Errors-To: owner-ecto@athos.rutgers.edu
Reply-To: ecto@athos.rutgers.edu
Sender: ecto@athos.rutgers.edu
From: ecto@athos.rutgers.edu
To: ecto-request@athos.rutgers.edu
Bcc: ecto-digest-outbound@athos.rutgers.edu
Subject: ecto #138


                            ecto, Number 138

                       Wednesday, 29 January 1992

                             Today's Topics:
                           *-----------------*
                               Newsgroup?
                             indy cassettes
                          To be or not to be...
                              mbox_extract

========================================================================

Date: Wed, 29 Jan 92 11:59:55 EST
From: clsi!hank@uunet.uu.net (Henry Kilmer)
Subject: Newsgroup?

Hi all!  This is my first posting although I have been following the
group for a little bit.  I had a thought about the mailing
list/newsgroup discussion that has been going on...what if the
newsgroup was moderated to keep out the flames etc that has altered
gaffa?  Just a thought :)  I can't wait for the concert in Philly!!!

Hank

When it dawns on you - what it takes from you - livin on the clouds of
pain.  There's a storm in you - you don't know what to do - just when
you think you're going insane - you're like naked in the rain. -CN


========================================================================

From: drk@leland.stanford.edu
Subject: indy cassettes
Date: Wed, 29 Jan 92 10:12:05 PST

Hi everybody.  Clayton Glad (glad@Csli.Stanford.edu) is assembling
a "monthly post of music on independantly-produced cassettes" for
rec.music.misc.  The first one just appeared -- *sans* Happy.
Ordering information is accompanied by a short description/review
of the music. Could someone (Jessica? Vickie?) send this guy the
appropriate information regarding Happy.  I'm not qualified --
still waiting for the 1st4 on CD.

<this post FDA certified 99.5% free of Blue Fuzz>    8^)

David.


========================================================================

From: kyrlidis@athena.mit.edu
Subject: To be or not to be...
Date: Wed, 29 Jan 92 13:31:15 EST

Fuzzy Blue...

Hi,

This issue came up in several posts recently. I like ecto the way it is now.
I wouldn't want it to change. Ecto can be many things:

a) A great source for Happy news and information and discussion (which is why
it was created in the first place).

b) A great source for new music information. I have made several purchases
based on ecto-recommendations and haven't regretted a single one.

c) A fuzzy blue wonderland. Where we can compile birthday lists, run silly
polls, ask silly questions we wouldn't ask anywhere else in fear of getting
flambe'-ed, and feel comfortable.

I don't think I'll run another snow-poll since some people find it annoying,
but if ecto changes and degenerates to a gaffa, I will feel very sorry.

Angelos
+=====================================+
|'My ears have parasites'-hApPy RhOdEs|
+=====================================+


========================================================================

Date: Wed, 29 Jan 92 13:40:26 -0500
From: sae@cmpsci.suffolk.edu (Sam Alan Ezust)
Subject: mbox_extract

Hi folks. Some of you may already have this program, but I thought I'd
post it anyway, for the benefit of those who missed it back in 1990.
This program will work on unix type mbox files, and it extracts messages
from it and places them into unix mbox files of whatever name you wish,
depending on what strings are found in the message header. It's useful
if you want all your ecto messages to be in one file, to go through at your
leasure while you are reading the messages which are specifically to you
immediately.

The below document contains the sourcecode which I am using now. It has
worked fine without modifications since January 1990. I am sitll using
it today, but please be aware that the reference to "I've been using it
for 11 months" should really be "I've ben using it for 3 years".

mbox_extract version 3.0     16 January 1990
By S. Alan Ezust   sae@cmpsci.suffolk.edu

This program is public domain, and may be distributed freely. No guarantees
come with this program, and if any data is lost as a result of its use, I can
not be held responsible. However, I have been using it for about 11 months and
have never lost any data as a result of its use, and to my knowledge, there
are NO bugs in this program.

