Language Reference

General Service and Utility Procedures

All of the procedures described in this chapter that do not belong to the xlf_fp_util module belong in this category. To ensure that the functions are given the correct type and that naming conflicts are avoided, use these procedures in one of the following two ways:

1. XL Fortran supplies the module xlfutility, which contains the interfaces and data type definitions for these procedures (and the derived-type definitions required for the dtime_, etime_, idate_, and itime_ procedures). XL Fortran flags arguments that are not compatible with the interface specification in type, kind, and rank. These modules enable type checking of these procedures at compile time rather than at link time. The argument names in the module interface are taken from the examples defined below. The following files are supplied for the modules xlfutility and xlfutility_extname:
   

   File names	File type	Locations
   xlfutility.f xlfutility_extname.f	source file	/usr/lpp/xlf/samples/modules
   xlfutility.mod xlfutility_extname.mod	module symbol file (32-bit)	/usr/lpp/xlf/include_32_d10 /usr/lpp/xlf/include_32_d7 Note: The files in these directories are exact copies of one another.
   	module symbol file (64-bit)	/usr/lpp/xlf/include_64

   You can use the precompiled module by adding a USE statement to your source file (see USE for details). As well, you can modify the module source file and recompile it to suit your needs. Use the xlfutility_extname files for procedures compiled with the -qextname option. The source file xlfutility_extname.f has no underscores following procedure names, while xlfutility.f includes underscores for some procedures names (as listed in this chapter).

   If there are name conflicts (for example if the accessing subprogram has an entity with the same name as a module entity), use the ONLY clause or the renaming features of the USE statement. For example,

   USE XLFUTILITY, NULL1 => DTIME_, NULL2 => ETIME_
   

2. Because these procedures are not intrinsic procedures:
   • You must declare their type to avoid potential problems with implicit typing.
   • When compiling with the -U option, you must code the names of these procedures in all lowercase to match the names in the XL Fortran libraries. We will show the names in lowercase here as a reminder.

   To avoid conflicts with names in the libc library, some procedure names end with an underscore. When coding calls to these procedures, you can:

   • Instead of entering the underscore, use the -brename linker option to change the name at link time:

     xlf -brename:flush,flush_ calls_flush.f
     

     This method works best if you need to rename only a small number of procedures.

   • Instead of typing the underscore, use the -qextname compiler option to add it to the end of each name:

     xlf -qextname calls_flush.f
     

     This method is recommended for programs already written without the underscore following the routine name. The XL Fortran library contains additional entry points, such as fpgets_, so that calls to procedures that do not use trailing underscores still resolve with -qextname.

   • Depending on the way your program is structured and the particular libraries and object files it uses, you may have difficulty using -qextname or -brename. In this case, enter the underscores after the appropriate names in the source file:

     PRINT *, IRTC()  ! No underscore in this name
     CALL FLUSH_(10)  ! But there is one in this name
     

   If your program calls the following procedures, there are restrictions on the common block and external procedure names that you can use:
   

   XLF-Provided Function Name	Common Block or External Procedure Name You Cannot Use
   mclock	times
   rand	irand

Note:

The mvbits subroutine that was in XL Fortran Version 2 is now an intrinsic subroutine, MVBITS(FROM, FROMPOS, LEN, TO, TOPOS).