Having both the summary and whiteboard columns enabled causes them to fight for space. This is because both columns are set to width=100%. This is silly, obviously. The solution is to make the tables width="100%", and let *all* the columns autosize themselves. The routine causing this mess is on line 498 of buglist.cgi: http://lxr.mozilla.org/mozilla/source/webtools/bugzilla/buglist.cgi#498 Lines 500 to 504 should, IMHO, be removed. The following should be inserted as a replacement: $tablestart .= "<TH align=left valign=top>"; Further, line 493 should be adjusted to read my $tablestart = "<TABLE WIDTH="100%" CELLSPACING=0 CELLPADDING=2> ...(note the addition of the width attribute). While I'm at it, another bug in buglist.cgi is on line 582. Because of line 191, if serverpush is disabled, line 582 results in the HTTP header being printed out twice, and the Set-Cookie being added as part of the document and not the header. That would be harder to solve than the main bug reported here, though.

This kind of bug report is really hard to read. Referring to line numbers of source files is a bad idea, because you never know if someone is going to edit the file before your bug report is looked at. (Which in fact has happened here.) Could you please do me a favor, and submit a context diff that makes the changes you want? I don't care if you actually test it or anything. Just quickly make a copy of the sourcefile, edit it to appear about what you think would be better, run "diff -c" between the original and your new version, and paste the results in here. Thanks!

Oops! Sorry about that. The code in question is the part that says 'foreach my $c (@collist) {' followed by some '$tablestart .= "<TH WIDTH=100% valign=left>";'. (It's the second such foreach in the file at this time.) The entire 'if ($::needquote{$c}) {' block including its 'else' part should IMHO be replaced by: $tablestart .= "<TH align=left valign=top>"; Further, the only 'my $tablestart' line in the file should be changed to read: my $tablestart = "<TABLE WIDTH="100%" CELLSPACING=0 CELLPADDING=2> ...(note the addition of the width attribute -- it is needed because the first part of the change removes the WIDTH=100% on the columns, so instead the _table_ must be so sized). The above change would fix the racing caused by both 'summary' and 'status' columns being given 100%. The second part of the error report refers to the 'my $toolong = 0;' line. Because of the 'print "Content-type: text/html\n\n";' line after the 'quietly_check_login();' statement, if serverpush is disabled, line the 'my $toolong = 0;' line results in the HTTP header being printed out twice, and the Set-Cookie being added as part of the document and not the header. This is clearly visible in IE. It would be harder to solve than the main bug reported here. I am afraid I cannot do a diff because the package that would let me do that is on my main machine, and the only decent web browser on my main machine is Lynx, and, well, see bug 3078.

Terry: this is a simple fix, and how to fix it has been included, any chance of it being fixed? Pleeeeeease..... :-)

FWIW, the HTML spec says it's not legal to specify a percentage width on a table cell. It's perfectly valid on the table itself, but cell widths must be specified in pixels or left out to be autosized. Netscape of course violates this and allows it, but it is against the HTML spec. (And every HTML checker I've seen flags it as an error). I was just going to do this and supply a diff, since a diff was asked for here, and was never provided, but in doing so, I discovered this has already been fixed at some point or another. This bug is fixed. The code in the current version of buglist.cgi matches what is suggested here. Whoever did it forgot to change the status on the bug.

dave: It isn't fixed, sorry. The part I am referring to is currently line 959 of the file buglist.cgi: http://lxr.mozilla.org/mozilla/source/webtools/bugzilla/buglist.cgi#959 ...and the changes I suggest still need doing. Namely, remove these lines: 958 my $h = ""; 959 if ($::needquote{$c}) { 960 $h .= "<TH WIDTH=100%>"; 961 } else { 962 $h .= "<TH>"; 963 } ...and replace them with just: 958 my $h = "<TH>"; That has not yet been done in version 1.96 of this file, which I believe is the latest one.

ah, well, what you're quoting now is a different section of code that what was quoted at the beginning of this bug report. What was quoted at the beginning here is already fixed. The table itself already contains the WIDTH="100%". I made the adjustment you mentioned, one diff coming right up.

tara@tequilarista.org is the new owner of Bugzilla and Bonsai. (For details, see my posting in netscape.public.mozilla.webtools, news://news.mozilla.org/38F5D90D.F40E8C1A%40geocast.com .)

Chris: This is the one I meant. Patch is attached, although it may have bitrotted since it is now old.

checked into trunk. we should take a look at this on landfill with various columns and widths to make sure that it looks okay.

Sorry for the spam, but I needed to be able to query for all of these correctly.

Can't verify this as whiteboard isn't turned on on Landfill. Gerv

Moving closed bugs to Bugzilla product