Archived GeneralDiscussion

calling title --DeanGoodmanson, 2003/05/01 03:20 GMT
Sorry for the poor attempt to point out "security.declareProtected(Permissions.View, folder)", it doesn't appear on any of the util.py function calls.

Trying to reach Simon --2003/05/01 16:07 GMT
Simon - simon AT joyful.com doesn't appear to work ("message delayed 24 hours", is what I'm getting). Is there another email address that might work for you? I've got a question that I'd like to ask. (maybe simon AT zwiki.org ?). (this is TonyRossini, rossini@u.washington.edu )

Zwiki 0.18 released --Simon Michael, 2003/05/01 16:26 GMT
Summary: Full Plone and CMF skin, miscellaneous fixes.

More information: http://zwiki.org/ReleaseNotes , http://zwiki.org/KnownIssues , http://zwiki.org

Download: http://zwiki.org/releases/ZWiki-0.18.0.tgz or http://zope.org/Members/simon/ZWiki/ZWiki-0.18.0.tgz

Peace, --Simon

Trying to reach Simon --Simon Michael, 2003/05/01 16:34 GMT
Eh, what's this courtesy copy business..

Tony writes:
> Simon - simon AT joyful.com doesn't appear to work ("message delayed 24
> hours", is what I'm getting). Is there another email address that might
> work for you?

Hey Tony.. that sucks, any more details in the error ? My mail provider has been giving intermittent trouble. I'm receiving other mail this morning.


> I've got a question that I'd like to ask.

Shoot.

also changing create-page "?" to square brackets --Simon Michael, 2003/05/01 16:47 GMT
Hi Bill - glad you are exploring the issues with this. I want this option.


> I've always hated the question-mark because it confuses non-wiki people)

Weird.. I find the question mark simple and elegant. A single click reveals the meaning to a newcomer. I dislike wikis where existing and non-existent page links are the same or similar.

structured text or hyper-link regex? --DeanGoodmanson, 2003/05/01 18:54 GMT
The following URL fails on the asteric: http://www.edmunds.com/apps/uvl/idp.do?vehicleId=795205&tid=edmunds.u.stdalone.uvl.results.1.Honda*&zip=54469

..let's see if it works in a STX link Maybe next year*&zip=54469

re changing "?" to square-brackets --BillSeitz, 2003/05/01 22:25 GMT
Well, one place where I think it's really confusing is where you print a page for someone! Second place would be where you copy a chunk of body into an HTML email.

I agree that making both kind of links look identical is bad.

In my plan, existing pages will look like Bill Seitz while uncreated potential pages will look like :

 [Bill Seitz]

(but with the latter bracket, or maybe both, being "hot").

JordanCarswell --2003/05/01 22:39 GMT
Bill, see:

http://zwiki.org/ZwikiPrinting

You can use this page to print or extract an HTML version of your page. I use it at http://www.commonplacebook.org where it works quite nicely.

solution to getslice problem? --BillSeitz, 2003/05/01 22:40 GMT
Anyone familiar with this? I thought upgrading Zope would take care of it, but my laptop's z2.5.2 seems to display the same problem.

http://lists.imeme.net/archives/imeme-users/2003-February/002406.html

ZwikiPrint? and Linking pondering --DeanGoodmanson, 2003/05/02 03:36 GMT
The /print method prints the ? create this page links. :-(

The only way I can see to correct this is to be able to pass in a parameter (such as bare) that would turn off wiki linking. In that case, people still might guffaw at the camel case.

A win in this scenario is that this link has a specific CSS class attribute, so you can format that ? link to a more sub-dued form. Make it super/subscript, text color the same as background... /print can call it's own stylesheet.

Although I believe WikiNames are a much niftier/geekier function over form, I'm grateful for freeform links in the same way I'm grateful for embeddable HTML. Brackets replacing ?'s are a keen idea, as they support this form linking. As I may have said before, I was surprised when peer put URL's in brackets out of his personal intuition. I was then pleased that they worked as expected.

Bill: What do you think of, when the "bare" attribute is "1", to remove all wikilinks, including new page links? Could your current extension support that?

Your feedback on the current /print functionality is also appreciated.

StructuredText and locales --ArnoPucher, 2003/05/02 08:35 GMT
Due to my Issue (#462 Text formatting with special chars like @ or german "Umlauts") I tried like described the problem in the zope howto (see comment in issue).

Anybody tried this already too - successfully ?

ad StructuredText ... --ArnoPucher, 2003/05/02 09:20 GMT
Hell, it worked :)

the mentioned "locale" incorparation rendered perfectly. I has been a browser refresh problem after restarting Zope with "- L "de_de@euro"".

So I closed issue #462.

solution to getslice problem? --Simon Michael, 2003/05/02 16:58 GMT

> Anyone familiar with this? I thought upgrading Zope would take care of

Not me.. I'd try #zope and file a zope collector issue.

Re: [IssueNo0506? How do I upgrade / update zwiki external method in plone site from 0.17 to 0.18?] (new) --Simon Michael, 2003/05/02 17:02 GMT

> Can anyone help me on this? How do I do to upgrade and update zwiki
> inside plone from 0.17 to 0.18?

If you want your wiki to start using the zwiki_plone skin, follow the steps in ReleaseNotes. Otherwise I think you don't need to do anything.

ZwikiPrint? and Linking pondering --Simon Michael, 2003/05/02 17:04 GMT
Someone should figure out how to make and best integrate a printing stylesheet.

Thank you Simon. Still need your help on IssueTracker used inside Plone site --2003/05/02 17:27 GMT
Really thank you for the instructions. So embarrassing that I forgot to check the ReleaseNotes.

Regarding IssueTracker inside plone..... I really wanted to use zwiki style IssueTracker inside plone. I just like zwiki IssueTracker. I think it is the best and most user friendly and clear arranged IssueTracker compared to all others. Could you please spend some more time on it and give some instructions on the installation of IssueTracker inside plone?

