From:	"Oliver Roese" <oroese AT edina DOT xnc DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Rhide: How to configure the searching facilities!
Date:	Mon, 23 Nov 1998 15:31:32 +0100
Organization:	Xenologics Networks & Communications GmbH
Lines:	100
Message-ID:	<73brtu$av5@tim.xenologics.com>
NNTP-Posting-Host:	ip206.xnc.de
X-Newsreader:	Microsoft Outlook Express 4.72.3110.5
X-Mimeole:	Produced By Microsoft MimeOLE V4.72.3110.3
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

Since the beginning Rhide had a interface to grep, but it never worked for
me.
Recently i found out what went wrong:
Grep.exe is started with a parameter missing (the -H parameter is needed.)
To correct this you have to overwrite the "RHIDE_GREP" variable in Rhide.
For example, if you add the following two line to your [rhide]-section of
your djgpp.env
(or to your rhide.env):

SEARCH_DIRS=%DJDIR%/include %DJDIR%/lang/cxx $(INCLUDE_DIRS) .
RHIDE_GREP=grep.exe -Hnwi $(prompt arguments for GREP) $(addsuffix
/*.h,$(SEARCH_DIRS))

you can search case-insensitive for words in all *.h-files in all standard
includedirectories and in the current one.
That works, but there are two free search tools available, that may
integrate
with Rhide: "ID-Utils" and "Ctags".
("Id-Utils" is available from any djgpp-mirror for example from
ftp://teeri.oulu.fi/pub/msdos/programming/djgpp2/ Look for "Idu32b.zip" and
"Idu32d.zip".
A good tagfilegenerator is available from
http://fly.hiwaay.net/~darren/ctags/.)

Wouldnt it be nice to use them?
It is possible to use "Id-Utils", but there are some drawbacks:
1) "lid.exe" can handle only one database  at a time.
2) "lid.exe" dont understand C-syntax. If you search a common symbol your
message-window will be flooded  with a lot of references.

The most attractiv tool is ctags.exe.
However i had difficulties to convert tagfiles to errormessages, suitable
for rhide.
At the end i wrote a gawk-script.
If you want to use ctags in this way you have to download and install gawk.
(Look for "gwk303b.zip" and "gkw303d.zip" at any djgpp-mirror).
A different awk will not work!
Sorry for the unconvenience, but i couldnt figure out a better way.
But i am sure you never want to miss it, after using it once.
Then add the following to your djgpp.env:

[rhide]
....
SEARCH_DIRS=%DJDIR%/include %DJDIR%/lang/cxx $(INCLUDE_DIRS) .
#RHIDE_GREP=grep.exe -Hnwi $(prompt arguments for GREP) $(addsuffix
/*.h,$(SEARCH_DIRS))
#RHIDE_GREP=lid.exe -Rgrep $(prompt arguments for LID)
RHIDE_GREP=gawk -f%DJDIR%/bin/tag2err.awk rexp=$(prompt arguments for GAWK)
$(addsuffix /tags,$(SEARCH_DIRS))

#if you use "lid.exe"
[lid]
IDPATH=%DJDIR%/include/ID


Convert the following to a file tag2err.awk and place it in a convenient
place, presumably %DJDIR%/bin.

#tag2err.awk
#Searches a regular expression (given on the commandline) in a tagfile
#produced by ctags and generates "gnu-style" errormessages,
#which Rhide can handle.
#"gawk -f<path>/tag2err.awk rexp=<regular_expression> <tagfile> ..."
#searches for <regular_expression> in all <tagfile>'s.
BEGIN\
{
#A Gawk extension.
  IGNORECASE=1
}

#Modify given regular expression to match only whole words.
$1~("^" rexp "$")\
{
  if (rexp == "")
#Avoid annoying delays.
    exit
  fn = $2
#This works for strings generated by "Exuberant-Tags" too.
#You can get it for free from http://fly.hiwaay.net/~darren/ctags/.
#"gensub" is another Gawk-extension.
  pat=gensub(/[^\t ]+[\t ]+[^\t ]+[\t ]+\/\^?([^$/]*)\$?\/.+/, "\\1", -1,$0)
  lc = 0
  while ((getline < fn) == 1) {
    lc++
    if (index($0,pat) == 1) {
      print (fn ":" lc ":" $0)
    }
  }
#Make the file available for the next cycle.
  close(fn)
}



I hope i have helped you. :-)

- Raw text -