--- name: "fflow" section: "1" platform: "linux" package: "forutil" version: "0.62-4" release: "Debian 5.0 (Lenny)" platforms: ["linux"] url: "https://new.carta.tech/fflow" --- # fflow(1) generate a flowgraph for fortran files. ## NAME fflow - generate a flowgraph for fortran files. ## SYNOPSIS **fflow \[-cefhinpqtuv\]** **-E***ext* **-I***dir* **-d***depth* **-l***level* **-o***file* **-s***name* **-x***name* **\[files\]** ## DESCRIPTION **fflow** Generates a (partial) flowgraph of a collection of fortran files. When fed with a couple of directories and/or files, **fflow** creates a flowgraph indicating the flow in your program. The output is written on stdout by default but can be written to file. **fflow** is very usefull for large programs where you only want to see the flowgraph of a certain part of a program. ## OPTIONS - **-c** Using this option instructs **fflow** to generate a complete flowgraph for each subroutine or function it encounters. By default, **fflow** will place a == Line XX == in the output for each subroutine or function for which it has already generated a flowgraph. Beware though that using this option can significantly increase the size of the output file and that it can take a little longer to generate all flowgraphs. - **-e** Using this option will cause **fflow** to not include functions without calls in the output. Helps to decrease the size of the output file. - **-f** Include the full path of files in the output. The default is to print the name of the file only. Using this option increases the memory **fflow** uses to store its data, so it might be unwise to use this option on msdos systems with their 640kb memory limit. - **-h, --help** gives a short description of every option you can use. - **-i** Using this option causes **fflow** to not print unknown function calls in the output. Usefull if you only want to have a flowgraph of a part of a source tree. - **-n, -n-** Using this option will cause **fflow** to tell what subroutines are never invoked. If you only want to check your files for this and do not want to generate a flowgraph, use -n- - **-p** Starts the flowgraph at the PROGRAM definition. When used with **-u** , **fflow** generates a flowgraph for PROGRAM only. Note: if you use the **-u** option, **fflow** will generate a flowgraph for the first PROGRAM statement it encounters. - **-q** Be really quiet (usefull if called from a script). Only error messages will be shown. - **-t** Use tabs instead of \| as level indicator. - **-u** Do not print unused functions in the output file. Using this option will cause **fflow** to not print functions and/or routines which are not referenced. Use the **-n** or **-n-** option if you only want to print unused subroutines. - **-v** Be verbose. Using this option shows a lot of usefull information about the files being scanned. - **--version** Show the version number of **fflow** - **-E***ext* Extensions **fflow** must use to determine which files it has to scan. Up to eight -E are allowed. The default extension is .f - **-I***dir* Using this option tells **fflow** which directories it has to search for f77 files. Only the contents of that directory are scanned. Subdirectories are ignored. Up to 64 -I are allowed. - **-d***depth* Sets the maximum recursion depth to depth. The default value is 64. This recursion depth is the maximum depth of function calls, and is mainly used as a test for detecting recursion. If **fflow** manages to get to the default value, this probably indicates that somewhere in your program infinite recursion might occur. If **fflow** aborts saying it has reached the maximum recursion depth, it means that this number has been hit. If you are absolutely sure that recursion does not appear in your program, you can increase this number to the maximum value of an unsigned integer on your system. The default value of 64 should be enough for any program. - **-l***level* Using this option will cause **fflow** to generate a flowgraph which is only *level* levels deep. This option can be very usefull to get a general idea on the flow in a very large program. It also will help reduce the size of the generated flowgraph. - **-o***file* file to write flowgraph to. The default is stdout. - **-s***name* This option causes **fflow** to start the output file with the named function. When used with **-u** , **fflow** generates a flowgraph for this function only. - **-x***name* Tells **fflow** to exclude the named file when scanning the files found with the *-I -E* options. Up to 64 -x options can be given. ## EXAMPLES **fflow -p -u fortran-files** generates a flowgraph for your program only. **fflow** **-s***name* **-u fortran-files** generates a flowgraph for *\[name\]* only. **fflow** **-I***dir1* **-I***dir2* **-E***.f* **-E***.f,v* generates a flowgraph from files ending in *.f* and *.f,v* from the directories dir1 and dir2. ## DIAGNOSTICS The following warning messages can occur: - **WARNING: no program entry point found in file** *\* a call statement was encountered before a PROGRAM statement was found. **fflow** will use the name of the file where this warning occurs as a program name and continue. - **Write failed** This error can occur when **fflow** is writing a flowgraph to file and a write error occurs. The most likely reasons for this error to occur are *disk full* or *no space left on device* The flowgraph generated by **fflow** can be very huge. We have seen a complete flowgraph of a single function which was in excess of 50MB, so it's easy to fill up your disk. When you encounter such large flowgraphs, you might want to use the **-l***level* option in order to generate a useable flowgraph. ## DISTRIBUTION POLICY All utilities in this fortran utility suite are distributed under the GNU Public License. ## COMPATIBILITY **fflow** has run under HP-UX 9 and 10, SGI IRIX 5.2 & 5.3, DEC OSF1, SunOs 4.1 and 4.3, Cray Unicos, IBM AIX 3.2 & 4.1, Linux and MSDOS 6.2. **fflow** also runs in a dos box under Windows 3.1, 3.11 and Win95. An MS-Windows version, Winfflow, is also available. ## SHORTCOMINGS The output of **fflow** contains a flowgraph for each subroutine or function found in the input files. When you give **fflow** the *-c* option, it can take a long time to write the flowgraph to file if you have fed it a collection of source files with a huge number of calls. This can be annoying especially when running **fflow** on MSDOS. Currently, **fflow** only looks at CALL, so although FUNCTION is found, these do not show up in the flowgraph. ## BUGS The MS-DOS versions of these tools can cause problems under MS-DOS if you feed them with a large collection of files, few files with a lot of calls or a combination of both. This is totally due to the 640kb memory limit MS-DOS has. If you think you have found a bug, you can send email to the author, describing what the bug is, how you generated it and if you can reproduce it. Also specify the platform where this bug occured. ## AUTHOR Koen D'Hondt ripley@xs4all.nl (C)Copyright 1995-1996 by Ripley Software Development ## SEE ALSO [**ftags**(1)](/ftags/1.md), [**ffscan**(1)](/ffscan/1.md), [**scan\_commons**(1)](/scan_commons/1.md), [**get\_common**(1)](/get_common/1.md), [**list\_commons**(1)](/list_commons/1.md)