Fix quoting problem which has appeared with new version of cvs2cl.
Update CGI's to know about new site.
Use static menu instead of SSI one.
Urg - now the image is generated using SSI the things that include the include manually won't work. We can kludge these CGI's... but what about the viewcvs stuff? This needs a rethink.
A few fixes: - Need to reset the section marker for each new module. - Should escape HTML in the log messges too :-)
Final round of XHTML/tidy'ing CGI scripts. They both produce XHTML that validates correctly. The code is a little better, but a touch more messy in a few places.
Commit new website. The old site is tagged, so this won't change the live site... but it does move HEAD on to the new site. Too many changes to list really. General points are: - Moved to a XHTML CSS compliant site. - Reorganised the site into a more multi-project based look. - Removed a lot of cruft. Still to do: - Fix all the zillions of bugs stopping the whole site from validating :-) - Tidy up the HTML in terms of layout and indentation. Thanks to AJ for his help this weekend in doing this.
Sort out the revision matching regex. Merge the two regex's, and make it work with revisions that have more than one tag associated with them.
Update scripts that use cvs2cl. Now using cvs2cl 2.49, still with some custom patches (but less than before). Diff included here. --- cvs2cl.bak Sat Sep 27 16:57:59 2003 +++ cvs2cl Sat Sep 27 17:10:55 2003 @@ -1036,7 +1036,8 @@ elsif ($No_Wrap && !$Summary) { $msg = &preprocess_msg_text ($msg); - $files = wrap ("\t", "\t", "* $files"); +# tdb: no-wrap shouldn't wrap ANYTHING! + #$files = wrap ("\t", "\t", "* $files"); $msg =~ s/\n(.+)/\n\t$1/g; unless ($After_Header eq " ") { $msg =~ s/^(.+)/\t$1/g; @@ -1328,7 +1329,8 @@ # there are multiple qunks in the first place! if ((scalar (@qunkrefs)) > 1) { foreach my $tag (keys (%non_unanimous_tags)) { - my $everyone_has_this_tag = 1; +# tdb: we don't want to have unanimous tags + my $everyone_has_this_tag = 0; foreach my $qunkref (@qunkrefs) { if ((! (defined ($$qunkref{'tags'}))) or (! (grep ($_ eq $tag, @{$$qunkref{'tags'}})))) { @@ -1521,7 +1523,8 @@ unless ( exists $fileinfo_printed{$b} ) { if ($fbegun) { - $beauty .= ", "; +# tdb: hacked to put a newline after each one + $beauty .= ",\n"; } else { $fbegun = 1; }
Fix a typo I noticed on the website.
Fix minor perl bug.
Merge in changes from branch created for new server. The old server is now history... so we can make the new server use HEAD.
Nothing changed, just wanted to log this relevant diff... This is a diff against the latest cvs2cl, version 2.46. It makes it work how we require for cvslog. There's no new features, it's just a new diff covering all the previous alterations on 2.38. Oh, and it's a unified diff, which is much more readable :-) --- cvs2cl.orig Mon May 12 14:12:50 2003 +++ cvs2cl Mon May 12 14:13:14 2003 @@ -879,7 +879,8 @@ elsif ($No_Wrap) { $msg = &preprocess_msg_text ($msg); - $files = wrap ("\t", " ", "$files"); +# tdb: no-wrap shouldn't wrap ANYTHING! + #$files = wrap ("\t", " ", "$files"); $msg =~ s/\n(.*)/\n\t$1/g; unless ($After_Header eq " ") { $msg =~ s/^(.*)/\t$1/g; @@ -1042,7 +1043,8 @@ # Keep track of whether all the files in this commit were in the # same directory, and memorize it if so. We can make the output a # little more compact by mentioning the directory only once. - if ((scalar (@qunkrefs)) > 1) +# tdb: hacked to turn off the commondir feature + if ((scalar (@qunkrefs)) < 0) { if (! (defined ($common_dir))) { @@ -1086,7 +1088,8 @@ # there are multiple qunks in the first place! if ((scalar (@qunkrefs)) > 1) { foreach my $tag (keys (%non_unanimous_tags)) { - my $everyone_has_this_tag = 1; +# tdb: we don't want to have unanimous tags + my $everyone_has_this_tag = 0; foreach my $qunkref (@qunkrefs) { if ((! (defined ($$qunkref{'tags'}))) or (! (grep ($_ eq $tag, @{$$qunkref{'tags'}})))) { @@ -1237,7 +1240,8 @@ next if (defined ($$qunkref{'printed'})); # skip if already printed if ($fbegun) { - $beauty .= ", "; +# tdb: hacked to put a newline after each one + $beauty .= ",\n"; } else { $fbegun = 1; @@ -1286,7 +1290,8 @@ # todo: still have to take care of branch_roots? - $beauty = "* $beauty:"; +# tdb: hacked to remove the prefixed * + $beauty = "$beauty:"; return $beauty; }
