I have a site in Django-CMS 2.4.1. I often go into the admin panel and change a page there, e.g. dragging a plugin from one placeholder to another, or just change the details of a plugin.
If I press "Publish draft" straight after this, I often get an error:
PublicIsUnmodifiable: The public instance cannot be published. Use draft.
Alternatively, sometimes I am told that I cannot submit the form as the slug is already taken (by the same page).
I have found that if I navigate away, come back, make the changes again, and press "Save and continue editing", and only then "Publish draft", it often works - but not always.
Also, I have found that I often cannot delete plugins from pages.
Has this happened to anyone else? Has my database somehow become corrupted?
Sorry I can't be more specific, but I thought I'd ask in case others have had the same problem. I am still trying to work out the pattern. The new management command cms check doesn't detect anything wrong.
One thought - I am using mptt version 0.5.5, which is higher than that loaded automatically by Django-CMS (which uses 0.5.2).
Related
The way that Page Editor handles versions has been causing issues for in one of our Sitecore solutions for a client. I've posted about them here:
Sitecore instance showing incorrect workflow state in Page Editor
Expected usage of Page Editor's Experience date
I didn't get much response on those and so far I've surmised that this is just how Sitecore works. This is less than ideal for our instance, as when publishing restrictions are set, authors don't know which version they're editing which is causing various issues for them.
I'd like to implement a solution(s) to improve the experience provide the following functionality:
Something in the Experience tab that shows the number of the version being viewed.
A button on the Experience tab that resets the Date to the default (this is not simply setting the date to the current date/time, but resetting it to act as it did before a date was manually set).
A custom button in Content Editor which allows an author to open a specific version in Page Editor...set the date automatically when it opens, I guess.
An any one give me some clues on how much of this is possible and where to start?
Thanks.
For the first bullet "Something in the Experience tab that shows the number of the version being viewed", you can add version item in Core database.
Move to "Core" DB in Desktop Mode
Navigate to "/sitecore/content/Applications/WebEdit/Ribbons/WebEdit/Experience/"
Create "Versions" item like this, http://imgur.com/bPEDm7R
Create "Compare" item under "Versions" like this, http://imgur.com/dG8dz2M
The result like this, http://imgur.com/HPu3XAL
The content author can see which version they are using and they can also compare with previous versions.
We've been noticing an increase in number of broken links on our sitecore website.
Some it is due to
User Training
Publishing Issues (linked page is not yet published)
Maybe content editor issues
It's been hard to verify some of these but sometimes the link might have the authoring page URL (which means someone didn't follow the SOP), sometimes they have a strange url like /shell/Content Editor/...
So we are trying to proactively fix these before the pages go live.
I had a couple ideas like writing a Handler that would look through all "Rich Text Fields" and looking for inconsistencies (like authoring server URL). Also using a crawler-type of validator could help us (http://home.snafu.de/tilman/xenulink.html#Description) but we were wondering what the community was doing to address this issue.
The "internal link checker" usually works well but doesn't detect many of this erroneous setups (as I believe it sees them as 'external links').
Your input is greatly appreciated!
If you have RichText fields and create internal links, your internal link looks like this in the HTML view: "a href="~/link.aspx?_id=EB3AD128E7BF4F3C9F3812F701D7724E&_z=z" and when you hover with your mouse over it, is show "/Sitecore/Shell/Controls/Rich Text Editor/" before the ~/link.aspx. This is normal behavior. This link is modified to a normal link during rendering of the RTE field.
However, be sure to use the Sitecore controls like to render these RTE fields and to render links. Also using sc:fld() in XSLT instead of sc:field() can create strange links because sc:fld is rendering the raw value of the field.
In all of the Sitecore projects I have been working on, I didn't have much issues with broken links.
There is a known bug with copy-pasting links from a rich text editor, where path info is appended to the link (generating the /shell/Content Editor/ stuff).
Sitecore have a fix for it here:
http://sdn.sitecore.net/Products/Sitecore%20V5/Sitecore%20CMS%206/ReleaseNotes/KnownIssues%20Recommended/Copying%20and%20pasting%20link%20in%20rt%20fields%20may%20break%20the%20link.aspx
I would suggest a new Validation Action added to the workflow command before the items are finalized. Then you could actively stop them from being published and give immediate feedback. If you're not using workflows, you could add a new item level validator, but those often get ignored in my experience -- too many false positives on the existing validators.
