Comments on: Google Groups KillFile 3.4.3 Released

By: Giraffe

Giraffe — Fri, 16 May 2008 18:45:52 +0000

There are certain topic titles that refuse to be killed. One example is here:

http://groups.google.com/group/sci.math.symbolic/browse_thread/thread/4a159a9c58e3af16#

By: Mike

Mike — Sat, 26 Apr 2008 12:00:17 +0000

I am unable to kill this user
Bill Bonde ( ‘the oblique allusion in lieu of the frontal attack’ )
in this or any other thread
http://tinyurl.com/6cjtug
I tried manually adding the user to the string, but had no luck.
Unfortunately, this poster is the main reason I tried the string.

By: Mike

Mike — Fri, 25 Apr 2008 11:58:27 +0000

For some reason I can’t killfile Bill Bonde in this (or any other) thread:
http://groups.google.com/group/alt.fan.cecil-adams/browse_thread/thread/ff45bfd450705bb8/b7e64282432214fe?hl=en#b7e64282432214fe

By: Mark Widmer

Mark Widmer — Wed, 19 Mar 2008 17:13:13 +0000

Weird that it works for you and not me. Well, rather than post my whole killfile, I have modified it to be just
—–%FFtt
so that “tt” is the one and only entry, and still have the problem of blocking Tim Wescott’s posts.

FYI, I am using:
Windows XP Professional
Firefox 2.0.0.12
Greasemonkey 0.7.20080121.0
and re-clicked on the link to Google Groups KillFile 3.4.3, just to be sure.

Mark

My current killfile, in which I have now removed %FFtt in order to see Tim’s posts, is
—–%FFSohaib%FFRadium%FFcheapNdisco…@yahoo.cn%FFPhil%20Allison%FFSkybuck%20Flying%FFTopic%3AM%60I%205.*%FFTopic%3AM%27I-5.*%FFTopic%3AM%2CI.5.*%FFTopic%3AM-I%205.*%FFTopic%3AM.I%2C5.*%FFTopic%3AM%20I.5.*%FFTopic%3AM.I%275.*%FFTopic%3AMI5.*%FFglobwholesale%FFcheapforwholes%5C.%5C.%5C.@126%5C.com%FFcheapforwholes%5C.%5C.%5C.@yahoo%5C.cn%FFcheapforwholesal%5C.%5C.%5C.@yahoo%5C.cn

By: Damian

Damian — Wed, 19 Mar 2008 15:41:29 +0000

I see… okay I’ll put this into a release and push it tonight. I barely know what I’m doing myself seeing as I’m a java programmer :)

By: Tim

Tim — Wed, 19 Mar 2008 15:33:21 +0000

Yes. The screen was getting corrupted below a certain point.

For a while I was hacking my regexs so that there was never more than one matching but it was getting impossible to maintain and when it was a new post that broke it rather than a new regex it was hard to track down what needed fixing.

IIRC this was easy to reproduce, just kill the author and the thread of the same post.

I think the problem was lines like this one:
parentTable.removeChild(parentRow);
that would get called multiple times for the same candidate.

But it was a bit of a guess. I don’t know javascript or, in particular, what document.evaluate() actually does. I’m a C and C++ programmer really.

Tim.

By: Damian

Damian — Wed, 19 Mar 2008 14:09:51 +0000

Yeah I got it, makes sense. Was this actually causing problems ?

By: Tim

Tim — Wed, 19 Mar 2008 14:05:43 +0000

Thanks!

Something has gone wrong with the HTML formatting of that code patch.

Hopefully it’s obvious that I’ve simply added ‘!blocked &&’ to the for() loops and moved the initialization of blocked and not done anything really weird like j=0 in the condition of the loop.

Tim.

By: Damian

Damian — Wed, 19 Mar 2008 13:16:06 +0000

Comments need to be approved before they show up Tim, and I just approved it over on the 3.4.2 post. Thanks for the update and I’ll add it for the next release.

By: Tim

Tim — Wed, 19 Mar 2008 13:12:38 +0000

There is a problem when multiple entries kill the same post.

I’ve made the following changes that seem to fix it for me:

— googlegroupskillfile.user.js.orig 2008-03-19 12:45:16.791559000 +0000
+++ googlegroupskillfile.user.js 2008-03-19 12:45:16.081662000 +0000
@@ -179,8 +179,7 @@
var user = cand.textContent;
var blocked = false;

- for (var j=0; j - blocked = false;
+ for (var j=0; !blocked && j

topicRE.compile(“^” + list[j]);
if (topicRE.test(topic) || topicRE.test(user)) {
@@ -275,9 +274,9 @@
for (var i=0; i var cand = candidates.snapshotItem(i);
var user = cand.textContent;
+ blocked=false;

- for (var j=0; j - blocked = false;
+ for (var j=0; !blocked && j= 0) {
wrappingTable = document.evaluate(“ancestor::table[2]“, cand, null, XPathResult.ORDERED_NODE_SNAPSHOT_TYPE, null);

Tim.

p.s. I’ve tried to post this several times before but it always seems to fail.