the portland > group fortran compiler (I am using the Intel Fortran compiler). Apparently, the Portland Group Fortran compiler doesn't care how many times a symbol is declared as long as the declarations are the same. > Looks to me like a compiler problem. some dependency between intel fortran and > the subroutines of the UDUNITS library. > > Did anyone else has encoutered such a problem? > Any ideas for possible solutions? > > I will appreciate any help, > > Best regards, > > Nir. Regards, Steve Emmerson