Thank you so much. JY Lee

... --2003/05/02 19:47 GMT
When I try to create a Wiki Folder item in my Plone site, it fails with this exception: Bad Request: The id "standard_error_message" is reserved. It looks to me like CMFCore.PortalFolder._checkID() refuses to create an object with that id because the portal root object has an attribute with that same id. Products/ZWiki/content/basic/ has a standard_error_message.dtml file that causes ZWiki to create a standard_error_message object itself. I'm stumped -- how do I get past this? Context: ZWiki-0-18-0, Zope 2.6.1, CMF-1.4beta1, CMFPlone? 1.0.1.

ZwikiPrinting Stylesheet --2003/05/03 19:45 GMT

> Someone should figure out how to make and best integrate a printing stylesheet.

  1. Should it use the wiki_stylesheet object, or a seperate wikiprint_stylesheet ? (If no wikiprint_stylesheet, use wiki_stylesheet?)
  2. What should be contained on a default version?
    • class new style. Subdued ? link...to what extent?
    • Subdued a tag rendering? Same as p ? (No hyperlink font or underlining notation.)

..DeanG

Purple Numbers --2003/05/03 20:20 GMT
The good people at http://www.planetwork.net/ would like to enable purple numbers in ZWiki. Purple Numbers: http://www.eekim.com/software/purple/purple.html Has anyone done work on this?

Re: Purple Numbers --JordanCarswell, 2003/05/04 01:34 GMT
For more info, see PurpleWiki, which uses the "purple" markup. It's pretty cool.

http://purplewiki.blueoxen.net/cgi-bin/wiki.pl

Stylesheet bugs --2003/05/04 06:44 GMT
When a wiki_stylesheet node is created, zope hangs when pages are requested. Some python process uses 100% of the CPU and never actually gives a page... Also in skins/default, wikipage.zpt refers to "stylesheet" while standard_wiki_header.dtml refers to "wiki_stylesheet". It is the one in the header that causes the hang. The one in wikipage.zpt does not cause my browser to request the stylesheet from the server.

Thank you Simon. Still need your help on IssueTracker used inside Plone site --Simon Michael, 2003/05/05 16:20 GMT

> Regarding IssueTracker inside plone..... I really wanted to use zwiki
> style IssueTracker inside plone. I just like zwiki IssueTracker. I think
> it is the best and most user friendly and clear arranged IssueTracker
> compared to all others. Could you please spend some more time on it and
> give some instructions on the installation of IssueTracker inside plone?

Hi JY, I'm glad you like IssueTracker (aka ZwikiIssueTracker).

I'd love to spend more time on it but must also pay bills and limit my time at the screen. So you could help move this along by funding it, or by taking responsibility for the task and making it happen (see below).

I did try to get the zwiki tracker working in a plone site the other day but gave up before adding all the necessary indexes and metadata fields to the site catalog (portal_catalog). I'm always a little reluctant to add what feels like bloat to a site catalog. Someone should do this and report on the remaining problems in detail (maybe you have already ?).

Aside: I'd like to collect some guidelines for the issue cheerleading role, which is an important way to help. (What do other projects call it ?) It includes: checking for relevant past discussion, finding/opening an appropriate issue, investigating the situation fully and communicating it clearly, setting up screenshots if needed and demos whenever possible, requesting technical guidance where needed and making it easy for me or others to provide same.. etc.

bad request: standard_error_message reserved --Simon Michael, 2003/05/05 16:32 GMT

> When I try to create a Wiki Folder item in my Plone site, it fails with this
> exception: Bad Request: The id "standard_error_message" is reserved. It looks
> to me like CMFCore.PortalFolder?._checkID() refuses to create an object with
> that id because the portal root object has an attribute with that same id.
> Products/ZWiki/content/basic/ has a standard_error_message.dtml file that
> causes ZWiki to create a standard_error_message object itself. I'm stumped --
> how do I get past this? Context: ZWiki-0-18-0, Zope 2.6.1, CMF-1.4beta1,
> CMFPlone? 1.0.1.

Hi.. I'm surprised you're still seeing this. ZWiki/CMFInit?.py line 73:

            and not filename == 'standard_error_message.dtml'):

was added a few releases back to omit standard_error_message within CMF. I can't test right now, but maybe this gives a clue. If your installation looks ok, find or open an issue for this if you can.

Purple Numbers --Simon Michael, 2003/05/05 16:39 GMT
That's interesting, and with that name, it's bound to succeed. :)

Shouldn't be hard to add a prototype implementation to zwiki. My dumb approach would be to add a final pass in the render_ method that adds the anchors to all paragraph and heading tags and optionally (user cookie ?) adds the visible purple numbers. This may tie in somehow with our need for permanent comment links.

editform.pt --JordanCarswell, 2003/05/05 16:43 GMT
The editform page template released with 0.18.0 still has the DTML methods for the form and input tags. Here, I think, is the correction:

   <form method="post" action="page_url" enctype="multipart/form-data" tal:attributes="action oldid">
     <input name="timeStamp" type="hidden" value="editTimestamp" tal:attributes="value here/editTimestamp" />
     <input name="page" type="hidden" value="id" tal:attributes="value id" />

editform.pt --Simon Michael, 2003/05/05 16:58 GMT
Thanks, opened IssueNo0507? .

single-page FAQ ? --SimonMichael, 2003/05/05 17:31 GMT
I think FAQ was broken into separate pages prematurely. I'm thinking of putting it all on one page for simplicity, preserving sections so that we can split it up again if needed.

Re: Purple Numbers --2003/05/06 00:35 GMT
I'm going to attempt to implement purple numbers in ZWiki. I expect to have a question now and then so I've built page called PurpleNumbers. Please chime in on the discussion there (hope I don't have to talk to myself too much!) When I'm finished with the work I hope it will be useful to a wider audience.

My first question is for Simon: what is the vision for "permanent comment links"?

