comments:

... --Simon Michael, Sat, 07 Aug 2004 22:33:43 -0700 reply
bill.page1@sympatico.ca wrote:

Wed Jul 21 22:47:55 CDT 2004 bill.page1@sympatico.ca * Allow empty RemoteWikiURL name

Thank you! I had just pulled this from your repository.

FYI in case of larger changes in the future, current Zwiki policy is that contributions to my repo should have copyright assigned to me. ( http://www.gnu.org/licenses/gpl-faq.html#AssignCopyright ). I figured you wouldn't mind in this case.

Best, -Simon

... --Bill Page, Sun, 08 Aug 2004 16:29:54 -0700 reply
Simon,

You are welcome. I don't mind about copyright. How do I "assign copyright to you"? What specifically does this mean I should do that I did not do this time?

Regards, Bill Page.

... --Simon Michael, Mon, 09 Aug 2004 22:49:05 -0700 reply
You don't need to do anything, at least for these small fixes. If I ever find myself pulling something big from your repo, I'll check with you. Thanks again.

<> URL delimiting -- Tue, 10 Aug 2004 10:33:05 -0700 reply
I've found that URL's at even UNC(?) (\\domain\etc\etc) when enclosed in <>'s allow for email friendly delimiting. Auto-linking readers (Yahoo, gmail, Outlook) will properly acknowledge and escape spaces and following punctiation (ever clicked a bad URL simply because it was followed by a paren, comma or period?).

Having struggled with how to allow UNC links‡ in zwiki I would like to discuss how adding angle brackets to the Zwiki and STX formatters could alleviate linking woes, from general auto-linking to STX targets.

‡ I currently have added backslashes to the URL regex, and have added file to the list of protocols. I would like to add a delimiter to remove the need for space encoding. Linking to UNC names outside of Windows IE is still frustrating.

SearchPage? questions -- Tue, 10 Aug 2004 11:00:06 -0700 reply
Somewhere I recall seeing a note that links should be removed from the catalog. Did this happen? When experimenting (many months ago) with searching the catalog vs. full text I was frustrated that both the URL and link text were overlooked. Thus a "good searchable description here":http://example.com" link would fail the search with a key error

Secondly, when searching for "remove link" the SearchPage? failed on me ::

<> URL delimiting --Bob McElrath?, Tue, 10 Aug 2004 11:09:59 -0700 reply
anonymous [zwiki-wiki@zwiki.org]? wrote:

I've found that URL's at even UNC(?) (\\domain\etc\etc) when enclosed in <>'s allow for email friendly delimiting. Auto-linking readers (Yahoo, gmail, Outlook) will properly acknowledge and escape spaces and following punctiation (ever clicked a bad URL simply because it was followed by a paren, comma or period?).

Having struggled with how to allow UNC links‡ in zwiki I would like to discuss how adding angle brackets to the Zwiki and STX formatters could alleviate linking woes, from general auto-linking to STX targets.

‡ I currently have added backslashes to the URL regex, and have added file to the list of protocols. I would like to add a delimiter to remove the need for space encoding. Linking to UNC names outside of Windows IE is still frustrating.

So you want: <your@email.here.com>

This is virtually impossible, because: 1) STX allows mixing with HTML. 2) There is no proper escape sequence in STX.

STX is not fully aware of HTML, but rather ignores it entirely. You could change the dtmlorsgmlexpr in Regexps.py to also ignore anything between angle brackets, but this would cause other failures. So the examples above would get placed on web pages unmodified, where they would be invisible or screw up the HTML parser. We cannot convert them to <your@email.here.com> because then we must enumerate all possible things that angle brackets can contain (which must be a disjoint set from HTML) so that we can escape them.

The attempt to get a more formalized syntax for STX has resulted in the reStructuredText project, which you may want to investigate.

P.S. What is "UNC"?

SearchPage? questions --Simon Michael, Tue, 10 Aug 2004 11:40:20 -0700 reply

anonymous wrote:

Somewhere I recall seeing a note that links should be removed from the catalog. Did this happen? When experimenting (many months ago)
Hi there.. this is something each site admin has to do manually at present. I felt that automatically blowing away catalog fields, not necessarily used only for this wiki, was overreaching. Or not ?
Secondly, when searching for "remove link" the SearchPage? failed on me ::

I think you meant to include a traceback..

It worked for me just now: http://zwiki.org/SearchPage?expr=remove+link ?

... --Simon Michael, Tue, 10 Aug 2004 12:58:01 -0700 reply
Bob McElrath? wrote:

I would like to see IssueTracker spit off into a separate product, as I'm not sure what it has to do with a wiki...but it is a nice bug management system that one might want to optionally add onto zwiki.

