home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
OS/2 Shareware BBS: 35 Internet
/
35-Internet.zip
/
extract4.zip
/
extract.doc
< prev
next >
Wrap
Text File
|
1996-11-06
|
5KB
|
106 lines
Usenet File Extractor Version 4.0
(C)1994-96 Turgut Kalfaoglu <TURGUT@VM.EGE.EDU.TR>
SHAREWARE
EXTRACT/MERGER package is shareware. Usage over two weeks requires a
registration fee of $20 be sent to the author:
Turgut Kalfaoglu
1378 Sokak 8/10
Alsancak, Izmir 35210
Turkey
I highly recommend that you register the associated software as well.
HIGHLIGHTS
* Converts Usenet .binaries. group data directly into files you can use
such as .ZIP, .GIF, .JPG, .ARJ, etc.
* Retries files that have missing segments.
* Accepts files with up to 30 segments.
* Expires old segments.
* Easily customisable.
* Requires: OS/2 2.x or above with Rexx loaded
IBM TCP/IP (for DUMPGRP.EXE)
UUDECODE.EXE (shareware or public domain)
FORALL.EXE (shareware or public domain)
10 MB or more of disk space, depending on how much
news you receive at once.
INSTALLATION
You need three directories for this program. Although everything
is configurable, here is how I have mine set up:
C:
|
+-- DOCS
+--- S
+---- PARTIAL
I recommend this hierarchical setup. In this case, put everything
into DOCS (or somewhere on your PATH..), leave the other two blank.
When DUMPALL terminates, you will have your binaries in the 'S' directory.
The package contains two .CFG files, EXTRACT.CFG and MERGER.CFG. These
two need to be configured - simply use your text editor such as 'E' or
'EPM' to configure these files. If you have not used DUMPGRP.EXE before,
you need to configure its configuration file as well - see its
documentation.
Examine and/or Edit DUMPALL.CMD file to get an idea of what is done.
DESCRIPTION
This package will collect news from an NNTP server (from any .binaries.
group) and extract the received information into files. Meaning, it will
'reconstruct' files sent in UUENCODE format, and in several pieces,
automatically. It will keep the incomplete sections for 20 runs,
attempting to complete the files at each run.
The package is run by DUMPALL.CMD that invokes each of the programs in
the necessary order. First, the old incomplete sections are backed up
into another file, and DUMPGRP.EXE is invoked. DUMPGRP.EXE will collect
the "new news" from your NNTP server, and place them into a file, (by
default, into "S.UUE") Then, S.UUE is processed by EXTRACT, which will
split the files it finds in this file into its temporary "partials"
directory, where segments of each file are kept. (Default:
"/DOCS/S/PARTIAL"). Information on these files is also kept in a file
called EXTRACT.SUM that has such trivia as filename,number of pieces,
and the number of retries for this partial file. In other words,
EXTRACT.SUM is like a fancy "DIR" for EXTRACT/MERGER package.
To avoid problems, leave this file alone.
Extract will recognise three types of headers in the news items it
scans: BEGIN / END pairs are the first category: They are the surest to
determine. Type two is 'CUT / blank' category: These messages contain
the word 'cut' at the beginning of the message, and sometimes nothing at
the bottom - it is normal to see the warning 'hit the separator' warning
for such files. Type 3 contains either a 'Section...' at the start of
the file, or nothing at all. Depending on how the file begins, it will
expect it to end with one of these three. Even if they don't, your
separator will save it from continuing onto the next file.
MERGER is the second step: It takes pieces found in the partials
directory, and puts them together whenever possible. It will ignore
others if segments are missing. Otherwise, it will create
the resulting file, and delete the partial files. Merger attempts to
produce a "clean" file with no blank lines, and no unnecessary headers.
(The list of files processed by EXTRACT are kept in EXTRACT.SUM. Normally
it shouldn't be changed, except by EXTRACT and MERGER programs.)
Finally, the resulting files, characterised by their two-letter
filetypes, are passed thru a UUdecoder which will convert the files to
their binary form, and delete the two-letter version. It is normal to
see warnings while UUDECODE runs - at least my version of it complains
but creates good files.
LAST WORDS
Extracting news into files is an inexact science: I recommend that you
report cases where EXTRACT/MERGER pair failed to process, or other
problems you may notice with the software. As with any other software,
I cannot be held liable for its behavior, but would be very grateful for
suggestions and bug reports.