[RegExp] Finding Fortran functions in a directory

[guerom00]

Hi there,

Given a directory full of Fortran files, I want to list all the functions using grep. A Fortran function is declared as follow :

CODE

[type] function the_name(the_args,another_args,...)

where “[type]” is optional. To sum up, one has to match
1) the beginning of the line
2) any number of spaces
3) any characters up to a space
4) the word “function”
5) the name of the function
6) the opening brace

Steps 5) and 6) are here to avoid matchind the “end function” statement

In Oniguruma, the following does the job :

CODE

(?ix)^\s*(.*)\s*\b(function)\b\s*[a-zA-Z0-9_]*\(

but I don't know which RegExp format is used by grep (damn : it does not seem to know “(.*)” ) but the above expression does not work.

Any help would be much appreciated

TIA

[Philjfry]

Hi guerom00 !
Why dont you just
grep "function" | grep -v "end"
?
You might have to add stuff to weed out "function" written in comments.

This post has been edited by Philjfry: Oct 18 2006, 01:51 PM

[guerom00]

Hi guerom00 !
Why dont you just
grep "function" | grep -v "end"
?
You might have to add stuff out "function" written in comments.

Thanks Philjfry Didn't know about the -v flag

It appears that

CODE

grep "function [A-Za-z0-9_]*(" *.f *.f90 | grep -v "end"

do what I want.

Here is the end result

CODE

#!/usr/bin/env csh
echo "Finding tag "$2" in directory "$1
cd "$1"
if(-e .tags) goto exist
grep "subroutine [A-Za-z0-9_]*(" *.f *.f90 | grep -v "end" > .tags
grep "function [A-Za-z0-9_]*(" *.f *.f90 | grep -v "end" >> .tags
grep "entry [A-Za-z0-9_]*(" *.f *.f90 | grep -v "end" >> .tags
echo >> .tags
set nlines=`cat .tags | wc -l`
@ nlines = $nlines - 1
if($nlines == 0) then
echo 'No functions or subroutine found'
exit 1
endif
set files=`cat .tags | cut -f 1 -d :`
set names=`cat .tags | cut -f 1 -d :`
set subfun=`cat .tags | cut -f 2 -d :`
rm .tags
touch .tags
set i=0
set j=0
while(1 < 2)
@ i ++
if("$subfun[$i]" == "subroutine" || "$subfun[$i]" == "function" || "$subfun[$i]" == "SUBROUTINE"\\
|| "$subfun[$i]" == "FUNCTION" || "$subfun[$i]" == "entry" || "$subfun[$i]" == "ENTRY") then
@ j ++
@ k = $i + 1
set names[$j]=`echo "$subfun[$k]" | cut -f 1 -d "("`
@ i = $k
echo "$files[$j]""|""$names[$j]" >> .tags
endif
if($j == $nlines) break
end
exist:
set file_name=`cat .tags | grep "|$2\b" | cut -f 1 -d "|"`
echo "opening "$file_name
if($file_name == '') exit
set line=`grep -n "subroutine $2" "$file_name" | grep -v "end" | cut -f 1 -d :`
echo 'line '$line
if($line == '') set line=`grep -n "function $2" "$file_name" | grep -v "end" | cut -f 1 -d :`
if($line == '') set line=`grep -n "entry $2" "$file_name" | grep -v "end" | cut -f 1 -d :`
echo $line
if($line != '') then
mate -l $line $file_name
else
mate $file_name
endif

An equivalent, not perfect, very ugly () but which works not bad…, of the “etags” command and its integration with XEmacs, for TextMate
I know : it could be more easy/beautiful/efficient using Ruby or Python but what can I do : I only know csh scripts

Edo : The last refinement is

CODE

grep "^[^c\!].*function [A-Za-z0-9_]*(" *.f *.f90 | grep -v "end"

because I picked up some spurious things before…

This post has been edited by guerom00: Oct 18 2006, 10:38 PM