CommonPlace --JordanCarswell, 2003/05/06 02:16 GMT
There is an updated version of the CommonPlace "skin" available for your perusal. Added page templates, printable page, slightly new interface. It has not been widely tested, but works with 0.18.0 ;-)

to-be-named role --DeanGoodmanson, 2003/05/06 03:12 GMT
"role, which is an important way to help. (What do other projects call it ?) It includes: checking for relevant past discussion, finding/opening an appropriate issue, investigating the situation fully and communicating it clearly, setting up screenshots if needed and demos whenever possible, requesting technical guidance where needed and making it easy for me or others to provide same.. etc." - SimonMichael

Facilitator, Janitor, Wrangler, Researcher, Fireman, Sentinal, Curator?

More details would be appreciated. This is an area that's hard to do, and rarely done. Mostly for technical and time reasons. Describing guidelines and workflow may help focus the traffic and reduce noise.

Personally, I'm guessing that you are talking about a product support type of role, not a WikiGardener?, and am looking forward to hearing the differences. This may even help define a basic social pattern for a project focused wiki.

Lurker first, Weed pulling gardener second, and what-not later, - DeanG

content-type in MailOut? - UTF --FlorianKonnertz, 2003/05/06 09:39 GMT
How can the content-type in MailOut?-s be set to UTF encoding, so that the mail client can display international chars properly? (Is there a property for the language or has Mailout.py be patched or another way to do it?)

content-type in MailOut? - UTF --Simon Michael, 2003/05/06 13:54 GMT
Set it in sendMailToSubscribers ?

ahhh.. --SimonMichael, 2003/05/06 23:58 GMT
nothing like a good round of persnickety issue updates. :)

Re: [IssueNo0399 add a doctype and charset to the default skins to make them ValidHTML]? status ? --PieterB, 2003/05/07 00:38 GMT

> Not sure how our latest default and zwiki_plone skins measure up here. Pieter
> maybe you could test/close this or send me your SF id if you want to check in
> some fixes.

sourceforge id is pieterb. Unfortuately I don't have much time until next month. I didn't succeed in convincing my companys management in investing time into the use of Zope portal/wiki products, so I'm affraid I only can do some zwiki'ing/zope/pythoning in my free time.

Pieter

Note for Simon --2003/05/07 20:11 GMT
Simon - zwiki.org and joyful.com still don't accept (blackhole?) mail from blindglobe.net (they did from u.washington.edu). Anyway, the numbers you mention are reasonable, and I just need to work through the details here (which unfortunately, are non trivial, so no ETA), but it will hopefully happen. best, TonyRossini

Re: [IssueNo0399 add a doctype and charset to the default skins to make them ValidHTML]? status ? --Simon Michael, 2003/05/07 20:18 GMT
Ok, no worries. It's in the tracker so we'll resolve it eventually.

Note for Simon --Simon Michael, 2003/05/07 20:55 GMT
Hi Tony,


> Simon - zwiki.org and joyful.com still don't accept (blackhole?) mail
> from blindglobe.net (they did from u.washington.edu).

Bummer. :( I don't know who's blackholing your blindglobe.net - imeme ? I get the same problem with some servers refusing mail from my smtp provider (verizon).


> Anyway, the numbers you mention are reasonable, and I just need to work
> through the details here (which unfortunately, are non trivial, so no
> ETA), but it will hopefully happen.

Excellent. I'd love to make it simple for people to fund zwiki itches they need scratched quickly.

skin cleanups in progress --SimonMichael, 2003/05/08 21:22 GMT
Pardon the breakage.

... --guaka, 2003/05/08 21:30 GMT
Hi all, I'm a complete ZopeNoob?, and I want to get a ZWiki running. I have apt-get installed zope-zwiki, and am trying to understand the Zope management thingy. When I try to add a "ZWiki web" I fill in "test", "test" for Id, Title and click the Create wiki button. But then I get this message:

Error Type: EmergencyUserCannotOwn? Error Value: Objects cannot be owned by the emergency user

What am I to do?

... --guaka, 2003/05/08 21:31 GMT
Hi all, I'm a complete ZopeNoob?, and I want to get a ZWiki running. I have apt-get installed zope-zwiki, and am trying to understand the Zope management thingy. When I try to add a "ZWiki web" I fill in "test", "test" for Id, Title and click the Create wiki button. But then I get this message:

Error Type: EmergencyUserCannotOwn? Error Value: Objects cannot be owned by the emergency user

What am I to do?

... --guaka, 2003/05/08 21:32 GMT
Hmm, I guess the first thing I am to do is not press the "add a comment" button twice :) Konqueror problem?

... --guaka, 2003/05/08 22:46 GMT
I guess I found the solution:

When installing Zope you are asked to create a user account. Apparently you're not supposed to manage pages with this account. So the first thing to do is create a new user (click "acl_users"), say bertje. Then you should log in as bertje and create a ZWiki.

Zope emergency user --DeanGoodmanson, 2003/05/08 22:47 GMT
guaka: The ZopeBook should help you out.

In this case, use the emergency user to create real user(s), with yours as owner, then re-login and proceed.

NoobNoob? EmergencyUserCannotOwn? solution --guaka, 2003/05/08 22:47 GMT
I guess I found the solution:

When installing Zope you are asked to create a user account. Apparently you're not supposed to manage pages with this account. So the first thing to do is create a new user (click "acl_users"), say bertje. Then you should log in as bertje and create a ZWiki.

... --guaka, 2003/05/08 22:51 GMT
Oh, and what I mean with "Konqueror problem" is that it takes quite a while to update the page. So then I think that the page hasn't been updated yet, and I update again.

Thanks, Dean. I think this should be mentioned at the installation-howto page. At the moment I am not interested in the ins and outs of Zope, I just want to have my Wiki :)

StructuredTextSucks --2003/05/09 05:05 GMT
Here's another structured text/HTML bug that is bothering me: structured text will mark up things that occur inside a tag. For example: <strong> test </strong>.

What's the status of Structured Text and getting these kinds of bugs fixed. I see lots of complaining about it on zwiki.org, but where is it going?

StructuredTextSucks --simon, 2003/05/09 05:55 GMT
Hey, cool gif.

