I have a list of full file paths. All of the full file paths look like "/dir1/dir2/dir3/s..". I want to completely remove the s. from the filename. There is the possibility of a filename being plural, for example s.asdfs.cpp. I do not want to remove the second occurence of s. since that is part of the actual filename and not a reoccuring theme in every full file path in the list.
Running the following in shell works as I want it to:
echo /dir1/dir2/dir3/s.filenames.cpp | sed 's#\(.*\)\/s\.\([^\/]*\)#\1\/\2#g'
Gives the desired result of:
/dir1/dir2/dir3/filenames.cpp
But if I run the following in php:
$formatted_filename = exec("echo ".$filename." | sed 's#\(.*\)\/s\.\([^\/]*\)#\1\/\2#g'");
where
$filename = /dir1/dir2/dir3/s.filenames.cpp;
And then in my bash shell run
php -q script_name_that_contains_command_above.php > test.html
and refresh my firefox browser that displays test.html I get very strange results. In place of where this edited file path should be listed I get
<strange box>/<strange box>
where
<strange box>
is a small box with 2 rows and 2 columns consisting of 0's except for the bottom right cell. The first occurence has a 1 in the bottom right cell, and the second occurence has a 2 in the bottom right cell.
The sed command works, but php, or the exec command is interpreting it incorrectly I believe. Any ideas?
The solution for using exec with this particular regular expression was to use the php function, preg_replace()
preg_replace("/\/s/./", "/", $filename);