#!/usr/bin/gawk -f
#
# Use: dtdiff new_file.html old_file.html > diff_file.html
#
# Program dtdiff compares two -style resource record files, and
# produces a 'diff' report in HTML format. The comparison is done in
# memory, so it is not neccesary that either input file be sorted.
# D.Wells, NRAO-CV, 13Feb94.
# --------------------------------------------------------------------
END {
for (i = 1; i <=2; i++) { # read "new_file", then "old_file":
if((val = read_header(ARGV[i])) <= 0) {
printf("ABORT: read_header(%s) returned %d.\n", ARGV[i], val);
exit(13);
}
titles[i] = title;
links[i] = link;
j = 0;
do {
if ((val = read_record(ARGV[i])) <= 0) {
if (val < 0) {
printf("ABORT: read_record(%s) returned %d for j=%d.\n",
ARGV[i], val, j);
exit(13);
}
break;
}
j++;
records[i,j] = rr;
urls[i,j] = url;
} while (1);
num_recs[i] = j;
close(ARGV[i]);
#
print "file=" ARGV[i];
print num_recs[i] " records";
print "=" titles[i];
print "<LINK>=" links[i];
print "URL[3]=" urls[i,3];
print "records[3]=" records[i,3];
print " ";
}
#
# At this point we have the two sets of records and URLs in memory.
# First we will compare the lists and delete identical records.
# Then we will produce several listings:
#
# Delete identical (same URL, same text) resources:
num_delete = 0;
for (i = 1; i <= num_recs[1]; i++) {
for (j = 1; j <= num_recs[2]; j++) {
if ((i == 3) && (j ==3)) print urls[1,i], urls[2,j];
if (urls[1,i] == urls[2,j]) {
print "equal: " i, j;
for (m = i; m < num_records[1]; m++) {
records[1,m] = records[1,m+1];
urls[1,m] = urls[1,m+1];
}
num_records[1]--;
for (m = j; m < num_records[2]; m++) {
records[2,m] = records[2,m+1];
urls[2,m] = urls[2,m+1];
}
num_records[2]--;
num_delete++;
}
}
}
#
# NOTE! above logic has subscript wrong by 1 as soon as delete is done.
#
printf ("%4d identical resource records deleted.\n", num_delete);
exit(1);
# Merge cases of identical URLs but different text:
num_merge = 0;
for (l = 1; l < n; l++) {
if ((key[l] == key[l+1]) && (resource[l] != resource[l+1])) {
# append resource[l+1] to [l] with <DD> to make indentation:
resource[l] = resource[l] "<DD>||" resource[l+1] "</DD>||";
for (m = l+1; m < n; m++) { # move rest of list up one place:
key[m] = key[m+1];
resource[m] = resource[m+1];
} # resource[l+1] has been deleted:
num_merge++;
n--; # decrements count of resource[];
}
}
printf ("%4d similar resource records merged. (n=%d)\n",
num_merge, n) | "cat 1>&2";
printf ("%4d resource records edited:\n", num_edit) | "cat 1>&2";
for (i in hit) printf("\t%4d cases of |%s|\n", hit[i], i) | "cat 1>&2";
print "</DL></BODY></HTML>";
printf ("%4d resource records written.\n", n) | "cat 1>&2";
}
# ------------------------------------------------------------------
function read_header(filename)
{
# This function reads the <HEAD> lines from a specified input
# file, and returns the <TITLE> and <LINK> strings as global variables.
#
head = "";
do {
if ((ret_val = getline <filename) <= 0) return (ret_val);
if (NF > 0) head = head $0; else break;
} while (1);
u1 = match(head, "<[Tt][Ii][Tt][Ll][Ee]>") + RLENGTH;
ul = match(substr(head, u1), "</[Tt][Ii][Tt][Ll][Ee]>") - 1;
title = (u1 > 0) ? substr(head, u1, ul) : "";
u1 = match(head, "<[Ll][Ii][Nn][Kk] .* [Hh][Rr][Ee][Ff]=\"") + RLENGTH;
ul = match(substr(head, u1), "\">") - 1;
link = (u1 > 0) ? substr(head, u1, ul) : "";
return(1);
}
# ------------------------------------------------------------------
function read_record(filename)
{
# This function reads the next resource record from a specified input
# file, and returns the URL and the text as global variables.
# Canonical resource records begin with blank line followed by <DT>.
# Therefore, we skip all input lines until we get a left-justified <DT>,
# then we concatenate the resource record lines, up to the next blank
# line:
do {
if ((ret_val = getline <filename) <= 0) return(ret_val);
if ($0 ~ /^[<][Dd][Tt][>]/) break; # detect <DT> line
} while (1);
rr = $0 "||";
do {
if ((ret_val = getline <filename) <= 0) return(ret_val);
if (NF > 0) rr = rr $0 "||"; else break;
} while (1);
# The lines of this resource record are now concatenated into one string,
# separated by the code '||'. Now we will extract the URL:
u1 = match(rr, "[Hh][Rr][Ee][Ff]=\"") + RLENGTH; # find start of first URL
ul = index(substr(rr, u1), "\"") - 1; # get its length
url = (u1 > 0) ? substr(rr, u1, ul) : "";
return(1);
}
# ------------------------------------------------------------------
function print_record(resource, filename)
{
while (length(resource) > 0) {
p = index(resource, "||"); # find end of current line,
printf("%s\n", substr(resource, 1, p-1)) >filename; # print it,
resource = substr(resource, p+2); # and delete it.
}
printf("\n");
}