Structured Text is as ever - many complainers, few bug submitters, no developers. RestructuredText or something similar seems likely to replace it.

retiring old page types --simon, 2003/05/09 05:59 GMT
I'm going to retire a bunch of old page types, I think, including removing their render methods from the product. Here's how it's looking - shout if you see this causing any big problems:

        ('rstprelinkfitissue', 'RST + links + fit tests + issue'),
        ('stxprelinkdtmlfitissuehtml',
           'STX + links + DTML + fit tests + issue + HTML'),
        ('stxprelinkdtmlhtml', 'STX + links + DTML + HTML'),  # deprecated
        ('stxdtmllinkhtml',    'STX + DTML + links + HTML'),  # deprecated*
        ('dtmlstxlinkhtml',    'DTML + STX + links + HTML'),  # deprecated*
        ('stxprelinkhtml',     'Structured text + links + HTML'),
        ('stxlinkhtml',        'STX + links + HTML'),         # deprecated*
        ('stxprelink',         'Structured text + links'),
        ('stxlink',            'STX + links'),                # deprecated*
        ('wwmlprelink',        'WikiWikiWeb markup + links'),
        ('wwmllink',           'WikiWikiWeb markup + links'), # deprecated*
        ('prelinkdtmlhtml',    'links + DTML + HTML'),        # deprecated*
        ('dtmllinkhtml',       'DTML + links + HTML'),        # deprecated*
        ('prelinkhtml',        'links + HTML'),               # deprecated*
        ('linkhtml',           'links + HTML'),               # deprecated*
        ('dtmlhtml',           'DTML + HTML'),
        ('html',                'HTML'),
        ('textlink',           'text + links'),               # deprecated*
        ('plaintext',          'plain text'),
        #('issuedtml',          'old issue page type'),        # deprecated

python 1.5 too ? --simon, 2003/05/09 06:22 GMT
This might be the release that lets go of python 1.5 compatibility too.

bug: adding comment in Konqueror --guaka, 2003/05/10 01:13 GMT
When adding a comment in Konqueror (3.1.1) the page is not refreshed. This is really annoying, since you can't tell whether the comment was added or not. Hence my double messages yesterday.

bug: adding comment in Konqueror --simon, 2003/05/10 01:58 GMT
Yes that is annoying. Thanks for the report - drop it in the tracker which I have just finished sprucing up (various fixes & code cleanups; bold is now used for open bugs older than 60 days). I don't have konqueror handy though so I'll need someone to pursue this. Tweaking the action url of the comment form might help.

catalog is messed up --simon, 2003/05/10 03:34 GMT
Just ignore it.. I should have quit while I was ahead.

Doing the right thing about accented characters --Luciano Ramalho, 2003/05/11 04:47 GMT
I'm a great fan of ZWiki, but the current implementation has a major problem for Latino and European users: the mangling of accented characters in page id's. The transformation of accented characters in hexadecimal equivalents is OK from a technical viewpoint but very bad in usability terms, because it generates unfriendly URLs?, negating one important feature of Zope and most Wikis.

If most people agree with the solution proposed below, I will produce a patch implementing it in Zwiki.

PROBLEM --------------

For everybody who uses the Roman alphabet, replacing an accented character with an hex code in the URL is a very bad fix. It's much better to replace accented characters by their ASCII equivalents. For instance, these links:

     [Ação]    # Portuguese word "Action" rendered in ISO-8859-1
     [Büro]    # German word "Office" rendered in ISO-8859-1

should lead to pages called:

     Acao
     Buro

which read much better than:

     A_e7_e3o
     B_fcro

SOLUTION ----------------

Python 2.2 and later provide an easy way to find out the ASCII equivalents of any Unicode character, through the unicodedata module. Here is how:

     from unicodedata import decomposition

     def ascii_equiv(uchar):
         parts = decomposition(uchar).split()
         try:
             char_code = int(parts[0],16)
         except (IndexError, ValueError):
             return None
         if char_code < 0x80:
             return chr(char_code)

The decomposition function returns a description of how a character may be composed from other Unicode characters. For instance, the decomposition of Á is "0041 0301"; ñ decomposes as "006E 0303" and so on. If the first part of the decomposition corresponds to an ASCII character, we can use that as the ASCII equivalent of the Unicode characters.

The exceptions are: - decomposition() returns an empty string when there is no composition mapping in the Unicode database (this generates an IndexError? trying to get parts[0]?) - the decomposition string may have a composition format tag (such as <super>) as the first token (this generates a ValueError? when converting to int)

Note that the ascii_equiv function is useful not only for the ISO-8859-1 characters, but for any other Unicode character with a Roman alphabet equivalent.

With the ascii_equiv funtion, it's easy to generate translation tables for use with string.translate(), providing a very fast way of converting Ação into Acao.

-- Luciano Ramalho Hiperlógica

Doing the right thing about accented characters --Luciano Ramalho, 2003/05/11 12:36 GMT
I've edited my previous post to fix the encoding mess. Please read it online:

http://zwiki.org/GeneralDiscussion

-- Luciano

Need Hook to edit page content for PurpleNumbers --MikeMell 2003/05/12 02:52 GMT
Writing a module for PurpleNumbers is a fairly straightforward task. One key part of the task is to embed the purple numbers in the page text, but this is giving me trouble. (See the discussion on PurpleNumbers for why embedding is necessary.)

There seem to be only three cases which change the text of a page:

In each case, the entire POSTed? text needs to be scanned for purple numbers, adding them where necessary.

I'm a Zope newbie and have not been able to find where I could edit the POSTed? text before it is stored. Can any one point me to the module/method where I can add my hook? Thanks.

Doing the right thing about accented characters --AlastairBurt, 2003/05/12 11:51 GMT
Native Germans might correct me, but I think the best translation for German characters is:

I think this is what most Germans would type, given an ASCII keyboard.