It's a separate app built on top of zwiki, but so useful that I deemed it worthwhile baking it into python code and shipping with the core.

It's helpful to ship such things along with the core product to ease installation and version skew problems.

How would this be: move some of those extras (purple numbers, rating, issue tracking, maybe outline support) to a plugins directory that works like pagetypes -

# A Zwiki plugin is a module providing a mixin class that extends # ZWikiPage, thereby adding extra features to all wiki pages. # Mixins which used to be in the main ZWiki package are gradually being # moved here. Like page types, it should be possible for zwiki plugins to # be provided by other products also.

I think it could be made to work.

SearchPage? questions --DeanG, Tue, 10 Aug 2004 13:56:18 -0700 reply
http://zwiki.org/SearchPage?source=search&expr=%22remove+link%22 breaks. Huge trackback. I was trying to do an AND search. At some point I need to get up to speed on using some of the more modern Zope search products within Zwiki...one came out that supported regex's a few weeks ago, right? I'lll shoot some links in the air... ZopeWiki:SearchTools or ZopeWiki:SearchProduct

Hi there.. this is something each site admin has to do manually at present.

OK. I couldn't remember context, either. It's disappointing that catalog searches need to name each field, thus get redundant responses (IssueTracker) or lose fields (links), and full-text is potentially slow and complex-query inhibited.

diff --DeanG, Tue, 10 Aug 2004 14:05:01 -0700 reply
Just used the "Revert this change and all successive changes" button on the diff screen. Very handy. The "all successive changes" does the job of causing the user to ponder the implications while they figure out what it means. I didn't mean that left handed. It's succinct, precise, and I can't think of anything better.

Diff no longer shows up in the last edited n minutes ago by text. Was this due to spiders or safeguard to users? I get this mixed up with "changes"...but here again, I'm tired and note capable, lacking on the suggestions. Somebody must have turned on the "open for feedback" light. ;-) This content is more appropriate for IRC, my apologies.

diff --Simon Michael, Tue, 10 Aug 2004 18:17:12 -0700 reply
Thanks for the report Dean. This is Bob McElrath?'s diffform tweaks. I'm so dumb I haven't figured out what "revert all successive changes" is good for yet.. can someone enlighten me ?

diff -- Tue, 10 Aug 2004 22:34:08 -0700 reply

what "revert all successive changes" is good for?

Whacking test comments and spam into the bit-bucket. No-traces time travel...lest a search engine gets to the page first.

diff --simon, Tue, 10 Aug 2004 23:15:03 -0700 reply
Is it the same as "restore THIS version" ?

Duplicates in Digest Mode -- Wed, 11 Aug 2004 09:20:04 -0700 reply
Anybody else seeing duplicate comments mailed out in the digest mode for the whole wiki? (I haven't seen that on single page subscription)

diff --Bob McElrath?, Wed, 11 Aug 2004 09:52:36 -0700 reply
Simon Michael [simon@joyful.com]? wrote:

Thanks for the report Dean. This is Bob McElrath?'s diffform tweaks. I'm so dumb I haven't figured out what "revert all successive changes" is good for yet.. can someone enlighten me ?

"revert all successive changes" = "revert the change you're looking at and all the changes you had to look at to get here"

i.e. if it says "Revision 3" at the top it will also revert revisions 2 and 1. Hmmm okay my original patch put "Revision N" at the top of the page, but you can see it in the URL as currentRevision=N.

If you can think of a more intuitive way to word/organize this, I'd appreciate it... I thought this way was better than simply "revert" and/or "cancel".

Duplicates in Digest Mode --Simon Michael, Wed, 11 Aug 2004 10:10:26 -0700 reply
I'm getting duplicates, though I'm not in digest mode. I thought it was just me.

SearchPage? questions --Simon Michael, Wed, 11 Aug 2004 10:16:10 -0700 reply
As far as I know, zwiki.org search is basically working pretty well ? Correct me if that's not the case.

As for zwiki installations in general, the problem is unless we do expensive brute-force (non-catalog) searching, it is very dependent on which catalog and index products are used and how they are configured. So it's hard to make searching always work just the way we want out of the box. Ideas welcome.

diff --DeanG, Wed, 11 Aug 2004 14:24:12 -0700 reply

I thought this way was better than simply "revert" and/or "cancel".

