CVS log for 2002/scribe/Attic/scribe.perl

Up to [Public] / 2002 / scribe

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

migrated to github.com/w3c/scribe

... but do not allow slashes, for backwards compatibility.

Made $speakerPattern much more permissive, so that names with
non-ASCII letters will match, too, e.g., "François".

scribe.perl has many templates. The "manually" disclaimer had to be added
to each of them.

Added a statement to say that minutes are manually created, not a transcript.
See https://www.w3.org/mid/C27C433C-6E19-4B41-821F-82B8C7636FB2@w3.org

Replaced once forgotten meta-http-equiv with <meta charset=utf-8>.

Some clean-up of unused variables.

Now determines default present list before replacing <name> with
<scribe>.

Now always guesses the same scribe if there are two or more
candidates, thanks to sorting the names. (Not important in itself, but
makes regression testing possible.)

Fix error in previous check-in: error in "Meeting:" regexp.

Removed unusued function ProcessRalphURL.

Now actually does what the previous check-in comment said:
s/@@/http://example/foo/ treats the last "/" as a delimiter.

Fixes ambiguity when a "]" follows a URL. And also removes white space
after a URL if it is probably only there to delimit the URL.

Recognize https URLs.

Use lc instead of tr/A-Z/a-z/ and also use it in more places so that
more names that differ only in case are reocgnized.

Now allows slashes (or vertical bars) in the replacement of s//
commands. E.g., s/@@/http://examples.org/foo now replaces @@ with
http://examples.org/foo because that is probably what the users
expect. A final slash, however, is still a delimiter:
s/@@/http://example.org/foo/ replaces @@ with http://example.org/foo

Small improvements to regular expressions to avoid generating <p>
elements for empty lines. Small improvements to generate fewer empty
lines (thanks to using the /m switch and the ^ pattern, instead of
adding redundant \n at the start). Commented out or removed some lines
that appear not to be used. Removed the isBotLine function, which
isn't used anywhere.

Now also recognizes lines such as "<amy> Present+", i.e., without a
name at the end, as an instruction to add the speaker himself.

Removed multiline @229

Switch to HTML5

Make links to CSS and Icons absolute https ones so that scribe output
can be stored on lists.w3.org
If people prefer them to stay relative, tell me and I'll add a config
option.

treat horizontal ellipsis as continuation

Also allow the original elho theme for IRSSI (of which Bert's theme is
a variation). Allow nicknames to be preceded by @, + or % (which
indicates their status: operator, voiced or half-op).

Added patterns for the log format I'm using for my IRSSI client.

Enhancements by Arnaud Le Hors

Allowed ",", "[" and "]" in a URL. Do not allow "&" (except as part of
"&"). Removed the local $urlPattern from MakeLinks() and used the
global one instead. Not allowing "&" fixes a bug where
"<https://www.w3.org/foo>." would become a link to
"https://www.w3.org/foo&gt;."

Done in answer to
https://www.w3.org/Help/Requests/Ticket/Display.html?id=8366&results=0a0d78286e0ac30598201f9205c86b2d

Make all css links relative, per https://lists.w3.org/Archives/Team/sysreq/2013Apr/0255.html

Add patch to fill empty link text from Matt Womer.

Recognize https: URIs in Agenda:
reported 2011-05-12 by Coralie

Ensure /^https:/ lines are not attributed to a person named "https".

fix (?) problem with "use strict" vs. perl 5.10

escape the attendee list reported by Zakim as well

xss vulnerability test cases

Remove some XSS vulnerabilities; escape the values of
Chair Log Meeting Present PreviousMeeting Regrets Scribe ScribeNick Title

Put leading zeros on SV_MEETING_DAY and SV_MEETING_MONTH_NUMERIC
as I believe was intended

Make missing or empty template file a non-fatal error;
use previously set (or default) template.

XML-escape formatted actions when inserting them into the template.

add CONTINUES as synonym for CONTINUED

typo

Add an Id close to the top for ease of maintenance

Replace the text of the W3C Software Notice and License with the
current (Rev 1.11 2004-07-06) version.  Primarily because the
orignal date of the previous notice had not been quoted as a
CVS tag and had been lost.

Don't die on unparseable date errors.  These often occur when some mail
headers are pasted into the channel.

Improved the recognition of Irssi log timestamps.

Fixed ISO8601 timestamp pattern matching.

Saving a checkpoint.  LOTS of changes for converting to an improved
processing model, but not yet enabled because the new processing model
is not yet fully debugged.

s/--/&mdash;/ in <title>

Fixed bug that caused i/// to fail if the command appeared on the next line
after the desired insertion point.

P.S. Also fixed the RRSAgent_HTML_Format processing so that it no
longer causes duplicate links.

Added support for s||| substitution and i||| insertions, so that URLs can
be edited.  Also improved the processing of those editing commands.

Moved action processing into a function (ProcessActions).

Ignore "next agendum" lines.

Improved the MIT template to give instructions for finishing the minutes.

Avoid warning if new list of zakim attendees is a superset of old.
Make URLs clickable in embedded diagnostic output.

Added support for XChat_Timestamped_Log_Format.

Accommodate new RRSAgent message syntax.  Accept both formats of the
"logging" message (with and without "is"):
<RRSAgent> is logging to http://www.w3.org/2005/01/31-arch-irc
<RRSAgent> logging to http://www.w3.org/2005/02/14-arch-irc

Add a link from each action item, " [recorded in http://...#action01]",
to the location in the minutes (or log) where it was recorded, so that
it is easy to find the context of the action.  This depends on knowing
either the final URL of the minutes (preferably), or at least the log URL.
If RRSAgent is used, then normally the minutes URL will be guessed
from the log:
    <dbooth> rrsagent, draft minutes
    <RRSAgent> I have made the request to generate http://www.w3.org/2005/01/07-swcg-minutes dbooth
However, I also added the command-line option
"-minutes http://www.w3.org/2005/01/07-swcg-minutes.html" (for example)
to explicitly tell scribe.perl the URL of the minutes.
Incidentally, if only the log URL is known, then at present the action
item will only be able to point to the entire log, rather than using
a fragment identifier ("#action01") to point to the specific location
within the log, because scribe.perl is not currently able to determine
what fragment identifier is associated with each action in the log.

Also made a minor improvement to action status processing.

Prevent "to" from appearing twice when alternate ACTION format
is used: "ACTION DanC: to propose to postpone"
Changed the default to not break long ACTION lines.
No longer warn if new Regrets or Present list is superset of old list.

Added another link to scribedoc.html

Added i/where/line command to add a line at an earlier location.
Reduced the number of warnings.  No longer warn if no agenda or no regrets.
Improved the reporting of failed s/old/new/ substitutions.

Recognize Ralph's syntax for link text:
-> http://foo text
becomes:
<a href="http://foo">text</a>
to avoid having so may visible URLs.

Explicitly close STDOUT and STDERR on exit, in case some systems don't.

Remove $ from scribe.perl version info in generated minutes, to prevent it
from getting re-munged if minutes are checked into CVS.

Added -embedDiagnostics and -noEmbedDiagnostics options

Permit alternate action syntax: "ACTION dbooth: fix this bug"

Misc minor improvements and wish list update.

Added -asxhtml option to tidy, to generate XHTML when the -tidy option is used.

Added -draft and -final options.

Added code to highlight resolutions

Reverted change as it produces invalid HTML

Replacing at least 20 dashes by an hr

Replacing at least 20 dashes by an hr

Fixed a minor date formatting bug.  Noted (but did not fix) another formatting bug.

Removed extra CSS class

Improved a bit the HTML produced

Fixed bug in date parsing error message.

Improved date parsing and error messages.

Fixed a bug that caused the ScribeNick to be inferred as "scribe" even
though a Scribe was specified that matched an IRC handle.
Made scribe name comparisons case insensitive.
Removed a little dead code.

Added action status CLOSED as a synonym for DONE.

Added support for postponed action items

Corrected broken string terminator

Update wish list.

Added W3C Software License.

1. Now supports multiple scribes.
2. Added Mirc_Timestamped_Log_Format.
3. Other internal changes: ParseLine now always returns command
names in lower case, for easier comparison.

Shortened $maxLineLength to 67 because the style sheet indents them
3 spaces and Eudora was wrapping them at about 70 characters.

Improved action item processing, so that action items pasted in from previous
minutes are much more likely to be properly recognized.  In particular,
when the action items are summarized at the end of the minutes, they are
wrapped to fit within 76 characters (except for long words), and the
continuation lines now start with "... ", so they are recognized
properly as continuation lines when they are pasted in later.

Turned off $allowSpaceContinuation by default, because it is too error prone.
Updated the wish list.

Added -disallowSpaceContinuation to prevent leading space from indicating
a continuation line.

Made it "use strict;"

Keep zakim question reminders, i.e., when zakim says:
<Zakim> dbooth, you wanted to suggest mentioning semantics

Changed the order in which s/old/new/ substitutions are processed,
so that they are now processed first to last, because that's what
users expect.

Improved the MIT template to be more like what you get from
http://www.w3.org/Team/2004/04/06-MIT.html,new
.

Fixed another bug in action processing when -trustRRSAgent option is used.
Turned off warning about HTML being invalid, because it now appears to be valid.

Turned of $debugActions which was accidentally left on.

Fixed a bug in action item processing.

Delete duplicate Topic lines.
Changed $preferredContinuation to "..." because the other form
wasn't being recognized in the code that formats the paragraphs into HTML.

Deleted -teamSynonyms option.
No longer ignore lines starting with "*" (except if $inputFormat is Mirc_Text_Format).
Added "ScribeNick: ircName" command.  (Not yet documented.)
Improved the logic for guessing the scribe/scribeNick.
Misc code cleanup.

untest

test

No change.  Just trying to sync the version number.

Added -noDashTopics option to turn off -dashTopics option.