-- Alastair

deployed WebSeitzWiki:ExpandingWikiWords --2003/05/12 13:55 GMT
I've been dicking around with various other things to integrate with the newer zwiki codebase, but I got impatient, so I just spent an hour implementing ExpandingWikiWords? on top of my live ancient zwiki installation. It's now working. I'd love some feedback on how this makes things look (esp to a WikiNewbie?).

converting to RST --2003/05/12 16:55 GMT
How can I convert a small ZWiki web (in a plone site) to use RST? I'm using a snapshot of the ZWiki CVS HEAD taken within the hour. When I edit a page I get a warning about 'this page's type (stxprelinkhtml) is not one of the allowed types for this wiki'. When I save my changes the page views correctly and when I edit again I get the same warning, meaning (I think) that the page type did not change. How can I change the page type? And how do I set the default page type for a folder/web? Is that just a matter of setting the allowed_page_types property of the wiki folder?

exception during subscription --2003/05/12 17:05 GMT
I get an exception when I try to subscribe to this page:

  File "Python expression "request.RESPONSE.setCookie('email',request.email,path='/',expires=(ZopeTime() + 365).rfc822())"", line 2, in f
NameError: global name 'ZopeTime' is not defined

deployed WebSeitzWiki:ExpandingWikiWords --Simon Michael, 2003/05/12 19:52 GMT
Looks good! Except for the Py Thon type names. Zwiki will support this when _renderLink gets cleaned up.

catalog is messed up --Simon Michael, 2003/05/12 19:55 GMT
Ah! I made lastEditTime return the creation time by accident. Fixed. Also, in case it affects site performance: reduced the subtransaction threshold from 10000 to 1000,

Actually I'm still trying to get a catalog update to complete without hitting our memory quota and restarting zope. Will set it at 500 and try again. This makes the site quite unresponsive for ~10 minutes, sorry.

converting to RST --Simon Michael, 2003/05/12 20:13 GMT
A standard_page_type property controls the type of new pages. allowed_page_types in CVS isn't yet documented except in ChangeLog?; it's this that causes the warning. Just go ahead and save as RST. You'll have to adapt the formatting manually (the headings, mainly). Any html or dtml will get quoted. You'll need to convert any zwiki comment headings to the non-html style used on RestructuredText.

catalog is messed up --Simon Michael, 2003/05/12 20:16 GMT
RecentChanges? looks more up to date, but I actually can't tell whether these long catalog updates complete successfully. This is not very good, tips welcome.

exception during subscription --Simon Michael, 2003/05/12 20:36 GMT

> I get an exception when I try to subscribe to this page::

Weird, it worked for me just now.

Or did you mean on your wiki ? #246 UserOptions syntax error after upgrade might help, or if you're running zope 2.6.0 check #345 clicking full/simple/minimal or subscribing gives a Name Error with zope 2.6.0

Doing the right thing about accented characters --Simon Michael, 2003/05/12 20:42 GMT

> For everybody who uses the Roman alphabet, replacing an accented=20
> character with an hex code in the URL is a very bad fix. It's much=20
> better to replace accented characters by their ASCII equivalents. For=20
> instance, these links::

Hi Luciano. I felt that users would most often want accented characters to be significant for linking, since they may change the meaning of a word. You think a straight conversion to the nearest unaccented character is a more useful behaviour ?


> Python 2.2 and later provide an easy way to find out the ASCII=20
> equivalents of any Unicode character, through the unicodedata module.=20
> Here is how::

This looks good, but zwiki currently supports python 1.5.2 and will need to support at least 2.1 for some time. I'd keep the simple cases in a hard-coded translation table. It may need to change based on locale, eg Alastair's rules for german. (But are these differences in zwiki behaviour from site to site going to lead to confusion ?)

canonicalIdFrom is the place to experiment.

You probably know this, but an alternative is to hack zope to allow accented characters in ids (see Regexps.py).

=20's, comment form width --SimonMichael, 2003/05/12 20:53 GMT
By the way, does anyone know why =20's appear in Luciano's mail ? We have an open issue for this, mailin.py needs to do something sensible here.

Does anyone know why this site's skins/default/wikipage.pt comment form is too wide lately ? Is it the subject input field ? Can we make it use the same small font as the textarea ?

Re: [IssueNo0514? Bad check for ExternalEditor in "standard_wiki_footer.dtml"] --Simon Michael, 2003/05/12 20:58 GMT
This is just for debugging purposes. So, it works when you remove the AUTHENTICATED_USER check ? Any idea why that might be ?

=20 --DeanGoodmanson, 2003/05/12 21:12 GMT

>> By the way, does anyone know why =20's appear in Luciano's mail ?

That's a common sight among Mailman based lists.. Not sure what causes it.

Another example: http://mail.python.org/pipermail/python-list/2002-September/121710.html

Doing the right thing about accented characters --Luciano Ramalho, 2003/05/13 02:20 GMT
Hello Alastair! As you know, I speak Portuguese. But I chose German as the second language in my examples because I wanted to trigger this discussion about the umlauts and the esset.

Adding the translations you propose complicates things quite a bit, because we wouldn't have a single way of translating those accented chars, and we can't use just the Unicode data anymore. (In Portuguese, for instance, umlauts are very rare, and a recent spelling reform has dropped most uses of them, so now ignoring the umlaut is the right thing to do even according to the dictionary, in many cases!)

I'd also like to hear from some German-speakers on this issue. While we wait for their opinion, I insist that Buro is a better rendering than B_fcro, even though Buero would be better for them.

Of course, I'm open to a solution that takes these and other special needs into account, as long we manage to keep it reasonably simple. Let's hear it from Germans, French, Swedes, Poles, Turks and other users of the Roman alphabet and variations thereof.

-- Luciano

approving some old mails.. --SimonMichael, 2003/05/13 02:27 GMT
that were held due to IssueNo0504?:

dtml to list a page's children? --Mike Beaton, 2003/05/13 02:30 GMT
I note that on the PeopleIndex page there is this question:

What is the answer?! Is there any way to do this? That I can cut and paste from somewhere? I'm finding ZWiki very useful at the moment for organising my personal research notes(!!), but I haven't got time to get into the code. ;-(

CVSBackend? --PieterB, 2003/05/13 02:31 GMT

> What benefits are you hoping to get from a CVS backend?