It is. And (sigh) still beats out "Excavate", "Scorch", "Zwack-a-history", "Sweep-out", "Rail", and "A curse on you and those who came after!"

diff --Simon Michael, Wed, 11 Aug 2004 15:08:11 -0700 reply
I will accept a patch which uses any of those labels. :)

<> URL delimiting --Simon Michael, Thu, 12 Aug 2004 09:05:31 -0700 reply
I'm interested in understanding your ideas better. Could you lay them out one at a time in more detail ? Angle bracket links first, perhaps. You'd like to be able to write bare urls, as now except within angle brackets ? And this would be good because.. when you receive such links in mailouts, your email program can do a better job of linking them, because..

<> URL delimiting --DeanG, Thu, 12 Aug 2004 12:40:32 -0700 reply
Your email justifications are correct.

Right now I have some time for more information, but not complete. I was also looking for them as delimiters for URL escaping. In my case users copy paths from Windows File Explorer to the wiki. In a browser things are properly escaped, but they're "clear text" (??) in this case.

I'm still workign on the exact HREF syntax. I think in some cases the backslashes will need to be converted to forward slashes. (<\\domain\Fancy Path\> --> file://////domain/Fancy%20Path ?) Learning by trial and error..need to stop, google, and as you say, fully report later.

Feature poll --DeanG, Tue, 17 Aug 2004 14:30:18 -0700 reply
a. Did any features come up on IRC that didn't make it to this wiki?

b. New one: "Notify subscribers of change" checkbox on the edit screen. Not sure if anything other than the log would be necesary. Other possibilities are summary or diff.

... --Simon Michael, Wed, 18 Aug 2004 14:36:00 -0700 reply
This might be of interest to some. I'm not sure whether Zwiki page hierarchy avoids excessive cleverness, I hope so.

Ned Konz writes on the squeak list:

I am in the middle of reading Giorgio Busatto's 2002 PhD? thesis paper, entitled "An Abstract Model of Hierarchical Graphs and Hierarchical Graph Transformation" (http://theoretica.informatik.uni-oldenburg.de/~giorgio/papers/thesis.ps.gz , 672Kb ).

It is quite interesting and probably should be required reading for anyone interested in hierarchical graphs of any kind. It is also quite readable for someone like me who doesn't have a math or computer science degree (though some familiarity with the concepts and notation of set theory is helpful), and appears to be obviously useful, neither of which I can say about all the graph theory papers that I've read.

He argues (convincingly, I think) that because one might want to do a number of graph transformations and/or apply various policies to how hierarchical graphs are constructed, it is a bad idea to make excessively clever graph elements (nodes and edges). In fact, his model views the connection relationship between graph elements as an attribute of the graph itself, going so far as to move the knowledge of connectivity from the edges themselves into the graphs (more specifically, into the graph skeleton and its packages).

I think this view would work better in the general case, as it avoids having a dual structure that must be maintained separately in the case of the Connectors themselves.

Changes to the connectivity of a Connector are the user's assertions and desires about changes to the visible part of the graph itself. However, the morphs in the current world or other container may themselves only be a reflection of part of a larger graph.

By making the graph package (Busatto's term for a particular subset of the nodes and edges of the entire graph, itself part of the "hierarchy graph") separate and responsible for reflection on connectivity, a number of the problems are eliminated.

So questions about connectivity are always delegated to a graph package (what I called a "graph context"). This lets us have the same graph element (nodes or edges) shared, with some elements visible in some contexts and not others.

Authentication thought -- Fri, 27 Aug 2004 22:41:03 -0700 reply
I marveled at the social simplistity of this thought, of putting a pass-phrase in the FAQ/charter that is required to be enclosed in your sign-on or first post.. http://enthusiasm.cozy.org/archives/2004/08/denial-of-service-attacks-directed-at-communtities/

I came to know conclusion how this could be integrated into a Zwiki.

ZopeWiki:AllPages summary -- Sat, 28 Aug 2004 07:17:34 -0700 reply
I added the summary call to the ZopeWiki:AllPages . Looks OK. Extra space is annoying, but summary feature works well here. Not sure of resource issues. Should be conditional? -DeanG

RecentChanges? clunk period=week --DeanG, Mon, 30 Aug 2004 12:25:13 -0700 reply
http://zwiki.org/RecentChanges?period=week

Traceback (innermost last):

