I have two pages that I use for inclusion in other pages:
http://bynnewalle.wikidot.com/inc:arts-table
http://bynnewalle.wikidot.com/inc:combat-table
Both consist of [[html]] blocks in which there is some javascript and some plain html code.
When I include both on the same page, for example http://bynnewalle.wikidot.com/test:two-includes, at first I see exactly what I ought to see: first the combat-table, next the arts-table. Reloading the page, reloading-with-shift to force a true reload, and closing and reopening the page all show me the correct thing. Thus, if you go to look at my example page, you will likely see the right thing. At first…
When I come back some hours later and reload or reopen the page, I see something very strange: the combat-table has been replaced with another copy of the arts-table. Reloading, reloading-with-shift, and closing-and-reopening do not restore the correct two-different-tables appearance. Editing to add a trivial change, re-previewing, and saving also does not restore correct inclusion, although the trivial change does appear. The wiki source appears correct (that is, one include of each kind); the html source has links to two different html iframe blocks — but the html source of those content blocks is exactly the same, being the arts-table (with correct substitutions from the second include) in each.
Because this does not appear immediately, perhaps it has something to do with caching that wikidot is performing?
I am running Mac OSX 10.5.8, Firefox 3.6.3.
On Windows XP with Firefox 3.6.3 I initially saw the two Arts tables. (I may have visited the site a few days ago when I read the forum thread). At the same time, Google Chrome showed the correct two tables. Clearing the FF cache corrected the problem. After around an hour reloading the page on Chrome produced the error. The file-names in the iframe tags were identical between the two browsers, and seemed the same as they were originally. Again, clearing the cache cleared the error.
A couple of hours later, refreshing every half-hour or so, neither browser showed the error.
I can confirm that clearing the FF cache seems to clear up the problem for me too… at least for a little while (several hours?), after which it comes back and I have to clear the cache again.
Just for the record I can see the bug in FF 3.0.19 under Linux: saw two arts tables, cleared cache and refreshed, saw combat and arts, left it a long time and refreshed, back to two arts tables
I thing this was connected to bug Include with variables in live template requires compilation of all pages. I have touched your pages (changed slightly and saved) - the "include" relation was not stored properly.
Right now this should be fixed.
Michał Frąckowiak @ Wikidot Inc.
Visit my blog at michalf.me
I had hoped they were connected so that fixing one fixed the other. Indeed, this problem has gone away on the test page and on another page where it appeared (after I touched the other page). It would probably be appropriate to mark this "completed" now. Thank you!
Drat. It's back. It seemed so fixed yesterday…
I would like to ask you for something. Once the problem appear, please check the HTML of the page, locate the <iframe>s in it and post them here.
Piotr Gabryjeluk
visit my blog
Start-up Firefox and see two arts tables:
(I removed the <P> tags)
Clear cache, reload, and see a combat table and an arts table:
You can see this yourself if you visit the page tonight before you leave work. It will show the error tomorrow morning.
As I recall, showing source in Firefox causes it fetch the page which could potentially "fix" the problem, so I would like you (once the problem appear for you) to select the page contents (Ctrl-A) and then right-click and choose "View Selection Source". Then please find the iframe tags and post the URLs again.
In the mean time please create a page with no includes, just put the HTML blocks (but with different content) in page content and try to recreate the bug.
Moreover I will try to reproduce the bug, but visiting the page (with Firefox 3.5.8 on Ubuntu) now and checking it tomorrow as you suggested.
Piotr Gabryjeluk
visit my blog
While seeing the issue, selecting all and viewing selection source gives me these iframes:
Both pages are Arts tables.
Just the two html blocks (no substitutions) are here:
http://bynnewalle.wikidot.com/test:two-htmls
Two html blocks where I have manually replaced include-variables with the ones that the two-includes page uses are here:
http://bynnewalle.wikidot.com/test:two-htmls-values
As I have just created them, I do not know yet whether they display the bug; I will keep checking.
These pages do seem to display the bug. So I guess it doesn't having anything to do with the include blocks — is it the html blocks? This is bizarre!
Edit: iframes from two-htmls-values while displaying the issue:
Both are the arts table.
View Selection Source from test:two-htmls while displaying problem:
And after clearing cache and not showing problem: