This implies to show the menu as we traverse the gopher tree with the
initial connection to site.C, and follow gopherspace to site.A, site.B,
and site.C, but don't follow anything from site.X out port 9999, while
building the initial database for future use with the -B flag.
Added support for -u username to allow jughead to act as a search engine
with the uid username. Thank riddle@is.rice.edu for this suggestion.
Fixed a problem of reading the database and tables if they exist in a
location other than jughead itself. Thank you rzakon@mitre.org for
bringing this to my attention.
Release: version 0.7.3, released: May 14, 1993.
doyle@liberty.uc.wlu.edu reported that "a AND b" returned "a" if
"b" did not exist. I believe things are now working.
Release: version 0.7.2, released: May 12, 1993.
doyle@liberty.uc.wlu.edu reported that "a AND b" returned "b" if
"a" did not exist, so I had to fix it.
Release: version 0.7.1, released: May 11, 1993.
Fixed a problem where "a AND b AND c" would be evaluated to
"a AND b OR c" if "a AND b" evaluated to nothing. Altered the way
the information gets logged by logging the port jughead is using, to
support determine which jughead logged the information when using
multiple jugheads are logging to the same file. The bug and logging
feature was reported and suggested by: doyle@liberty.uc.wlu.edu.
Release: version 0.7, released: May 8, 1993.
Fixed a problem that gave incorrect results when boolean searching on
3 or more words. Thank you doyle@liberty.uc.wlu.edu for bringing
this to my attention. Moved jughead out beta state.
Release: version 0.6_beta, released: May 7, 1993.
Fixed a bug that would dump core when talking to a PC Gopher
server that was not following gopher protocol. Added the use
of the -m flag. This flag will suppress printing of the menus
when not building the dataFile, indexes or running as a server.
This flag will also enable the printing of the menus when building
the dataFile. Fixed a bug that would not acknowlege an invalid
port for the exceptions hosts, which is the host and port not to
traverse. Gave support for porting jughead to A/UX, thank you
hagberg@cumc.cornell.edu for bringing this to my attention. Fixed
a bug that would dump core when attempting to retrieve an empty
directory from a gopher+ server, thank you lindner@boombox.micro.umn.edu
for bringing this to my attention. Added support for partial word
searches, where the form of partial word searches is "matchPart*", "*"
is invalid, and any characters following the asterik are ignored. Gave
support for a selector string to not traverse with use of the -X flag.
This flag supports a wildcard and will not traverse any of these items.
Example:
jughead some.gopher.site -X "ftp:*"
will not traverse any ftp items pointed to by some.gopher.site.
Fixed a bug the would dump core if jughead received more than one
SIGHUP signal before being shutdown. Did some code cleaning up and
optimization.
Release: version 0.5_beta, released: April 9, 1993.
Modified the code to support a secondary hash table. The table that
gets read into memory is much smaller than previously, while the time
to do the boolean operations and search the tables takes slightly longer,
however, I doubt users will notice the difference.
Note: This is a major change with the index tables. When running
jughead with the -B flag, you no longer direct stdout to a file.
What happens now is 2 files automaticlly get created, 1 with a ".ix" and
the other with a ".ih" extention on the file created with the -b option.
Furthermore when invoking jughead as a search engine with the -S flag
you only need to give the name of the database file created with the -b
flag.
Release: version 0.4_beta, released: April 3, 1993.
Added support for the 'a', 'A', and 't' options, which prints the
port, host, number of directories, or port, host, the directories, and
the time required to process a menu or build the index table or
load the index table into memory respectively.
Added the code from Mic Kaczmarczik: mic@bongo.cc.utexas.edu, which
supports an optional -p port flag when running as a search engine.
When using the -b option the datafile now gets sorted to remove all
duplicates, and no longer gets written to datafile.sorted.
The -o and -O flags are now meaningless and may soon disappear.
Release: version 0.3.1_beta, released: March 28, 1993.
Fixed a problem using vsprintf compiled on sun.
Release: version 0.3_beta, released: March 28, 1993.
Modifications include consolidating various #defines into the
Makefile. Now keeps track of hosts we can't connect to and will
not attempt to connect to these hosts again.
Added support for compilation on a sun.
Initial release: version 0.2_beta, released: March 25, 1993.
jughead can be acquired via gopher by gophering to gopher.cc.utah.edu
and looking in:
About U of U Gopher
Gopher Tools
jughead
as jughead.ReadMe and jughead.VERS.tar.Z
or via anonymous ftp from ftp.cc.utah.edu in /pub/gopher/GopherTools
as jughead.ReadMe and jughead.VERS.tar.Z, where ".VERS" is the current version.
The file jughead.VERS.tar.Z is a compressed tar file which contains:
About.jughead
dirTree.c
dirTree.h
getargs.c
jughead.1
jughead.c
jughead.h
jughead.survey
Makefile
searchCmnds.c
search.c
sockets.c
tree.c
tree.h
utils.c
utils.h
Create yourself a directory for jughead, which should not be in a
gopher data directory, and download the file jughead.VERS.tar.Z.
Once you have downloaded jughead.VERS.tar.Z, where "VERS" is the version
of jughead, the following command will uncompress and untar the files:
uncompress -c < jughead.VERS.tar.Z | tar xfv -
To view the man page, prior to running make install you can:
make viewman
To make the jughead application you should edit the Makefile and verify
the following values are correct:
THEVERSION The version of jughead, only edit this if you make changes.
THEHOST The fully qualified name of the machine running jughead.
JUGHEADHELPDOC The gopher server serving the jughead help document.
DEFBOOLOP The default boolean operator if none is given between words.
THECATCOMMAND The cat system command, the %s is required.
THERMCOMMAND The rm system command, the %s is required.
THESORTCOMMAND The sort system command, the %s is required.
THETMPFILENAME The location and name of jughead's temporary file.
THEPORT2USE The port to use when jughead is a search engine.
MANSUFF The suffix of the man page and where to install the manpage.
DESTDIR The location of the jughead executable. You may not want
this to be publicly accessible.
DESTMAN The location of the jughead man page.
PERMISSIONS The permissions for the jughead executable.
VARARGS Commented out if you have <stdarg.h>.
VSPRINTTYPE The type vsprintf() returns.
PROTOTYPES Uncommented if you want to use prototypes.
POSIXSOURCE Needed when compiling under A/UX.
NOWARNINGS Uncommented if you don't want to see any compiler warnings.
OPTIMIZATION Any optimization flags you want to have.
CC The C compiler to use.
LINT The lint program to use if need me.
LINTFLAGS Any lint flags you want to set.
You will want to place the "About.jughead" document on your somewhere
on your gopher server, and then give the proper definition of JUGHEADHELPDOC
in the Makefile. You may want to copy the file "About.jughead" in the same directory as the jughead link, so the users will have knowlege how to search
the tables built by jughead.
Create a link file somewhere in your gopher data directory. Mine looks like:
Name=Search all UofU menus using jughead
Numb=4
Type=7
Port=3000
Path=
Host=gopher.cc.utah.edu
Now do a "make", "make all", or "make install" to compile jughead. If
you get any compilation errors please inform me of the problem.
The steps I did to get jughead running as a server follow:
jughead -tb data -X "ftp:*" .utah.edu gopher.cc.utah.edu
jughead -tB data
jughead -tS -l data.log data &
jughead is not inetd compatible.
There exists a mailing list for discussion about jughead called:
jughead-news@lists.utah.edu To get on this mailing list send
mail to: jughead-news-request@lists.utah.edu
Things still to do:
1) Move the Makefile macro THEHOST and JUGHEADHELPDOC to a
.rc or .conf file or something like that.
2) Add support for the following special commands:
?pathway selectorString
returns the gopher pathway of
'selectorString'.
?type t
returns only items of type 't'.
Allow an optional [=n] to return the first 'n' items.
3) Make jughead inetd compatible.
4) Use the Path field in the link file to specify which database
to use, load the database into memory, do the search, and
release the memory. Maybe have this as an option.
How about initialy start up with no tables cached in memory,
when we receive a query read the appropriate table into memory
if there is none, and only release the memory if a new query
refers to an alternate database?
5) Add the ability to query other jugheads.
6) Consolidate the tree.c and dirTree.c files.
7) Either reenable or remove support for the -o and -O flags?
8) Use vfork() instead of fork() when acting as a search engine
to eliminate, as much as possible, any memory duplication.
9) Implement a no feed back flag which will give no feed back when
building the hash tables, which will make the creation of the
hash tables speed up by approximately 20 percent.
10) Implement the use of a synonym table.
Please email source changes, complaints, or suggestions to:
