Before adding bug reports, please check TODO List to make sure they've not already been reported.
Well as much as I can see I have the same problems and are currently trying to solve them. I found some other Problem while using a multi-threaded SBCL but it seems I'm not yet able to subscribe to the proper mailing list. (Friedrich)
Please add new bugs at the bottom. A name and a timestamp on each would be appreciated
20030305;jshrager@stanford.edu;http://nostoc.stanford.edu:8000/cliki/index -- I've run a number of times into a very damaging but hard to reproduce bug wherein when you press SAVE to submit an edited page, what actually gets to the CLiki appears to be NIL. I.e., you lose everything, including your recent edits; the "new" page appears to be completely empty, and if you re-edit you get the Describe ...? thing, as though the page was completely deleted. (As it turns out, you can go BACK a couple of times, assuming that you notice the problem, otherwise, .... well, you better be making backups often!) This appears to happen more often when the page contains a lot of complex HTML, but I've seen it as well in simpler pages. I've sent what is a reproducable case (at least on our installation) to Dan -- for obvious reasons I don't want to put it in here! - Jeff
20030319;jshrager@stanford.edu; In further experimentation I've determined that the problem appears to be the size of the page. If it's more than ~6100 characters +- ~50, it loses. (Hard to tell the exact number because of line turns etc.)
Write a CLIKI-HOWTO that has: 1) material on how to edit Cliki (not just "style"), 2) material on how to get a Cliki up and running. How is this a bug? More to the point, what's stopping you from writing it yourself? You could even make it a cliki page, though I'd request you think of a better name tham "CLIKI-HOWTO" first Well, I would if I knew enough about how CLiki works, I truly would...This is a "bug" in the "wishlist" sense. This is a "documentation bug." "CLIKI-HOWTO" just looked "traditional" (in a GNU/Linux sort of way).
Searching for
franzresulted in 20 hits, listed on 3 pages, with 10 hits per page. The third page looked like this:
20 results found, showing results 21 to 20. Result page: Previous 1 2 3
2004-07-23
It seems that CLiki is tranlating these entities when the page is edited, but the exact place in the code where this happens eludes me. I was hoping to hack in a fix on the version of CLiki that I use. Does anyone have suggestions on where it should go?
I almost have a patch for this, see http://lemonodor.com/archives/001039.html -- John Wiseman
And for CLikis over which you have no control, see http://lemonodor.com/archives/001151.html for a Grease monkey script that fixes the handling of HTML entities -- John Wiseman
See http://lemonodor.com/archives/001079.html for a patch that adds the ability to diff different versions of a page -- John Wiseman
David Douthitt
The links to alpha.onshored.com fail. This affects the UncommonSQL page in particular, but probably others.
Here is a small patch that allows cliki to work with CMUCL.
--- orig/cliki_0.4.3/defpackage.lisp Sun Aug 10 20:16:04 2003
+++ cliki_0.4.3/defpackage.lisp Fri Mar 11 16:24:49 2005
@@ -1,4 +1,5 @@
(in-package :CL-USER)
+(require :gray-streams)
(defpackage "CLIKI"
(:export cliki-view-handler cliki-edit-handler cliki-instance cliki-view
cliki-url-root
@@ -21,7 +22,7 @@
(:import-from #+sbcl "SB-GRAY" #+cmu "EXT"
"FUNDAMENTAL-CHARACTER-OUTPUT-STREAM")
- (:use "NET.TELENT.DATE" "COMMON-LISP" "SB-GRAY" "ARANEIDA" "SB-BSD-SOCKETS"))
+ (:use "NET.TELENT.DATE" "COMMON-LISP" #-cmu "SB-GRAY" "ARANEIDA" #-cmu "SB-BSD-SOCKETS"))
;;; language for inline searches and stuff
(defpackage "CLIKI-PAGES" (:use #| nil |# ))
With this change, CLiki works very nicely with CMUCL. It currently runs our site just fine.
When viewed in Safari, H3 headings on this CLiki appear larger than H2 headings.
I suggest that the stylesheet be changed to specify sizes for either all headings (H1-H6) or none of them. (It currently does for H3 and no others.)
After updating the page on Theorem Provers, I note that the text box did not expand to include the new material, nor did the version number increment. However, re-editing the page does have the new material in the box. Perhaps this is an issue with Safari?