If you make changes to this program, please write in the sourcecode comments
explaining what they are, and if you distribute a modified version, I'd
appreciate you sending it along to me as well.

----------------------------------------------------------------------------

How the program works:

mbox_extract {infile outfile1}

This program takes as input two filenames. The first is read in as a
unix-format mbox file, and the other file is opened for output. If this
file already exists, it will be opened for APPENDING.

When the program is finished, infile will remain UNCHANGED.

mbox_extract also uses an auxillary file called me3.dat, which
must be in the current working directory. This file must be a standard text
file where each line contains a 'search string', followed by a colon (:),
immediately followed by a unix filename which corresponds to that string.

the program works as follows: It searches the headers of each piece of mail
for occurrences of one of the search strings. If the header of a piece of
mail contains one of these search strings, that entire piece of mail is
appended to the file with that corresponding filename.
If no line in the header contains one of these strings, the piece of mail
is appended to <outfile1>.

How to use it:

I have the following lines in my .login file:
---------------
| rm .temp
| mv /usr/spool/mail/depeche .temp
| mbox_extract .temp /usr/spool/mail/depeche
---------------

My system mailbox is /usr/spool/mail/depeche, and I wish to take
all bulk-mailed pieces of mail and place them into mbox_l. My
me2.dat file looks like this:

---------------
4AD-L@:4ad_l
cure@euclid.enet.dec.com:mbox_l
nm-list-sender:mbox_l
other-voices@:mbox_l
---------------

This way, any nm-list or other-voices messages get appended to my
mbox_l file, and any message from the 4ad-list gets appended to a different
file called 4ad_l. Any other piece of mail gets written to
/usr/spool/mail/depeche, which is my system mailbox. .temp will have the
old contents of my spool file, which I can look at in case I am wondering
how the original file looked before the extraction was done.

If a particular message header contains more than one string from me3.dat,
the message gets appended to only one of the output files - usually
the one corresponding to the first search string found. It will
not be sent to TWO different files, ever.

------------------------------------------------------------------------
Caution:

In general, the "system mailbox" can be changed by the system at any instant
in time. Pieces of mail get placed into them at unpredictable times. I was
told by some unix-gurus that to move and create new system mailbox files
CAN be dangerous if, by a coincidence, a new piece of mail arrives while
this program is running. If this happens, two processes are appending to it
at the same time, and data can be corrupted.

The running time of this program on a spool file is never much more than a
fraction of a second on this sun/4. I am willing to take the risk of such
an event happening on MY spool file, but it has NEVER happened in the
11 months I have used it. I just wanted you to realize that there is that
miniscule chance each time you run it in the way I have suggested, so that if
something weird happens to your spool file, this is probably the reason.

I could put an flock() on the output files, but this is not a standard
unix call, and is only supported by recent versions of BSD. This I believe
should solve this problem, but I haven't had time to test the program with
this added feature.

questions? bug reports? suggestions? send mail to me!

----------------------------cut here-----------------------------
#define VERSION 3.01
#define DATE "91/1/18"

/* me3.c - mbox_extract by S. Alan Ezust */

#include <stdio.h>
#include <strings.h>
#include <ctype.h>
#include <sys/types.h>
#include <sys/stat.h>

#define MAXDAT 10
#define MAXHEAD 40
#define LINELEN 200

char match[MAXDAT][LINELEN];
int last, total, extracted[MAXDAT], newcount, numfiles;

/* This program requires a me3.dat file of the following format:

string:filename

where the strings should be unique but the filenames need not be.

The first argument to this program is the input file, which is a
mbox-format file.

The program scans through the header of each message in the input file
for occurrances of any of the strings in me3.dat. If a string is found
in a message, that message is appended to the file with the corresponding
filename specified in me3.dat.

Recommended usage - if you want all your bulk-mailed mail to be put into
smbox, and everything else left in your spool file, you would put this
in your .login file:

mv /usr/spool/mail/depeche temp
mbox_extract temp /usr/spool/mail/depeche

and place distinctive strings indicating a message is bulk-mailed into
~/me3.dat.

here is an example of my me3.dat file:

4AD-L@:4ad_l
cure@euclid.enet.dec.com:ov_l
nm-list-sender:nm_l
other-voices@:ov_l
cloud-zero@:lpd_l

*/

FILE *datfile, *ifile, *ofile[MAXDAT];
char ofname[MAXDAT][50];
int findex[MAXDAT];

void strtoupper(S)
  char *S; {
  while (*S) {
    if (islower (*S)) *S -= 32;
    S++;
  }

int strpos(s,c)
  char *s;
  int c; {
  int i=0;
  while (*s) {
    if (*s==c) return(i);
    i++; s++;
  }
  return(-1);
  }

void openfiles(argc,argv)
  char **argv; int argc; {
  int i,j,l;
  char inpline[LINELEN];
  char tfname[MAXDAT][LINELEN];

  if (argc != 3) {
     printf ("Error : Incorrect number of arguments.\n");
     exit(0); }
  if ((ifile=fopen(argv[1],"r")) == NULL) {
    printf ("Error : Can't open inputfile: %s\n",argv[1]);
    exit(0); }

  if ((datfile=fopen("me3.dat", "r")) == NULL) {
    printf ("Error : can't open file me3.dat\n");
    exit(0); }
  i=0;
  while ((fgets(inpline,LINELEN,datfile))!=NULL) {
    l = strpos (inpline, ':');
    if (l == -1) {
       printf ("Error: No ':' in datafile line :");
       puts(inpline);
       exit(0);
    }

    strcpy(tfname[i],(unsigned)(inpline+l+1));
    strncpy (match[i],inpline, l);
    strtoupper(match[i]);
    i++;
    if (i > MAXDAT) {
       printf ("Error: too many lines in me3.dat. Max=%d\n",MAXDAT);
       exit(0); }
  }

  last=i;
  for (i=0 ; i< last ; i++) {
     l = strlen(tfname[i]);
     tfname[i][l-1]='\0';
/*     printf ("match: %s\n",match[i]);
       printf ("fname: %s\n",tfname[i]); */
  }

  fclose(datfile);

  numfiles=findex[0]=1;
  strcpy(ofname[0],argv[2]);
  strcpy(ofname[1],tfname[0]);

  for (i=1; i< last; i++) {
    findex[i]=-1;
    for (j=0; j<i; j++)
      if (strcmp(tfname[i], tfname[j])==0)
        findex[i]=findex[j];
    if(findex[i]==-1)  {
       findex[i]=++numfiles;
       strcpy(ofname[numfiles], tfname[i]); }
  }
  numfiles++;

/*

  for (i=0; i<last; i++)
     printf ("match: %s  file: %s   %d\n",match[i],tfname[i],findex[i]);
  for (i=1; i<numfiles; i++) {
    printf ("ofname %d:",i);
    puts(ofname[i]); }
*/

  for (i=0; i<numfiles; i++){
    ofile[i]=fopen(ofname[i],"a");
/*    printf ("output file %d: %s\n",i,ofname[i]); */
    chmod(ofname[i],384); }
}

void closefiles() {
   int i;
   fclose (ifile);
   for (i=0; i <numfiles; i++)
      fclose(ofile[i]);

