README

 HUM -- HTTP Upload Module -- Instructions

 Copyright (c) 2000 Edward V. POPKOV
 Distributed under GPL, see COPYING for details

 "Edward V. POPKOV" <evpopkov@carry.neonet.lv>

 $Id: README,v 0.7 2000/03/01 02:32:43 root Exp root $

 $Log: README,v $
 Revision 0.7  2000/03/01 02:32:43  root
 Don't run as root.

 Revision 0.6  2000/02/18 20:29:31  root
 Check for install in configure, generate temp. filename if output file
 exists

 Revision 0.5  2000/02/09 22:01:11  root
 better memory usage, small file upload bug fixed

 Revision 0.4  2000/01/31 13:56:27  root
 Added file pump, cosmetic changes

 Revision 0.3  2000/01/28 02:28:05  root
 Bugfixes: filename, log fallback. Added: buffered read

 Revision 0.2  2000/01/25 23:23:09  root
 Added strict filename conversion, minor changes

 Revision 0.1  2000/01/25 16:26:28  root
 Fixed bug in configure script and missing include of errno.h

 Revision 0.0  2000/01/22 17:25:52  root
 Initial check-in


 Contents

 1. What is HUM?
 2. Requirements
 3. Installation
 4. Distribution
 5. Acknowledgements

 1. What is HUM?

 HUM is a CGI module that processes multipart/form-data input, extracts
a file being uploaded and saves it under compiled-in dir preserving original
filename.
  If the file already exists, HUM calls tmpnam() function to generate
unique file name. Note that this can result in saving the file in a totally
different location. This depends upon P_tmpdir constant found in <stdio.h>.
See tmpnam(3). This feature is pretty much experimental. There are several
ways to generate unique filename in *nix, and I have choosen tmpnam() just
because it's POSIX compliant. If you believe HUM must use tempnam, or mktemp,
or whatever other way to avoid overwriting existing files, please let me know.
 In many senses HUM is referemtial implementation and some parts must be
changed to make it more stable and usable in extra heavy load enviroments.
 Note that since version 0.7 HUM refuses to run with root (e)uid/(e)gid. It's
not configurable, don't even ask. You can change HUM behaviour by altering
the code, but you aren't encouraged to. Running your HTTP server under
root account is asking for troubles.

 2. Requirements

 At present stage development occurs on PC with Linux 2.0.38 and the following
software:

 -GNU libc5
 -GNU cc 2.7.2.3
 -Apache 1.3.1
 -bash 1.14.7(1)
 -w3m, Netscape Navigator

 All the code is written with portability in mind. If something doesn't work 
for you as expected, send me stdout and stderr of configure script and make. 

 3. Installation

 HUM is shipped together with configure script and Makefile to adhere
GNU software distribution guidelines. Your normal way is

# ./configure
# make
# make -n install

 Check dirs and permissions, then

# make install

 For configure script and Makefile to work you need to have cp, rm, sed,
and bash available.
 Read the bottom part of INSTALL to see how to tell configure about your
system.

 4. Distribution

 NUM is currently GPL'ed, see COPYING for details.

 5. Acknowledgements

 This project was rolled at eLance, proposed and sponsored by Bob PICKLES 

 In my work I'm using: 

  Linux, bash, Apache, w3m, Netscape, vim, autoconf, make, sed

 Thanks to developers of these and numerous other free and open source tools.

 Thx and take care

Edward V. POPKOV
Independent Developer

-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GCS/CC d-(+) s:- a C$ UL+++>$ P+ L++(+++)>$ E--- W--(+) N
o? K? w$ !O M? V PS+ PE Y PGP t? 5? X? R !tv b++@ DI D+ G++
e h* r++ y+++
-----END GEEK CODE BLOCK-----

End of README