Update of /cvsroot/mahogany/M/src/util
In directory usw-pr-cvs1:/tmp/cvs-serv28930/src/util
Modified Files:
matchurl.cpp
Log Message:
tweaks to the URL detection: don't include trailing ! and ? in it; try to
automatically detect the URL end if it ends with an extension; disabled CGI parameters
heuristic text
Index: matchurl.cpp
===================================================================
RCS file: /cvsroot/mahogany/M/src/util/matchurl.cpp,v
retrieving revision 1.18
retrieving revision 1.19
diff -b -u -2 -r1.18 -r1.19
--- matchurl.cpp 8 Sep 2002 19:23:12 -0000 1.18
+++ matchurl.cpp 19 Oct 2002 21:44:01 -0000 1.19
@@ -555,4 +555,7 @@
}
+ // this test doesn't reduce the number of false positives all that
+ // much, finally, but does result in non recognizing some wrapped URLs
+#if 0
// even with all the checks below we still get too many false
// positives so consider that only "long" URLs are wrapped where long
@@ -564,4 +567,36 @@
break;
}
+#endif // 0
+
+ // heuristic text for end of URL detection: consider that if it ends
+ // with an extension (i.e. ".xyz") then it's the end of the URL
+ if ( p - start > 5 && p[-4] == '.' &&
+ isalnum(p[-3]) && isalnum(p[-2]) && isalnum(p[-1]) )
+ {
+ // some special cases where the URL could still be wrapped: html,
+ // jpeg and tiff are common 4 letter extension
+ static const char *longExtensions[] =
+ {
+ "html", "jpeg", "tiff"
+ };
+
+ size_t n;
+ for ( n = 0; n < WXSIZEOF(longExtensions); n++ )
+ {
+ const char * const ext = longExtensions[n];
+ if ( strncmp(p - 3, ext, 3) == 0 && p[2] == ext[3] )
+ {
+ // looks like a long extension got wrapped
+ break;
+ }
+ }
+
+ if ( n == WXSIZEOF(longExtensions) )
+ {
+ // it doesn't look that extension is continued on the next line,
+ // consider this to be the end of the URL
+ break;
+ }
+ }
// Check that the beginning of next line is not the start of
@@ -580,5 +615,5 @@
// it might be a wrapped URL but it might be not: it seems like we
- // get way too many false positives if we suppose that it's already
+ // get way too many false positives if we suppose that it's always
// the case... so restrict the wrapped URLs detection to the case
// when they occur at the beginning of the line, possibly after some
@@ -596,7 +631,7 @@
break;
- // it did occur at the start (or after '<'), suppose the URL is wrapped and
so
- // continue on the next line (no need to test the first character,
- // it had been already done above)
+ // it did occur at the start (or after '<'), suppose the URL is
+ // wrapped and so continue on the next line (no need to test the first
+ // character, it had been already done above)
p += 3;
}
@@ -604,5 +639,5 @@
// truncate any punctuation at the end
- while ( strchr(".:,;)", *(p - 1)) )
+ while ( strchr(".:,;)!?", *(p - 1)) )
p--;
-------------------------------------------------------
This sf.net email is sponsored by:
Access Your PC Securely with GoToMyPC. Try Free Now
https://www.gotomypc.com/s/OSND/DD
_______________________________________________
Mahogany-cvsupdates mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mahogany-cvsupdates