}

int isbeginning (s)
  char *s; {
  if (strncmp ("From ",s,5)==0) return(1);
   else return(0);
  }

int isextract (s)
  char *s; {
  char cnv[LINELEN];
  int i,j,l,m;
  m = strlen(s);
  strcpy(cnv,s);
  strtoupper(cnv);
  for (i=0; i<last; i++) {
    l = strlen(match[i]);
    if (l < m) for (j=0; j < (m-l); j++)
      if ( strncmp (cnv+j, match[i], l) ==0 )
	return(findex[i]);
  }
  return(0);
  }

int message(topline)
  char *topline; {
  int i,j,filenum, done, isnew;
  char buffer[MAXHEAD][LINELEN], lastline[LINELEN];
  strcpy (buffer[0],topline);
  done = 0;
  isnew = i = 1;
  filenum = isextract(topline);
  while (done==0) {
    if (fgets(buffer[i],LINELEN,ifile)==NULL)
      done=3;
    else {
      if (strncmp(buffer[i], "Status:", 7)==0) isnew=0;
         /* new messages do not have the Status: field in the header */
      if ((isbeginning(buffer[i])) && (strcmp (buffer[i-1],"\n")==0)) {
	 /* the start of a new message is indicated by a blank line
	    followed by a "From " line */
	strcpy(topline, buffer[i]);
	   /* copy first line of next msg into topline */
        i--; /* we don't want this as the last line of the current msg */
        done=1;}
      if (filenum == 0) filenum=isextract(buffer[i]);
      if (strcmp(buffer[i],"\n")==0) done=2; /* are there more header lines? */
      if (++i>=MAXHEAD) done=2; } /* end of else */ /* are we out of memory? */
    } /* end of while */

  /* Done indications
    1 means that the entire message has been read (this shouldn't happen here)
    2 means that we have finished reading header lines
    3 means we have reached the end of file (this also shouldn't happen)
    */

  extracted[filenum]++;
  for (j=0; j<i; j++) fprintf(ofile[filenum],"%s",buffer[j]);
    /* send out the buffer, which should contain only the header */
  strcpy (lastline, buffer[i-1]);
  while (done==2) {
    if (fgets(topline,LINELEN,ifile)==NULL) done=3;
    else {
      if ((isbeginning(topline)) && (strcmp (lastline,"\n")==0))
        done=1;
      else fprintf(ofile[filenum],"%s",topline);
      }
    strcpy (lastline, topline);
    }

/* once again, done has the same meaning, but the other cases will happen
   here */

  newcount += isnew;
  if (done==3) return(0);
  else return(1);
  }

main(argc, argv)
  int argc;
  char **argv; {
  int t;
  char topline[LINELEN];
  printf ("Mbox extract %3.1f  by S. Alan Ezust (depeche@cs.mcgill.ca) 1990\n",VERSION);
  printf ("usage : mbox_extract infile outfile\n");
  openfiles(argc, argv);
  newcount=total=0;
  for (t=0; t < MAXDAT; t++) extracted[t]=0;
  if (fgets(topline,LINELEN,ifile)==NULL){
     printf ("Error : input file is empty.\n");
     exit(0); }
  do {t=message(topline); total++; } while (t);
  closefiles();
  printf ("%d messages in inputfile: %s, ",total, argv[1]);
  printf ("%d new\n ",newcount);
  printf ("Count of messages extracted: \n");
  for (t=0; t <numfiles; t++)
    printf ("sent %d msgs to file #%d: %s\n",extracted[t],t,ofname[t]);
  }

========================================================================

To join ecto, please send electronic mail to the following address:
	ecto-request@athos.rutgers.edu

To have your thoughts included in the next issue, send mail to:
	ecto@athos.rutgers.edu

The ecto archives are on hardees.rutgers.edu in ~ftp/pub/hr. There is
a README file explaining what is where. Feel free to send me (or leave
in the incoming directory, just let me know) things you'd like to have
added.

To subscribe to "Ecto", the printed fanzine, send $8 to:
	Ecto
	PO Box 11291
	New Brunswick, NJ 08906 
Ecto is issued 8 times/year, and will include photos and as much material 
from non-net members as we can get! Donations above the subscription cost 
are welcomed - all money goes to bringing you better issues!

	     I've been told I'm too far from humble to even
	     have it in quotes, so, just "your moderator"
		    -- jessica (jessica@athos.rutgers.edu)