 Module ZPublisher.Publish, line 180, in publish_module_standard 
Module Products.PlacelessTranslationService.PatchStringIO, line 51, in new_publish 
Module ZPublisher.Publish, line 131, in publish 
Module Zope.App.startup, line 204, in zpublisher_exception_hook 
Module ZPublisher.Publish, line 101, in publish 
Module ZPublisher.mapply, line 88, in mapply 
Module ZPublisher.Publish, line 39, in call_object 
Module Products.ZWiki.ZWikiPage, line 237, in __call__ 
Module Products.ZWiki.ZWikiPage, line 245, in render 
Module Products.ZWiki.pagetypes.stx, line 88, in render 
Module Products.ZWiki.ZWikiPage, line 288, in evaluatePreRenderedAsDtml 
Module OFS.DTMLDocument, line 128, in __call__
<ZWikiPage at /zwikib/RecentChanges>
URL: http://zwiki.org/RecentChanges/manage_main
Physical Path:/zwikib/RecentChanges 
Module DocumentTemplate.DT_String, line 474, in __call__ 
Module DocumentTemplate.DT_Let, line 75, in render 
Module DocumentTemplate.DT_Util, line 198, in eval
__traceback_info__: period 
Module <string>, line 1, in <expression> 
Module AccessControl.ZopeGuards, line 67, in guarded_getitem 

KeyError?

darcs retrieve problem -- Thu, 02 Sep 2004 06:19:35 -0700 reply
$ darcs get --partial http://zwiki.org/repos/ZWiki Copying patches... ....................................................................................... darcs failed: Failed to download URL http://zwiki.org/repos/ZWiki/_darcs/patches/20040711035747-e02d6-b1a3673e583523de8ec1fb7076c63b3cd6c33dd1.gz libcurl: HTTP error (404?) $ snarf http://zwiki.org/repos/ZWiki/_darcs/patches/20040711035747-e02d6-b1a3673e583523de8ec1fb7076c63b3cd6c33dd1.gz error: HTTP error from server: HTTP/1.1 503 Service Temporarily Unavailable

release today --Simon Michael, Thu, 02 Sep 2004 09:34:18 -0700 reply
Client work has been following the monthly iteration notion lately, so as usual, work deadline..

There are a number of silly bugs I'd like to squash - bogus contents in new wikis, /wikipage url giving double skin in cmf/plone, etc. - but the time is upon us. 0.34 will go out later today one way or another.

bug hunt report --Simon Michael, Thu, 02 Sep 2004 09:52:19 -0700 reply
During the first end-of-month bug hunt, 8/27 to 8/31, 2 issues were closed.

question about dtml interpretation in rst --stefan rank, Tue, 07 Sep 2004 01:45:16 -0700 reply
i was wondering why there is no dtml processing for rst? For me the following patch:

 [enable dtml interpretation in rst pagetype
 stefan(DOT)rank(AT)oefai(DOT)at**20040907081416] {
 hunk ./pagetypes/rst.py 23
 +    supportsDtml = yes
 hunk ./pagetypes/rst.py 43
 -        t = page.preRendered()
 +        if page.dtmlAllowed():
 +            t = page.evaluatePreRenderedAsDtml(page,REQUEST,RESPONSE,**kw)
 +        else:
 +            t = page.preRendered()
 }

allows dtml in rst. Using the idiom:

 .. raw:: html

    allows any dtml you like, like &dtml-pageCount; or other handiwork of the devil

well yes. this would be dtml/html only as a block-level element, inline is little bit harder and would probably require the addition of an "interpreted role" to restructuredtext.

thought this might be interesting for others too.

question about dtml interpretation in rst --Simon Michael, Tue, 07 Sep 2004 10:16:49 -0700 reply
Nice, Stefan. I have created HowToEnableDTMLInReStructuredText as a next step. Could you add an example or two of rst-friendly dtml snippets ?**

Comment/Note as change history -- Thu, 09 Sep 2004 11:15:28 -0700 reply
In chatting with a colleague about use of mail-out for page update notification, and trying to tweak the signal/noise ratio, I've come up with the following:

a. Every Edit with a Note gets appended to the page as a comment (without /diff .. usally meaningless, inspired by IssueTracker note field.)

b. Only comments with a Subject gets marked up and added as comments, all others are Appends (and get appended to the Document, not the comments.) Comments w/ subject sent out as usual.

Summary: If it matters, summarize it, and folks who care will be notified.

Anyone interested in pursing a new mail_out policy of "Logs" with me?

Side benefits: RecentChanges? and RSS filter.

Perhaps I should take this to WikiMailDiscussion... Also not sure if add a comment button name text would need to be changed, or just chagned to "Add" with the "subject" heading changed to "Comment subject:" (There's also a whole functionality tangent of prepend/append to document, or comment reverse order when the page has more of a log/journal/blog/story/update perspective, or comment insert on thread-mode. That would be considered in source, balanced by YAGNI and extensibility.)

p.s. I haven't looked into the plug-in architecture.

Diderot -- Mon, 13 Sep 2004 07:25:56 -0700 reply
via DailyURL?! http://sourceforge.net/projects/wikiwriter/ "a Python-based WikiWriter?. A graphic user interface with text editor and category browser for interaction with the free encyclopedia Wikipedia."

Anybody tried this? Might be Win32 only.

Diderot -- Mon, 13 Sep 2004 07:41:53 -0700 reply
there is nothing yet than a picture on the website .... no files ... no documention ... it might be a nice idea ;)

Diderot -- Wed, 15 Sep 2004 05:59:20 -0700 reply
I thought there were files in the CVS repository

RESPONSE ETag? / Last Modified -- Wed, 15 Sep 2004 08:53:40 -0700 reply
Perhaps this is a Zope discussion, but how would one go about setting the RESPOSNE.Etag or Last Modified value for a wiki page? I'm hoping to populate it with the page's last edit time. (Is there a function equivelent to this for a URL workaround?: e.g. http://..../Pagename/last_edit_time() )

http://www.pocketsoap.com/weblog/stories/2002/05/0015.html

HTML (DTML) mode w/ Epoz --DeanG, Thu, 16 Sep 2004 08:48:54 -0700 reply
DTML doesn't work in an Epoz edited page. Proposal:

Change: HTML (+ DTML) change to HTML (WYSIWYG)

Add: HTML (+DTML) for the StructuredText escaping non-Epoz'ers we currently use HTML mode for.

Should this be an issue page?

But...that little check-box seems to switch from raw to wysiwyg, but...

RESPONSE ETag? / Last Modified --Simon Michael, Fri, 17 Sep 2004 15:32:11 -0700 reply
If dtml is enabled, I would guess it's something like:

 <dtml-call "RESPONSE.setHeader('Last-Modified',value)">

You might find RESPONSE's api at ZopeWiki:APIDocs .

HTML (DTML) mode w/ Epoz --Simon Michael, Fri, 17 Sep 2004 15:45:32 -0700 reply

Change: HTML (+ DTML) change to HTML (WYSIWYG)

Yes I think this should be an issue. HTML + DTML is indeed a bit misleading when Epoz is in effect.

The other things may be separate issues or not, there are some related issues already.

String as Tracker field -- Tue, 21 Sep 2004 21:47:47 -0700 reply
I'm trying to add a free-form string as a tracker field.

I'm only get a few pages indexed for it to show up in the catalog, even when ALL issue pages have a value for that field.

Has anyone done this before? The IssueTrackerOwnerAndDueDate patch added the owner in as a FieldIndex?, as it's a selection box. Is that the only route to go?

String as Tracker field --Simon Michael, Wed, 22 Sep 2004 08:02:28 -0700 reply
You could fire up zopectl debug and try the catalog search your code is doing, to understand what's up.

MediaWiki? table format -- Thu, 30 Sep 2004 14:19:41 -0700 reply

The wikipedia/media wiki table format code looks interesting. Little column spacing requirements, no HTML noise, built in support for table formatting.

In my world STX tables are a pain (editing in browser), HTML is ugly but the norm. reST doesn't look much better.

To those whom parsing comes naturally: Does it look like a reasonable format to code around?

MediaWiki? table format --Bob McElrath?, Thu, 30 Sep 2004 14:32:21 -0700 reply
anonymous [zwiki-wiki@zwiki.org]? wrote:

The wikipedia/media wiki table format code looks interesting. Little column spacing requirements, no HTML noise, built in support for table formatting.

In my world STX tables are a pain (editing in browser), HTML is ugly but the norm. reST doesn't look much better.

To those whom parsing comes naturally: Does it look like a reasonable format to code around?

I've been thinking about this and I see the number of pagetypes exploding. stx with/without tables/discussion or with wwml or moinmoin or wikipedia markup. This is a bad thing.

I think it's nice to have lots of choice, but rather than "fix" the STX table problem I think we should implement other wiki syntaxes that do things better. There's no use in reinventing the wheel and there exist many people who already know one syntax or another. Taking a piece of one syntax and mixing it with another syntax is a recipe for user confusion. I think users should choose one pagetype for their site.

I can implement some kinds of interconversion. All these wiki types are pretty simple. Once a StructuredText DocumentClass? is defined for each wiki type, a renderer can be written to interconvert. Some info will be lost in some conversions (i.e. the attributes on mediawiki tables when converting to stx). But this will give users a way to "upgrade" the syntax of their wiki when they hit the limitations of stx. A STX + HTML pagetype could even be converted to mediawiki format without losing table attributes. Known lossless conversions could even be done by anonymous users. (i.e. I want to edit this page in wwml rather than stx -- via a preferences page)

All this, of course, requires a resonable AdminForm to make sense of the plethora of choices.