Added the ability to embed options in input, by saying "ScribeOptions: ...".
Fixed a bug in action status processing.

Improved the heuristics and warnings for the -dashTopics option.

Update wish list.

Changed the order of some action status words.

Added more action statuses from
http://www.w3.org/2001/sw/Europe/200401/actions

Changed the processing of s/// to be last first, so that you
can even correct previous instances of s///.

Added to WISH list.

Corrected warning about replacing list of people present.

adding markup/formatting as per http://lists.w3.org/Archives/Team/w3t/2003Dec/0158.html / http://www.w3.org/Team/2003/12/mockup-minutes

Made "Present:..." not give warning about replacing default list
of people present (picked up from zakim).

Checking in changes made earlier, but i forget what they were.

Added SCRIBEOPTIONS environment variable for default options

Updated version number???  (I thought CVS did that for me automatically.)

Fixed a bug in action continuation line processing.  Added a ParseLine()
function to make the parsing more rational.  Cleaned up the action code
a bit.

Fixed RRSAgent_Visible_HTML_Text_Paste_Format

Ignore empty argument.

Made -useZakimTopics the default.
Took out the long opening comment saying how to use the program.

Made sorting of action items case insensitive.

Join continuation lines in ACTION items.

Fixed a bug in Present/Regrets processing.

Updated the wish list.

Added Hugo_Log_Format and renamed the other Normalizer functions to
make them easier to read and make the documentation easier to write.
Filtered out completed actions when printing the list of people
with action items.

Added a wish list item for improving the guess of who attended.
Removed some extraneous debugging statements.

Fixed continuation line processing.  Added the -useZakimTopics option.

Added features: (1) Guess the scribe name; (2) -implicitContinuations option
(3) Recognize a leading space as indicating a continuation line.
Misc code cleanup.

Added -implicitContinuations option, though didn't test it yet!
Cleaned up misc code a little.

Added some debugging stuff.  Found a bug that puts "Scribe:" at the beginning
of the first scribed line.  Not yet fixed.

Added s/old/new/G for substitution forward and back.

Saving disk and bandwidth: removing commented out old text from templates

Fixed bug where Present could be considered a speaker
Added "Agenda:" in list of keywords
Better regexp for "Agenda: http://..."
Better warnings for present list changes

Added to the wish list.

Auto get attendee list from zakim statement.

Added version information

Fixed an incorrect warning message about agenda missing.

Added a feature wish list.

Added code to spot link to agenda

Added a -normalize option to output the normalized input, so that it
can be fed back into scribe.perl as input.  Not sure if it works right
yet.

Changed to print the people present after all are obtained.

Fixed ACTION highlighting.

Commented out debugging lines that were accidentally left enabled.

Added support for "Present+" and "Present-".

Improved ACTION processing to find the action status better and to pick up the
URL where the action was recorded.  You should normally use the -trustRRSAgent
option now.  It will probably become the default soon.

Also fixed a longstanding unnoticed bug
that prevented the last line of the input from being processed.

Reimplemented -trustRRSAgent from revision 1.18

Changed the order in which action item are listed.

Decided to list DONE and DROPPED items also.

Fixed action item processing in my script for formatting minutes:
http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Added NormalizerPlainText to recognize the format of plain text scribe notes.

Fixed a minor bug: "Date:..." line is now removed when it is recognized.

Problem: it is difficult to guess the definitive list of action items;
since CTCP actions are not logged, and RRSAgent gives feedback about action
items that way, it is impossible to handle "ACTION- 5" or
"ACTION 3 = New text for the action item" lines, which leads to duplicates,
and listing dropped action items as still in progress.

I added a -trustRRSAgent option:

#       -trustRRSAgent          Don't try to be smart about action items.
#                               Just trust what RRSAgent says

This tells the program to use RRSAgent list of action items instead of trying
to guess it.

Improved how "possibly present" names are collected.

Moved the doctype into the template documents

Now generating HTML 4.01
Only declaring the doctype used in one place and adding it to all the templates
  at the end
Using <h3 id="..."> instead of <h3><a name="...">: allows links in the topic

Smarter processing of the list of people present.
Added some warnings that give guidance on what to do.

Added warning for no topics.

fix rev

Added startup notice to incidate the version number and where to get
the latest version.

Added code to figure out the meeting date from the IRC log name.
Cool, huh?

Fixed a bug in NormalizerRRSAgentText (the function that recognizes
the .txt log format produced by RRSAgent).

Added separator: ----
Also improved input format detection.

Modified to accept several different input formats (HTML, Text,
square brackets, timestamps, Yahoo IM format, etc.).

Added a link, at the bottom of the generated output, to this program.

Clarified the documentation on what IRC log format is acceptable.
Moved template files to be internal to the program.
Made the PublicTemplate be the default.
No longer put the list of people possibly present into the
resulting HTML, since it was almost always wrong, and it's more
evident that it needs to be filled in if it is left empty.

changed

changed

Allow "Chairs:..." in addition to "Chair:..."

Initial version