Feel free to enter here the bugs you just found.
Make a new root
I created a new page, originally the child of my root. I decided to click on "make root" to test this feature. It is now the root - good (although there was an error reported when I did this - "clearall" function does not exist). However, I cannot make the original page the root again now (no "make root" button).
Any way around this?
Ahh - found it and fixed it - it was a bug in wiki.classes.php
You need to amend the line:
$cacheLite->c l e a n A l l (); (no spaces)
to read:
global $cacheLite;
$cacheLite->w c c l e a n A l l (); (no spaces)
(sorry spam filter will not let me post this)
David : thanks. Corrected. Yes spamSimpleFilter rejects the content because wcClean All contains a n a l
Easy to get an XML parsing error if you include an ampersand symbol
WiClear needs to change ampersands before storing them in the SQL database - change to & Particularly if they are placed in a section title you will get an XML parsing error when you try to view the page.
David : please provide an example (even if that breaks XHTML validity of this page)
Try this & cry out
Hmm - does not affect this one, but it does on my wiki - http://www.rwapadventures.com/ql_wiki
In fact I am having problems with a few XML parsing errors - see the ql_wiki at the moment.
Is it the fact I am running on a Linux server?
David : I've browse several pages without encountering any XML validation issue. Can you point me with a link to a page that has this problem please ?
Ok try: Sandbox on QL Wiki
David : OK, reproduced. This is because you are using the "allow html" option in your preferences. (well, I reproduced it in this mode, and it works ok without it. Like in this site, so I guess this is the likely explanation). If you allow html content and your html don't validate you will get that error because wiclear is served as xhtml. So your html must validates or xhtml capable browser like firefox will report those kind of errors. Perhaps in that case, wiclear should be modified to serve the page as html instead of xhtml...
Rich : Or perhaps wiclear could store ampersands and other special characters in a different way to overcome this. Unfortunately, even using
REDIRECT TO: http://wiclear.free.fr/index.php?title=%28the%20html%20ampersand%29%20instead%20of&from=Bugs
does not pass the test at present.
David: I've modified wiclear to be served as html when allow html content is used. It prevents the page from not validating in FF. You can then use & now...
Also, I see my version of WiClear lacks the controls boxes for editing text and also sub nodes is not working - am I missing something on the install?
David : for your nodes, I guess this is you who entered the page Tools (especially Nodes) not working. This is already fixed.
Rich : Yes thankyou - it was also down to me running an early MySQL (pre v4.1).
David : for the toolbar I don't know, do you have the plugin SimpleToolBar installed ? If you have it, and if you have created a custom theme, have you left the call to simpleToolbarSetup(); in prepend.inc.php in your theme ? As your page are not opened for edition, I cannot have a look to the javascript during edition to see if I can find something...
Rich: You need to register to edit. However we based the theme on WiClear - that does not include a prepend.inc.php reference or even file in the theme folder. Only Dokuwiki does that. I obviously need to add this to my theme.
MySql charset problem during install
Alas, I do not know how to resolve this one ?
- how can I detect if MySql supports charset or not !
- how do I know which charset is installed ?
- how can I change my sql requests to include or not charset information ?
David : should be fixed in latest snapshot. Install program has been modified to list available charsets when first connecting to database. If none are available, then installation proceed without using them. This needs to be polished but should do the trick. Nice side effect : it is now possible to choose charset at install, instead of having to use the UTF-8 converter later on.
All other reproductible and known bugs resolved
ACLS possible improvement
display the list of sub nodes of a particular node that do not have the same acl than current node.
Hide sub nodes if people doesn't have the right to see them
I think there is no use showing nodes that cannot be read.
David (03/02/07) : done (was possible after all)
Inherit ACL from parent node
Pages should have the same ACL as the parent node when creating/attaching a page to a node.
I think it should be great to automatically recurse apply ACL of the new parent node to the node (and it's sub nodes).
David (03/01/07) : done
Show and respect ACL in the site plan
At least show if a page is readable/writeable by anonymous. It will help detecting possible problem in the wiki. Pages that cannot be read, should not be shown in the plan.
David (03/02/07) : also done (was not that hard finally). Read protected pages are shown with a non clickable flag and a lock icon right aside
Button preview/publish/cancel should be at the same level
They act on the same part of the form. They should be at the same level. (it is UI things)
David (02/27/07) : done
Add a link wiki syntax help
Replace the wiki syntax help at the end of the edit page by a link which pops a new window. Because it is useless to have this explanation in most of the case and in other case, it is more useful to have it in a separate window that you can look into when you need it. In the current situation you must go to the end of the page and then go back to the textarea...
David (02/27/07) : done
Notification mail contains HTML entities
I am notified with �
David (02/27/07) : done
Remove the content at the beginning of the page
At least when people hit on edit, they should not see the content at the beginning of the page. It is long to scroll through the entire content (which people just have read before).
However, the content should appear again when people hit the "preview" button...
David (02/27/07) : done
Separate the parent/sub node tree functions
I have rewritten this function and the subnode function to get a simple li list not containing the current node. It helps me building my theme.
Patch is available for this
David (03/01/07) : waiting for patch
Some russi an symbols gets corrupted in v.1.0
Hi, i using php 5 and MySQL 5.2 I successfully installed Wiclear but found, thats anytime i using "ш" symbol in content or title it gets corrupted and showing as ?? I viewed my DB (in HEX mode) and found, that Wiclear write "D13F" bytes instead "D188" in DB I checked - my tables is in utf8_general_ci collation Moreover - i have no troubles with any other russi an symbols and have no troubles with Drupal (also UTF8) installation in same DB Interesting thing: If i type line of "ш" symbols and push preview button - page updates and i see valid result - row of "ш", but if i push Submit button - symbols gets corrupted.
Errors in the NodeParents results.
I'm using the content id's here.
1 is parent of 6, which is parent of 7, According to the website, siteplan and MYSQL database : wc_content_hierarchy
When I get on the 6 content, I get this result from the getNodeParents : you are here : 1 > 6 But when I get on the 7 content, I get this : you are here 1 > 44(?) > 6 > 7
I try to move this page to another parent, this is working, when I put the 6 parent back, I still get the 1 > 44 > 6 > 7 sheme...
I tryied to figure how the request was working, and maybe the problem comes from the n_left and n_right parameters ... They were crossing between the 44 content and the 6 ...
I changed ( by hand ) the n_right value of 44 to the higher value used by its children, and this seems to work again. how comes that it got a higher value, then included the other contents in it ?
Thank you in advance