MediaWiki? table format --Simon Michael, Thu, 30 Sep 2004 16:29:15 -0700 reply
Ah, you're making sense there Bughunter Mcelrath. There is indeed a need to simplify the available options and choose good defaults that will work for people.

The trouble is, for our purposes there is no one set of markup rules that gets it all right! Your exploration of the STX framework may give us more freedom to combine markup rules of our choice along with the wiki-linking features in a single more unified system. Be warned (I think you are) it's going to take time to work out all the issues with prerendering, i18n, supporting other modes, migration, etc.

About mediawiki (wikipedia) table format - as far as simplified table formats go, I'd say yes that's a good one to support. They have put some thought into making it usable for editors. And, it's a step towards a full media wiki page type.

# shortcut? -- Thu, 07 Oct 2004 09:07:41 -0700 reply
What's up with the # shortcut for issues? #001

feature warning -- Thu, 07 Oct 2004 09:15:10 -0700 reply
Is it reasonable to ask that a new feature (even in experimental mode) be added as an Issue?

This would also provide a replacement venue for DevDiscussion? if/when we go back to GeneralDiscussion.

MediaWiki? table format -- Thu, 07 Oct 2004 09:17:58 -0700 reply

it's going to take time to work out all the issues with prerendering, i18n, supporting other modes, migration, etc.

+ dtml and other Dynamic content integration.. :-/

Right now STX at Zwiki is still basically STX. What are considerations for crossing the line into full-blown ownership of ZwikiSTX??

feature warning --Simon Michael, Thu, 07 Oct 2004 10:18:20 -0700 reply
#924 short issue names and links

discussion --Simon Michael, Thu, 07 Oct 2004 10:23:03 -0700 reply
PS I do post major things on GeneralDiscussion/DevDiscussion? at times, but in general I assume that people who want to be fully up-to-date are subscribed to the whole wiki. That way things can be posted where they best belong.

Related: what about renaming GeneralDiscussion to GeneralDiscussion. This better reflects the reality that "development" topics can pop up there, and communicates that it's a catch-all, while still keeping DevDiscussion? as a place for in-depth dev topics.

MediaWiki? table format --Simon Michael, Thu, 07 Oct 2004 10:25:41 -0700 reply
I think Bob's current work is taking us into that zone..

MediaWiki? table format --Bob McElrath?, Thu, 07 Oct 2004 10:57:01 -0700 reply
anonymous [zwiki-wiki@zwiki.org]? wrote:

it's going to take time to work out all the issues with prerendering, i18n, supporting other modes, migration, etc.

+ dtml and other Dynamic content integration.. :-/

Right now STX at Zwiki is still basically STX. What are considerations for crossing the line into full-blown ownership of ZwikiSTX??

Simon Michael [simon@joyful.com]? wrote:

I think Bob's current work is taking us into that zone..

Prototype is here:

    http://bob.mcelrath.org/WikiStructuredText/

with a small test program.

There will be one Document Class for each page type that is handled by the StructuredText machinery (stx, moin, etc). The pre-parsed DOM tree will be stored as _preRendered. Parsing stx takes 99% of the time in rendering a page. In addition there will be a Document Class that does wikilink-ing only, so (for instance) reST (which does not use the StructuredText machinery) will generate HTML, and the wikilink document class will pre-parse it. This replaces the fragile <zwiki>WikiLink</zwiki> wikilink-marking mechanism, and is faster.

Note that this makes it trivial to mix and match features between pagetypes that use StructuredText. (i.e. STX with mediawiki table syntax)

Also I will write a LaTeX? backend, so printing a wiki page without the zwiki or plone headers will be easy. (latex -> ps)

If anyone wants to pitch-in, see the above prototype. For each syntax object (such as a moin table) there is one function which recognizes it that must be written. It's easy to merge such changes since they are self-contained. So we can do one piece at a time...

MediaWiki? table format --Simon Michael, Thu, 07 Oct 2004 11:53:38 -0700 reply
Is it a page type yet ? Do you want help ? Is it a darcs repo ? (I tried to pull from zwiki-test but got an error)

MediaWiki? table format --Bob McElrath?, Thu, 07 Oct 2004 11:57:57 -0700 reply
Simon Michael [simon@joyful.com]? wrote:

Is it a page type yet ? Do you want help ? Is it a darcs repo ? (I tried to pull from zwiki-test but got an error)

Still working...