Nothing changed, I just want to log a relevant change... This is the latest modification to cvs2cl, revision 2.38. This is an old and outdatted version of cvs2cl, but it's what we're using right now. I've got rid of unanimous tags in this change. The patch given below is a complete diff from 2.38. 777c777,778 < $files = wrap ("\t", " ", "$files"); --- > # tdb: no-wrap shouldn't wrap ANYTHING! > #$files = wrap ("\t", " ", "$files"); 934c935,936 < if ((scalar (@qunkrefs)) > 1) --- > # tdb: hacked to turn off the commondir feature > if ((scalar (@qunkrefs)) < 0) 978c980,981 < my $everyone_has_this_tag = 1; --- > # tdb: we don't want to have unanimous tags > my $everyone_has_this_tag = 0; 1126c1129,1130 < $beauty .= ", "; --- > # tdb: hacked to put a newline after each one > $beauty .= ",\n"; 1175c1179,1180 < $beauty = "* $beauty:"; --- > # tdb: hacked to remove the prefixed * > $beauty = "$beauty:";
A fix for version numbers with more than two parts - ie. branches (1.2.1.2)
It seems showing branches isn't actually working.
Added section to deal with users who don't have an e-mail address. Without this the line doesn't get any special treatment, which doesn't look pretty!
Fixed the link to diffs when there was a tag associated with the commit.
Added option to show commits since a given date: cvslog.cgi?period=since&days=2001/12/09 It should be noted that the defintion of "since" in this case does NOT include the day given.
Bug which meant the cvs update wasn't correctly executed before generating the cvslog output.
Major changes: - Now supports period=date&date=yyyy/mm/dd query strings. - The files/logs section of each commit has been broken into two seperate bits. - The filenames are linked to their relevant viewcvs page. - The revisions are linked to their diffs, allowing one to see the changes a certain commit has introduced. - The log message itself has been unwrapped, for neatness. Bugs: - fails to know about "deleted" files. They simply show up as a new revision. This is a fault in the underlying cvs2cl script. Notes: - This uses a modified version of cvs2cl. The diff is given here to avoid any confusion as to what this script expects in the future. 777c777,778 < $files = wrap ("\t", " ", "$files"); --- > # tdb1: no-wrap shouldn't wrap ANYTHING! > #$files = wrap ("\t", " ", "$files"); 934c935,936 < if ((scalar (@qunkrefs)) > 1) --- > # tdb1: hacked to turn off the commondir feature > if ((scalar (@qunkrefs)) < 0) 1126c1128,1129 < $beauty .= ", "; --- > # tdb1: hacked to put a newline after each one > $beauty .= ",\n"; 1175c1178,1179 < $beauty = "* $beauty:"; --- > # tdb1: hacked to remove the prefixed * > $beauty = "$beauty:";
Looks a bit neater I think.
Some fixes to sort the colours of links.
Added the title and bottom in the correct fashion. This is inline with the i-scream layout and style.
Added the left bar to the commit logs.
Improved the resulting HTML, making it vaguely readable.
Slight alteration to ensure that it only one module is checked by fullcvsupdate.sh if it is specified in the query string.
Now accepts a module in the query string, should that be preferable.
Urg, yellow. No...
Some changes to layout, incorporated from a modified version by Paul ;-)
Uses the new script that fully checks out and updates a cvs local copy.
Now been "updated" to support some more funky features :) Firstly the output is grouped by module, making it clearer to see what commits belong where. Secondly, query strings allow you to specify the period, rather than just the current day that the old one did. This are: Current Day: period=today Current Week: period=thisweek Last n Days: period=days&days=n There are also subtle location changes to suit the location on the new webserver rather than raptor.
Best update the checked out copy before doing the cvs2cl.
Now use viewcvs instead.
Changed the title on the page.
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, select a symbolic revision name using the selection box, or choose 'Use Text Field' and enter a numeric revision.