I'm currently working on a project which uses CVS as main repository. Adding a Zope database is no option (maintainability). I'm also looking at the possibility to use wget in a cron-script and to check those pages into CVS, but it gives me a bad feeling.

Main reason is dataformat (accessibility in future), so data should be stored either in txt or html.

Pieter

Doing the right thing about accented characters --Luciano Ramalho, 2003/05/13 02:43 GMT

On segunda-feira, mai 12, 2003, at 17:43 America/Sao_Paulo, Simon Michael wrote:


> Hi Luciano. I felt that users would most often want accented
> characters to
> be significant for linking, since they may change the meaning of a
> word.
> You think a straight conversion to the nearest unaccented character is
> a
> more useful behaviour ?

A missing accent changing the meaning of a word does happen, but is not so common, and I don't think justifies the nuisance of unreadable URLs?.


> This looks good, but zwiki currently supports python 1.5.2 and will
> need
> to support at least 2.1 for some time. I'd keep the simple cases in a
> hard-coded translation table.

I agree. My idea was just showing there is a a "canonical" way of generating ASCII renderings of accented chars. I'd use Python 2.2 to genereate a table for use in earlier Python versions.


> It may need to change based on locale, eg
> Alastair's rules for german.

Alastair's rules do make sense. They also complicate the solution considerably, by adding the requirements that (1) we change some chararacters by two (2) we add a different conversion rule according to locale and (3) we adopt rules that are not encoded in the Unicode database or in any standard Python module that I know of (please correct me if I am wrong).

Considering (1), (2), (3), and the fact that I believe that Buro is still better than B_fcro, I'd go ahead and implement my naive ASCII conversion, unless we get some strong oppostion from other users of accented Roman characters who may disagree.

Anyway, the ASCII convertion should be implemented with an easy configuration switch, and keeping current behaviour the default at least for some time.


> (But are these differences in zwiki behaviour
> from site to site going to lead to confusion ?)

I don't think so, because only very few of us will regularly contribute to Wikis in two different accented languages (most of us will usually write in English and at most one other European language), and the behaviour in question will only be noticeable once in a while.


> canonicalIdFrom is the place to experiment.

Thanks a lot for the pointer!


> You probably know this, but an alternative is to hack zope to allow
> accented characters in ids (see Regexps.py).

Yes, I know. But that does'nt solve the ugly URLs? problems. Instead of Ac_e7_e3o we get A%E7%E3o. Accented URLs? are really non-standard, and you are right in avoiding them in the first place.

-- Luciano

AddZwikiWebForm? --simon, 2003/05/13 04:53 GMT
I mocked up a better AddZwikiWebForm?. It needs some more organization and then the programming to make it work.

dtml to list a page's children? --DeanGoodmanson, 2003/05/13 13:57 GMT
I added a more detailed request for this over at DTMLCookbookDiscussion.

in case you missed the quote.. --simon, 2003/05/13 16:28 GMT
Another fine Paul Graham essay: Hackers and Painters

Doing the right thing about accented characters --Simon Michael, 2003/05/13 16:30 GMT
Luciano Ramalho <luciano@hiper.com.br> writes:
> A missing accent changing the meaning of a word does happen, but is not
> so common, and I don't think justifies the nuisance of unreadable URLs?.

Well this is good news. Unless someone speaks up with counter examples, let's do it. We should consider Alastair's idea separately (later).


> Anyway, the ASCII convertion should be implemented with an easy
> configuration switch, and keeping current behaviour the default at least
> for some time.

If we can get away without another configuration switch I'd prefer that. Also I'd be inclined to switch the behaviour asap, we can always reconsider if there's an outcry. We do need to figure out the upgrade/migration strategy, but if we're lucky the present upgradeAll/upgradeId code already has it covered.

Thanks for working on this. i18n is the last big item needed for a 1.0 release! (gasp :)

7 day pack --simon, 2003/05/14 16:45 GMT
850 (ouch) -> 360 Mb.

HTML and DTML in RestructuredText pages --DeanGoodmanson, 2003/05/14 18:10 GMT
Is there any hope for this?

If so, is there anything community members can do to help this effort?

STX may be a annoying, but I expect that at least 80% of the issues with it are worked around with HTML and (thankfully) the Zwiki patches.

hackers and painters --DeanGoodmanson, 2003/05/14 18:26 GMT
Tim O'Reilly followup: http://www.oreillynet.com/pub/wlg/3190

Doing the right thing about accented characters --Luciano Ramalho, 2003/05/15 06:05 GMT
I've finished the first phase of my research about converting accented chars to ASCII equivalents. Here you'll find conversion tables for all the ISO-8859 Latin alphabets:

http://www.ramalho.org/luciano/python/ascii_equiv/

Next, I will publish the Python 2.1 scripts I used to produce them, and then start looking into integrating the string.translate tables I generated with the id-producing code of Zwiki.

-- Luciano

plone related error --FlorianKonnertz, 2003/05/15 08:12 GMT
Hi all! I get the typical Zwiki-in-Plone error as described on the issue page but i am actually not adding the wikiweb in a plone site but just in the root folder. Any ideas? Z.2.6.1, ZWiki-0.18 ??

Doing the right thing about accented characters --Simon Michael, 2003/05/15 18:53 GMT
Nice tables!

WYSIWYG editor for Zope --2003/05/16 20:23 GMT
Any use for something like this in Zwiki?

Epoz - a cross-browser-wysiwyg-editor for Zope

re: WYSIWYG editor for Zope --DeanGoodmanson, 2003/05/16 20:30 GMT
Couple of notes:

The HTML it produced wasn't ugly (no whitespace) but it wasn't necesarily pretty (lots of carriage returns, no tabs) and I did not check to see how it handled round trip of mixed plain text&stx and HTML.

The browser requirements are pretty high. It didn't work in my 0.4 version of Phoenix.

Might make an interesting Comment box, or specific text entry box (blog entry, faq q/a, etc.). I'd like to hear of experiences using it in the edit page.