It requires extensive modifications to ZWikiPage.py.

zope hung again --Simon Michael, Thu, 07 Oct 2004 14:19:05 -0700 reply
Oh, and if I'm reading the reports right the server was down for about 8 hours last night. I'm becoming desensitized to this. :(

zope hung again -- Thu, 07 Oct 2004 15:59:27 -0700 reply
Next time it happens, I'll attach gdb and look for clues.

articles --Simon Michael, Fri, 08 Oct 2004 14:51:43 -0700 reply
Two fine articles:

http://www.pingwales.co.uk/software/cox-on-better-software.html

http://plone.org/Members/pupq/best_practices

using _prerendered in RSS feed? -- Fri, 15 Oct 2004 07:16:08 -0700 reply
I'm trying to improve my RSS v2 feed. I just dump the full text of my bloggish pages into the feed, but it's the raw structured text, and I'd like to get closer to HTML. Right now I use page="getObject()" and text="page.text()" - is there an equivalent "page.prerendered()"? or "page._prerendered"?

using _prerendered in RSS feed? --Simon Michael, Fri, 15 Oct 2004 08:16:13 -0700 reply
Use the source Luke.. is render(bare=1) what you're looking for ?

using _prerendered in RSS feed? -- Fri, 15 Oct 2004 08:54:13 -0700 reply
Doesn't that trigger the rendering? And hasn't that already been done and stored in the zodb?

By the way, I didn't mention before that this occurs in the context of dtml-in Catalog which was probably a significant omission....

using _prerendered in RSS feed? --Simon Michael, Fri, 15 Oct 2004 09:47:37 -0700 reply

Doesn't that trigger the rendering? And hasn't that already been done and stored in the zodb?

Yes. Usually text formatting has been done, but linking has not been done. You can get this intermediate rendering with getObject().preRendered(). You could maybe call renderMarkedLinksIn on that.

To render and link some arbitrary text, like summary which you can get from the catalog result, you could do renderLinksIn(renderStxIn(summary)).. but summary's markup may be incomplete since it's truncated.

using _prerendered in RSS feed? --Simon Michael, Fri, 15 Oct 2004 09:49:09 -0700 reply

Doesn't that trigger the rendering? And hasn't that already been done and stored in the zodb?

Yes.

Oh, but render() does make use of the prerendered data if it's there.

ZopeWiki and SearchPage? -- Mon, 18 Oct 2004 10:57:25 -0700 reply
Consider adding the "other sites" section back to the SearchPage?, in the least to add the ZopeWiki. Maybe even "Zopewiki via google" if you need to conserve resources.

...wonder if anyone has wanted SisterSites functionality now that ZopeWiki has developed momentum.

MVCC (multi version concurrency control) support -- Mon, 18 Oct 2004 11:14:57 -0700 reply
Saw this in the Zope 2.8 beta announcement

Is this interesting to Zwiki for either generic read-confict errors or handling edit conflicts?

MVCC (multi version concurrency control) support --Simon Michael, Mon, 18 Oct 2004 12:56:50 -0700 reply

Is this interesting to Zwiki for either generic read-confict errors or handling edit conflicts?

I haven't yet seen a zwiki where read conflicts were an issue, but it should certainly help with scaling to more traffic. It doesn't affect wiki edit conflicts, that's a higher level.

New issue number feedback --DeanG, Wed, 20 Oct 2004 07:03:53 -0700 reply
This morning I wanted to test editing and previewing an issue. I thought issue #1 was a good place to start. (Hey, that numbering prefix worked well in that context!)

I started by URL. Blech. http://zwiki.org/1 yields issue #100, same for 001 , and 0001 is not found.

I searched the Issues for 1. No luck. Maybe we need a "lookup by #" form.

So now I'm here, supplying feedback and adding #1 to a page so I can link to it.

Also, does the new sumbering system mean we can, effectively, now track more than 9999 issues?

issue edit preview --DeanG, Wed, 20 Oct 2004 07:09:14 -0700 reply
Issue edit preview works fine. The preview doesn't include the issue properties form, but that's not a big deal to me. It's even less a priority than preview on comments (Tangent: but would be same when comment area is part of issue form, which makes initial use understanding simpler (Do I fill in form adn comment? Which button?). It was a straightforward but code level mod when changing a personal wiki. I'm striking out w/ IssueTracker searches this morning. There's an issue related to this I can't find.

Revert button -- Wed, 20 Oct 2004 20:46:13 -0700 reply
3 cheers for the revert button.

Backlinks -- Sat, 23 Oct 2004 18:04:09 -0700 reply
I broke the Backlinks on the AboutThisSite page.

I was trying various options on multiple non-backlink parents. Will try to reproduce exact cause.