I have noticed several occurrences of pages appearing in the Recent Changes list, and obviously having been edited, while the "Revisions:" section lists no past revisions; there will be a single non-link number which is either 1 or 0.
The most recent occurrence of this is the page Nikodemus Siivola, which got spammed. web.archive.org shows that the page had previous non-spam versions (and I restored the latest of those), but they were not available within CLiki.
I have been informed that these are "pages which were last edited before versioning was implemented", in which case the bug is specifically that the pre-versioning text is not being saved as a revision when the page becomes a page-with-versioning.
I just downloaded and installed cliki. It seems when I try to create a new page SAVE-PAGE creates a CLIKI-PAGE object with a pathname but no file yet exists at the pathname. So later when VERSION-OPEN-P is called (by CAN-SAVE-AS-VERSION-P) it throws an error because both PAGE-PATHNAME and FILE-WRITE-DATE return nil to < where a number is expected.
The cliki object has a DATA-DIRECTORY slot initialised correctly to a directory which exists and is writeable (it's already created an admin directory within it). Is there some other kind of initialisation I've missed?
--update-- I fixed this along with a couple of other problems. I'm using cliki 0.4.3 rather than CVS (telent's CVS seems to be down) so the problems are probably already fixed in CVS. If not lemme know and I can submit a patch.
Package search currently results in a pretty spare, unordered list of results. E.g. the following CLiki tag:
:(search :term (package) :title "
results in something like this:
(See ASDF-INSTALL for an example.)
I think it would be a huge improvement if the returned list were at least displayed in alphabetical order. And while we're at it, why not give a short description of the package that's automatically generated from the package's CLiki page, minion style?
| ACL-COMPAT | ACL-COMPAT is a free set of compatibility layers for parts of Franz' ACL API for several Lisp systems. |
| CL-PPCRE | CL-PPCRE is a portable, Perl-compatible regular expression library by Edi Weitz. |
| CL-WHO | CL-WHO is yet another specimen of the Lisp Markup Languages category, that is, it's a Web library to create (X)HTML from S-expressions. |
| Cliki | A free collaborative hypertext (Web) authoring program, written in Common Lisp. |
| CLSQL | CLSQL is a Database interface Library for Common Lisp that can access a number of different SQL database engines. |
| clsql-pg-introspect | clsql-pg-introspect (aka cookies1) is a package for introspecting a PostgreSQL database. |
| RT | Short for "regression testing", an older (last update 1995) Test Framework from the CMU AI Repository. |
| TBNL | TBNL is a library to create dynamic web pages. |
| etc. |
How about making the (or a) "Text Formatting" link on edit pages open a popup window containing the editing instructions? I think it's pretty handy to easily be able to have a window showing the cliki syntax in view while editing a page.
2006-01-06 LV
On windows platforms the Internet Explorer browser always has trouble with the CLiki header area that shows in steelblue, doubling the blue color down to the title of the page (screenshot) until receiving a mouseover in that area. File admin/cliki.css: #navbar { margin: 1em 0 -0.5em 0 } - the problem is in negative margin. By adjusting the file this goes away. Sample page loads cleanly on IE and with Firefox.
Note: I think this bug should be considered as hight priority - it is very easy to fix while it leaves quite unpleasant impression about Cliki and whole Common Lisp community to newcomers (from my own experience).
Note2: [2007-Feb-11] This cross-browser compatability CSS issue style bug is perhaps moot now since most Microsoft IE browsers are generally being upgraded these days to versions that display the page less catastrophically. See this image for an updated example that is just barely usable: IE version 7.0.5730.11 http://img180.imageshack.us/img180/7042/clikibugspngmz1.png
Question1: Still, if there is a CSS expert that can re-confirm the correct way to deal with this I would like to hear about it. Is there a legitimate reason for insisting on the negative value, or was it a typo from the beginning?
2006-03-14 I tried to modify the pages with my name be it Friedrich or FriedrichDominicus the edit fails with: Proxy Error The proxy server received an invalid response from an upstream server. The proxy server could not handle the request POST /edit/FriedrichDominicus. Reason: Document contains no data I have no idea why this happens. Regards Friedrich
2006-03-15
Another strange thing the following in a pre form is not properly rendered