I have several Sitecore items (whose template is connected to an approval workflow) that will not go into Edit mode when clicking "Edit" from the "Review" tab. The same workflow is used elsewhere on the site successfully (the site is more than 3 years old), but clicking Edit in this case doesn't seem to affect the workflow state.
I've turned on verbose logging, which provides a lot of detail, but nothing to indicate any error conditions. I've been comparing how this template is configured with other "working" templates for a couple of days and everything looks Kosher, but surely I'm missing something obvious.
EDIT: The item(s) in question are imported by a data provider, and I've confirmed that for some reason Sitecore is not setting the "Workflow" and "State" system standard fields with a default value. That appears to be at least part of the issue.
EDIT 2: Followed the steps in this thread http://sdn.sitecore.net/Forum/ShowPost.aspx?PostID=45991 to use Sitecore Rocks to update all of the items to the correct Workflow State, but the update did not stick. For some reason workflow changes are not persisted for imported items.
EDIT 3: Clicking the Edit dropdown button under the Home tab displays the message "The item is currently not part of a workflow."
Standard values set correctly? Only have one Standard Values item for this template?
Generally, almost every time I run into a situation where X works on certain items, but not on others... it's a permissions issue. You've tested this as an Administrator I hope?
Sorry my suggestions are so vague, but it's hard to be specific here. We need some more information from you... When you click on Edit what exactly are you expecting to happen? I'm not entirely sure what you mean by "Edit mode" here. In Content Editor, all items are essentially in edit mode all the time... In my configuration, when I click on Edit, I am just locking the item. It does not change workflow state or create a new version.
These items are coming from a data provider... is your data provider handling Lock/Unlock calls properly? I know it's possible to use a data provider without implementing the entire Item API.
Is there anyway to detect if the tab app is being loading in the new timeline or not? I got clients asking to upgrade my layout but not all of them have upgraded so it makes it hard to do that. If I had a way to detect this it would be perfect.
I managed to do that by looking for an album called "Cover Photos" on the album list of that page. Here's a sample:
http://graph.facebook.com/-page_id-/albums?fields=name
You don't need a token unless the page isn't published yet.
The problem with this approach is that some pages still don't have cover picture yet, but it's the minority.
Looks like an issue was closed related to the question I'm about to ask, so I wonder if I'm going to be skating on thin ice, but here goes. I feel like even though this is kind of localized by nature, it could be a useful example for other developers dealing with the Like button.
I seem to be having issues regarding the Facebook like button. The infuriating part is that I'm pretty confident I have everything setup properly, and even though the linter says "hey, this all looks kosher!" the like button fails to get the correct content and uses cached info from a different page.
Here's the case: referlocal.com. We serve daily deals, and list them right off the homepage. There can be one of many deals listed on that homepage, so obviously Facebook is seeing just one when it goes to take a peek at the root directory. Now, on the homepage, whatever deal your viewing has a FB Like XFBML tag with href and ref attributes set. Every page that displays a deal also has OG tags pointing directly to the offer view page. Deals can also be viewed on user pages. So, on these three paths:
/
/{username}
/offers/{city}/{title-url-alias}
a like button is included as well as OG tags that point to /offers/{city}/{title-url-alias}. For about 3 weeks, the button worked like a champ. But recently, it's been performing strangely. For any deal, regardless of the 3 locations, it always uses the information cached from the previous days "/" deal. Regardless of the OG tags or origination.
Here's the wildly confusing part. The Facebook Linter is supposed to recache the information on the page, right? Well, it sort of does. It picks up on all of the appropriate information set in the OG tags, but when you click the Like button at the bottom of the page, no dice, still uses the deal FB saw on the homepage from the day before.
I know this is probably killing my argument for localization, but check this out:
http://developers.facebook.com/tools/lint/?url=http%3A%2F%2Freferlocal.com%2Foffers%2Fdallas%2Fget-20-of-authentic-italian-american-cuisine-for-only-10-at-leggios-italian-ristorante-dallas
Regardless of the fact that the linter finds all the appropriate information, the like button at the bottom still is liking a deal FB cached from the previous day.
Any ideas?
I had this problem too. The linter caches things in a very weird manner - seems like the best thing to do is either to add a query string and change it every time, or just rename the file every time you're checking.