Backlinks --DeanG, Sat, 23 Oct 2004 18:14:19 -0700 reply
Also, please consider: In the event that someone uses the re-parent link from hte footer (or URL) and there is currently 2 or more parents, the new parent would be added (if it could), the old would not be removed, and the user would be re-directed to the backlinks page.

Redirecting to the backlinks page is also the recommended behavior for failure to find a (non-null) backlink. (Related: I wouldn't suggest adding a search button to the page, as it does exist in the header, and using it would abandon any form changes the user had made.)

Backlinks --DeanG, Sat, 23 Oct 2004 18:31:10 -0700 reply
Created #952. (Pardon my test: [Reparent and Backlink errors with Additional Links]?)

Just lost the DevDiscussion? page's history by clicking on reparent with an empty form. Ack! It wouldn't revert from the first view on the /diff page either!

using _prerendered in RSS feed? -- Sun, 24 Oct 2004 14:50:57 -0700 reply
Does it render the entire page, including header/footer shmutz (append form, AnnoyingQuote, etc.)?

using _prerendered in RSS feed? -- Tue, 26 Oct 2004 09:42:41 -0700 reply
I'm not aware of your you're doing, but want to also point out: http://zwiki.org/Rss2/text

Catalog crash -- Tue, 26 Oct 2004 12:10:36 -0700 reply
The SearchPage? is giving mostly brute force searches lately. Is this a memory or other catalog issue? I thought a search I did actually used the catalog but didn't return results (used the URL) but every query after that was brute force.

I like the latest SearchPage? features when used with a catalog, but don't want to implement it if it's prone to failure. Maybe that failure isn't relevant on a smaller or differently configured wiki?

1000 Issues -- Thu, 28 Oct 2004 11:25:56 -0700 reply
In a quest to get to 1000 issues, I'd like to offer the idea that we add concise, thought-out wish-list items. There are many of these posted as comments on the site, along with in our head. Examples?: Ship with an RSS/Atom feed, Build the AdminForm, Ship with a plug-in "template"/scaffolding/non-invasive example, update and ship a Print method, Document Zwiki with ZSyncer, horizontal page heirarchy option, sub-topic scoped subscription.

Would this set a bad precedent?

1000 Issues --Bob McElrath?, Thu, 28 Oct 2004 12:36:57 -0700 reply
anonymous [zwiki-wiki@zwiki.org]? wrote:

In a quest to get to 1000 issues, I'd like to offer the idea that we add concise, thought-out wish-list items. There are many of these posted as comments on the site, along with in our head. Examples?: Ship with an RSS/Atom feed, Build the AdminForm, Ship with a plug-in "template"/scaffolding/non-invasive example, update and ship a Print method, Document Zwiki with ZSyncer, horizontal page heirarchy option, sub-topic scoped subscription.

Would this set a bad precedent?

Nooooo, please do.

I need my wishes coallated, copied and stapled, or they get lost under the pile of walrus intestines.

(?i) --Bob McElrath?, Sat, 30 Oct 2004 02:32:23 -0700 reply
The reason I removed the (?i) from dtmltag is that python regexes seem to take the set of ALL flags that appear in a regex. That is, flags like (?i) do not respect grouping. For example: re.match(r'a((?i)b)', Ab) should NOT match, but it does.

Therefore, creating regexes by r'(%s|%s)'%(r1,r2) where regexes r1 and r2 have flags in them is very dangerous since the resulting regex has ALL flags applied simultaneously.

I discovered this because I have a regex doing this with wikilink and dtmlorsgml, which makes wikilinks case insensitive. That kind of sucks.

(?i) --Bob McElrath?, Sat, 30 Oct 2004 14:28:52 -0700 reply
Dean Goodmanson [goodmansond@gmail.com]? wrote:

like (?i) do not respect grouping. For example: re.match(r'a((?i)b)', Ab) should NOT match, but it does.

FWIW. I did verify it matches via Kodos regex debugger, and a C# regex debugger. I believe both claim to implement Perl regex rules.

I checked in perl and it does NOT match. (perl 5.8.4)

Thus the moral is...don't use (?i) and similar modifiers in regexes, if you plan to compose regexes later like r"(%s|%s)"%(r1, r2) because they will infect the entire regex. Another similar issue is the use of !? in regexes (where you're trying to escape a wikilink). This should only appear in "final" regexes. (this is okay as it stands in the zwiki codebase)

FYI --simon, Sun, 31 Oct 2004 17:42:32 -0800 reply
Archiving and retiring DevDiscussion? page. Hopefully the discussion choices on FrontPage are a little clearer.