It has been raining this weekend ;) --PieterB, 2003/05/18 21:44 GMT
I've updated the ZopeRoadmap? and I created a page on ZopeContinuousIntegration? and I'm playing around creating a ZopeSandBox?.

It has been raining this weekend ;) --SimonMichael, 2003/05/19 04:12 GMT
ZopeRoadmap? is great. I saved the search forms from the old one and deleted it.

I've finally got rid of the old todo list pages, and started HowToHelp.

Dean thanks for the epoz review.

waiting for next episode --SimonMichael, 2003/05/19 07:31 GMT
When will WikiWiki:ChangeYourOrganizationDiaryPartTwo be updated ?

ZopeContinuousIntegration? --DeanGoodmanson, 2003/05/19 14:06 GMT
Python version of Gump?: http://cvs.apache.org/viewcvs/jakarta-gump/python/gump/ , seen at http://www.intertwingly.net/blog/1374.html

taking remote image URLs? and making img tag for thumbnail plus popup --BillSeitz 2003/05/19 22:59 GMT
I've wanted to easily inline images in pages without resort to HTML tags. Tweaked the wikilink code :

 # if it's an ordinary url, link to it
        if re.match(url,m):
            # except, if preceded by " or = it should probably be left alone
            if re.match('^["=]',m):                                   # "
                return m
            else: #Seitz 
                from mimetypes import guess_type
                try:
                  if guess_type(m)[0][0:5] == 'image':
                    return '<a href="%s" target="img"><img src="%s" height="72" alt="remote image" align="right"></a><br clear>' % (m, m)
                  else:
                    return '<a href="%s">%s</a>' % (m, m)                    
                except:
                  return '<a href="%s">%s</a>' % (m, m)

image followup --BillSeitz 2003/05/19 23:02 GMT
Some of that is wrapped in a try because sometimes the guess_type returns (None, None)

(Let me know if I've missed some basic reason not to do this.)

(I'm always out of synch with CVS, etc., so this is my way of returning code changes! :) )

Add Summary metatype to Zwiki? --JordanCarswell, 2003/05/20 05:06 GMT
I've been playing around with BlogFace and it seems like an interesting idea to get it to work with Zwiki. However (you knew that was coming, right ?), BlogFace creates entries by looking for a metatype from a ZCatalog and the only place I can see to grab text from Zwiki pages is from text. This makes entries extremely long for large pages. Seems like there ought to be a metatype called summary for Zwiki pages that encapsulated the content for a particular page. Wouldn't that be useful in other places besides BlogFace as well?

How to rewrite wiki urls --2003/05/20 07:29 GMT
How do i alias http://foo.com/foowiki to http://wiki.foo.com/? IE make http://wiki.foo.com pretend to be http://foo.com/foowiki IE make the foowiki the site root, instead of a subfolder ie the browser should never know about the existance of foowiki I tried creating a siteroot in zope management, but it don't work! I'm using apache to rewrite wiki.foo.com urls to the zope cgi HELP?

half there --2003/05/20 08:15 GMT
adding a siteroot in /foowiki, with path /, caused requests to wiki.foo.com/ to go to /FrontPage (good sign), but zope says Site Error. An error was encountered while publishing this resource. Resource not found. Sorry, the requested resource does not exist. Check the URL and try again. Resource: FrontPage GET

How do I recompile the file parents.py --2003/05/20 08:17 GMT
I need to "make" a new Parents.pyc without the big big big big used to emphasis the "Backlink". When I try to compile the file I get an Import Error of "Persistent" : there are no file of this kind in Zope's distrib for Windows. By any chance does anyone could tell me how to "make" a new Zwiki Product after I altered the source code ? Thanks in advance.

looking for the source --DeanGoodmanson, 2003/05/20 14:36 GMT
Sam Ruby pointed out a great quote ( http://www.intertwingly.net/blog/1419.html ), which I can't seem to get to the rest of the article:

" The creative anarchy of the wiki is the philosophical inverse of conventional corporate groupware software. Groupware's highly structured rules and processes do not always reflect the way people really work. Employees often ignore costly corporate-sanctioned software and revert to informal social networks whether simply e-mail or impromptu water-cooler discussions." - Amy Cortese

Anyone know where I can find it?

re: looking for the source --DeanGoodmanson, 2003/05/20 16:11 GMT
Found it, thanks . http://www.nytimes.com/2003/05/19/technology/19NECO.html

re: Compiling --BillSeitz, 2003/05/20 18:50 GMT
You don't need to "compile" a new .pyc - just put the .py file in the correct folder, and then either "refresh" the product or "restart" the entire Zope service. The mismatched dates between the .py and .pyc will be noticed and a new .pyc will get built/used automatically.

ZwikiPlonesSkin? and metal --FlorianKonnertz, 2003/05/21 11:25 GMT
I just started with ZwikiAndPlone and am checking the zwiki_plone skin right now. (I want to add the PageManagementPanel? to the footer.) - I wondered why the wikipage.pt doesn't use a macro with slots, any ideas? Forgive me if this idea is overkill as i am still a newbie to "metal" but i thought it would be nice to have just one template with many slots which are filled with the control panels or default mesages if permission or the UserOptions? cookie is missing. - Is anybody interested? Maybe I'll create one.

FAQ --DeanGoodmanson, 2003/05/21 15:41 GMT
The faq anchor links weren't working due to the links were ending with a space then question mark " ?" . I did a blind replace and that seems to have fixed the links. Wanted to warn you before the next pack in case you need to rollback.

Are there any plans to add a a question form back into this page? I couldn't find the original code. My pages "TopicFaqTest? and SimpleFaqTestBroke?" are based on versions of the original, but as you can see, contain killer flaws.

The auto-generation of the TOC is a nice addition.

ZWiki on Plone has no publish status? --MikeMell, 2003/05/25 16:26 GMT
I've installed ZWiki on Plone, but when I create a new wiki, it has null publish status. Attempts to set an Effective date result in error message "You must select a publishing action." Any tips on a fix or workaround?

NewsIsWiki? --JordanCarswell, 2003/05/26 00:11 GMT
Interesting idea from News is Free:

NewsIsWiki? is an experiment to see how news items can be organized, cross linked, categories, commented on, etc. On NewsIsFree?, the wiki logo will let you post any item to this wiki. All you need to do is provide a WikiWord?, edit the text to give a chance for links to appear, and you're done.

http://wiki.newsisfree.com/index.php

Re: ZWiki on Plone has no publish status? --PieterB, 2003/05/27 09:06 GMT
Somehow my reply on Zwiki on Plone has no publish status by email didn't appear on this page. Simon, should this be classified as a bug?

See http://zwiki.org/pipermail/zwiki/2003-May/001821.html

Repeating it here... (trhough the web this time)


> I've installed ZWiki on Plone, but when I create a new wiki, it has null
> publish status.
> Attempts to set an Effective date result in error message "You must select a
> publishing action."
> Any tips on a fix or workaround?

Which version of Plone and Zwiki are you running? Did you read ZwikiAndPlone. Have you tried using Plone 1.0.2 (see http://collective.sf.net/) and latest Zwiki?

Can you explain "has null publish status"?

PieterB

Re: ZWiki on Plone has no publish status? --PieterB, 2003/05/27 09:11 GMT
Plone 1.0.2 testing can be downloaded from http://sourceforge.net/projects/plone
The Zope/CMF/Plone collective can be found at http://sourceforge.net/projects/collective

Re: ZWiki on Plone has no publish status? --MikeMell, 2003/05/27 23:03 GMT
Looks like the problem was one of my own creation: I've added a new module to ZWiki in order to implement PurpleNumbers (ZWikiAndPurpleNumbers?). My implementation of the AccessControl? and Permissions modules was a little off.

bad dtml --DeanGoodmanson, 2003/05/28 15:03 GMT
I didn't think it was always this way, but I may be mistaken.

I'm familiar with the ZMI not saving a new dtml method/document when there is a parse error. I didn't think Zwiki did this. This morning I found that was the case in my zwiki also. I prefer that the document get saved and then won't render, so the revision history is available...

connectivity troubles --2003/05/28 15:31 GMT
Hi all.. I am in ireland for six weeks and my usual dial-up procedure is not working.. a bad card modem cable or something. The 0.19 release will come out as soon as I can get this resolved and get my laptop online. --Simon

last_log questions... --DeanG, 2003/05/28 20:50 GMT
In my local RecentChanges? with a catalog, the last_log field doesn't show up. But in some brute force search snippets it works just fine. It also seems to be working here at Zwiki.

Is this not showing up becuase a catalog search only shows vars that are indexed by teh catalog, and thus this var is missing in the catalog search?

workaround for AUTHENTICATE_USER == "Anonymous User" bug? --BillSeitz, 2003/05/28 21:18 GMT
I can't seem to get any pieces of forms working on my local ZWiki that depend on checking roles. (For instance: the manage-page panel)

I can see them here on zwiki.org fine. (Although when I spawned a BillSeitzTempPage? and tried to rename it, I got a Memory Error!)

canonicalUrl choices --BillSeitz, 2003/05/29 03:35 GMT
I can understand getting rid of special chars other than underscore from ids, since that's consistent with Python variable naming.

But why (a) get rid of underscores, and (b) push initial capitalization?

canonicalUrl choices --DeanGoodmanson, 2003/05/29 14:11 GMT

>> why .. (b) push initial capitalization?

This has been brought up before, and the only thing I can remember off-hand is that it helps avoid dtml & method vs. Zwiki page object collisions.

Regarding underscores, wouldy you mind explaning a bit why they're useful? Do they collide with StructuredText underscoring?

canonicalUrl choices - underscores --2003/05/29 16:08 GMT
Well, I'm thinking mainly of some sort of separator within a datestamp portion of an id. Just munging together 8 digits (20030529 vs 2003_05_29) seems atrocious to me. Hmm, 2003x05x29?

last_log update --DeanGoodmanson, 2003/05/29 19:23 GMT
By adding last_log to my catalog as a FieldIndex? Index and metadata, hte last_log started showing up in my RecentChanges?.

canonicalUrl choices - almost FreeLink? --2003/05/30 04:08 GMT
It seems weird to me that BillSeitzTempPage?, [billSeitzTempPage]?, and [bill Seitz Temp Page]? all end up auto-linking to the same page. What [billseitztemppage]? and [bill seitz temp page]??

canonicalUrl - migration of old pages? --BillSeitz, 2003/05/30 04:17 GMT
Is there any sneaky code someplace to re-id old pages to their canonical form?

Also, note that you can't sort on title_or_id() it seems to me (testing on my laptop). (Maybe you can get it to work with the Catalog - I'm doing it without a Catalog.)

Viewing Recently Added pages. --DeanGoodmanson, 2003/05/30 15:59 GMT
a view that has been helpful for managing a wiki has been a "Recent New Pages" list.

this is wonderfully easy by sorting on creation_time. RecentChanges? and the RSSDiscussion code can simply change:

 <dtml-in "Catalog(sort_on='lastEditTime' ...

to:

 <dtml-in "Catalog(sort_on='creation_time' ...

And the appropriate UI descriptions.

Quick question: Which is more effecient to restrict the length of a dtml-in: A size=# or python slice [:#]? ?

canonicalIdFrom hack for blogbits --BillSeitz, 2003/05/31 22:50 GMT
I tweaked the method code to return-without-change a name matching:

 blogid     = r'(?L)\bz[%s][%s][%s][%s]-[%s][%s]-[%s][%s]-[%s]+' % (N,N,N,N,N,N,N,N,U+L+N)

so it will leave alone a name/id that looks like z2003-05-31-TweakingCanonicalIdFrom - any spaces, any special chars other than the num-separating hyphens, it fails to match and falls back to the standard code.

I may change my mind about this later, if I discover something, or if someone tells me about a nasty bomb that I haven't run into yet.