Search                        Top                                  Index
HELP VED_GREP (Unix Only)                      Aaron Sloman October 1990
                               Revised (Birmingham version) October 1995


Provides a facility for searching a set of files for a specified
string, and reading the result into a VED buffer. The name "grep" comes
from the name of the corresponding Unix search library, and is derived
from the concept of a Generalised Rexgular EXPression matcher.
(VED now also includes such a matcher. See REF * VEDSEARCH.

         CONTENTS - (Use <ENTER> g to access required sections)

 -- Introduction
 -- The <flags> argument
 -- -- Example: searching the TEACH directory
 -- Use "man" to find out about options avilable
 -- Getting the <search-string> from the VED buffer
 -- Getting ALL the arguments from the VED buffer
 -- Removing colons after file names
 -- Specifying which search program to use
 -- Using <ENTER> getit, to read in a file found

-- Introduction -------------------------------------------------------

The formats available for invoking ved_grep are as follows:

<ENTER> grep [<flags>] <search-string> file1, file2 ...
<ENTER> grep [<flags>] <search-string> <file-pattern>
<ENTER> grep [<flags>] <file-pattern>
<ENTER> grep [<flags>]

These formats are explained below.

The <ENTER> grep command takes the <search-string> (if provided,
otherwise a string obtained from the VED buffer) and searches through
all the files specified, looking for occurrences of the string.

The flags, if provided, specify how the searching should be done, e.g.
whether it should be caseless, and the format of the output, e.g.
whether line numbers should be shown, etc. as described in MAN * EGREP

If no files or <file-pattern> are provided on the command line then the
files to be searched are specified on the current line in the VED

All output goes goes into a temporary file in VED. The program uses
-vedgenshell- and -pipein- for this purpose. (See HELP * PIPEUTILS)

There are several ways the user can control how the search is done
or what information is printed out.

-- The <flags> argument -----------------------------------------------

The optional <flags> argument consists of one or more substrings
starting with '-'. Exactly which flags are available to control
the search will depend on which search command is used.

Often the <flags> argument is not provided. Commonly used flags are

        InVert the search. I.e. report only lines NOT containing the
        search string.

        List the files containing the string, but don't report every
        occurrence. Only the file names are listed.

-- -- Example: searching the TEACH directory

To find every occurrence in the TEACH directory of the string 'grammar'
in upper or lower case, do

    <ENTER> grep -i grammar $usepop/pop/teach/*

This runs a Unix file searching command (default egrep) with the

        -i grammar $usepop/pop/teach/*

The "-i" flag tells the search command to ignore case of letters.

The user can specify which Unix search command to use, as there are
sometimes faster versions available than "egrep".

-- Use "man" to find out about options avilable -----------------------

The Unix "man" command will provide information on flags availble and
the forms of string patterns that can be in the <search-string>. E.g.

    <ENTER> man egrep

if you use the "egrep" option, or

    <ENTER> man ggrep

if you use "ggrep" etc.

See below on how to specify which search command is used.

-- Getting the <search-string> from the VED buffer --------------------

If no <search-string> argument is provided and the <file-pattern>
contains no spaces (e.g. <ENTER> grep -l *.p) then the item to the right
of the cursor in the VED buffer is taken as the <search-string>, where
the item is delimited on the right by any of the following

        space tab newline ( ) ; , [ ] { } | > < ~ & * ? ` ' "

and a trailing . or , is ignored.

For example, if the cursor is just to the right of the arrow on the next
    -> prmishap,

then the command

    <ENTER> grep $popautolib/*

is equivalent to

    <ENTER> grep prmishap $popautolib/*

and will print out information about all files in $popautolib that
contain the string 'prmishap'.

-- Getting ALL the arguments from the VED buffer ----------------------

If no <search-string> or <file-pattern)> is given, then it uses the
whole of the current VED line as argument. Thus the above commands would
be equivalent to simply giving the command

    <ENTER> grep

while the VED cursor is on a line containing nothing but

    prmishap $popautolib/*

-- Removing colons after file names -----------------------------------

Normally the output of commands like "grep" includes a colon immediately
after each file name. This can be inconvenient if you wish to use the
"^f" option described in HELP * VEDEXPAND to give a command using the
file name. So by default, ved_grep removes the first colon in each

This can be suppressed by putting the following command in your
vedinit.p file:

    global vars grep_leave_colons = true;

The default is false: i.e. colons are removed.

-- Specifying which search program to use -----------------------------

The file-searching command is specified by assigning a string to the
variable -grep_search_command-.

E.g. depending on what is available on your machine, you could put one
of these in your vedinit.p if desired.

    vars grep_search_command = 'ggrep';
    vars grep_search_command = 'ngrep';

Otherwise the command name defaults to 'egrep', because that allows
flexible patterns.

Usually if the Free Software Foundation's GNU grep is available
(sometimes called "ggrep") it is the fastest search program to use.
Otherwise if 'bm' is available it is fastest. Otherwise try 'egrep' and
'grep' and decide which is faster. Note that 'grep' usually allows fewer
options in the search string.

The corresponding 'man' file (e.g. MAN * EGREP) should tell which flags
are available to control the search, and what search patterns can be
used in the <search-string>.

-- Using <ENTER> getit, to read in a file found -----------------------

If you have this in your vedinit.p file, then

    global vars grep_leave_colons = true;

<ENTER> grep will leave colons after filenames in its output file.

You can then use the <ENTER> getit command to go to the file found and
the line found. See HELP * VED_GETIT. (Available at Birmingham.)

--- C.unix/help/ved_grep
--- Copyright University of Sussex 1990. All rights reserved. ----------

--- $poplocal/local/help/ved_grep
--- The University of Birmingham 1995.  --------------------------------