Yes, thanks Partha for the update. At least we know someone is listening.

Thanks for this reply. That gives us at least some hope and let us stay tuned - but please hurry up. Databases are growing and so many workarounds have been already implemented. Cheers.

Dear Fred, Carl and Klaus,

Thank you for your feedback.

As Klaus pointed out, we are not able to work on this issue for the 3.x branch as the new editor in 4.0 (still being developed) changes everything.
We believe the issue may be fixed as part of the 4.0 work. If not, we hope to address it soon after the 4.0 release.
It is definitely high on the bug fixing backlog.

Please accept my apologies for the inconvenience.

Kind Regards,
Partha Kamal

For us and my customers this feature is mandatory too. From my investigation about this issue it seems that Atlassian either like to drop this mandatory feature or that they can only solve this issue with 4.0 new RichText and WikiMarkup editor. I hope the later is at least the case. If I would be a programer at Atlassian I wouldn’t try to fix it in 3.x as 4.0 will be a new Editor from scratch on as I did understand. You see this missing feature on pages at Atlassian confluence documentations – they have broken links too – so they must do know it. They don’t assign this bug, don’t really answer questions about it and they don’t talk about officially. Frustrating for me is not that this bug exists – it is that they do not communicate about it and leave us in a doubt about confluence state – I always keep an I open for any system that provides this feature so far.

Hey Atlassian!

Imagine my frustration as I look at my own comment from June '08 describing how serious and BASIC this bug is ... and two years, 4 duplicate tickets, 40 pleading voters, and several Confluence versions later it hasn't even been assigned to someone to work on.

I notice by pure coincidence that it is almost 3 years to the day since this issue was first reported.

Is there anything we can do to make this a priority?

Watching this task as the one I was watching was marked as a duplicate of this.

This is still an issue in v4, and is preventing at least 1 interesting plugin - numbered-headings from functioning.

This is an issue for us when it comes to panels:

,

,

.

This deficiency has bitten me again in a bad way, this time with a user macro with an unrendered body that emits wiki format, which I thought didn't have this problem.

We're using 3.1.1.

Come one! Help!

come on Atlassian, this should be fixed!

(or at least investigated so that someone could come up with a workaround for plugin developers to implement)

I am seeing this problem in a warning macro. Very bad! Please fix it ASAP!

Thats why I use just the standard panel macro.

Using reporting plugin to get informations about links the report does not include links that are within a info, note, warning or tip macro. Links witin a panel macro are reported.

This information might be helpful to find the problem.

This bug still exists in Confluence 3.2.1_01 when a link to a renamed page is inside a {info} panel, but it works with a {panel}.

I hope that this gets fixed soon because these broken links are really annoying because IMO this is a basic functionality that should always work. For most users, this is more important than fancy gadgets etc.

Please add 2.8 and 2.8.1 to the affected versions - if you close tickets as duplicates you should take the information from the tickets into this one ...

Hi,

a new incident of amnesia:
After doing xml export import multi-excerpt plug in got amnesia on all excerpts included elsewhere. The problem with it is, that you dont see not inserted content as no error message occurs – its just not there. I had to manually open each page with source content to force rerender.

I was forced to use multi-excerpt-include instead of a combination of include and no-include because of another major insane behavior which is identification in the way of assuming something equal which issn't. Content from another space onto a new space will be fully included if both spaces have the same pagetitle.

Confluence has great potential and is about to became genius and a critic on this is only worth if it offers a constructive solution.

A solution might be an automatic rerender of all pages after xml import, as you do for indexing.

To avoid further amnesia and miss-identification I suggest to:

1. bound any linking and plug-in to nodes like pageid and spaceid
2. rerender on xml import and similar operations
3. create a database table where pageid and spaceid are stored in combination with the errormessages a plug-in generates
4. all plugins effecting content should relay on ID`s
   1. If an error occurs the plugin just generates an entry like the indexqueue.
   2. at admin level within one space I could see all entries
   3. at global admin I could see all errors on every space
   4. lists would contain links for ease of access
   5. used plugins within a page will be listed inside the page
   6. on each render the list gets first deleted and new entries are done if still errors do exist.
   7. as it would be a case that after render the page notes the plugins in use it would generate a number of new admin possibilities.

• With above solution I could even manage to get aware of errors within content which occur due to disabled plug-ins. If a plugin is able to manage a message on a page, I expect it to be possible to get this message in a list of errors.

• Why should someone sit down and search for some wrong content or errors while he has a computer who is able analyze and to know whats wrong with his data.

Maybe its a hard job to do (and may it exist already partly - but I did not found it) but you did it once with linking and indexing, so I expect the knowledge to do get this into code and database is available.

Think about it,

Cheers, Klaus

Hi all,

I got into Confluence a month ago because I thought that I could set up a large nested knowledge-base. I know now, that I was wrong, because I see issues that are of mayor interest regarding nesting that have no progress since years.

A collaborative solution is something that has its own live because it is a collaborative mind - that grows and most content will not stay the same forever - their is some evolution on this - so names of pages need to change sooner or later, to correspond to their content.

Nesting is not possible if macros dont follow changes.

A WIKI is originally a syntax based solution running on a relational database. The links in relational databases never change - you see it on the pageid. Macros for nested content should not work only using pagenames or pagetitle, they should work using pageid (the ID No. of a page). Links should do also.

The advantage: Links and Content cant break even if a page is moved or renamed.

I took a glance into the MySQL database I use for Confluence and into the tables – it looks like that each page has its own ID which is unique.

Every node you create is like a synapsis in a brain. It is a start and end point of a communication line. It is actually a terminal, which is able to send and receive communication. As things change – nodes, synapsis, terminals, pagenames, changes too. They got terminated if dont neded, they got created as some communication about something thought to be useful to have access too. Content could change too. But what never changes - unless deleted is a node inside a database that gets an ID that you call to be the pageid.

Syntax based solutions are good for Wikipedia, good for dictionaries, do searches, listing etc. – they are no good for enterprise solutions.

So why dont using pageid to keep macros and linking working with stable terminals - like relational databases do – and keep syntax based operations for web based publishing as it is now working?

Why is it all so complicated - while a easy way is known ? – maybe there is something I still dont know.

Knowledge in a human mind is nested. What you might not want is a system with amnesia which is getting more and more insane.

So…get your mind out of the code - look at that what you could achieve with Confluence - a collaborative mind that can remember because it does not loose what it did know once!

Think about it! – I mean YOU - not someone else!

Cheers,

Klaus
Copyright © 2010 Klaus Feldmann

Is there any progress on this bug? It's now been more than two years.
We still get people bitten by it.

Again, this is a Major bug! It causes links to break seemingly at random.

This continues to be a problem in 3.0.1, including (amongh others):

• {float}
  • the body contents is not updated
• {info}
  • title= parameter obscures the rename
  • interestingly, the body contents is updated

I had higher hopes for 3.x.

It's a major issue for us because user do not think about renamen reference. Every testcase page is lost in the testplan after renaming. And it's not testet than.

Yes this is a pretty major issue for us too. We are running 2.7.1
-Joseph

(Another case http://jira.atlassian.com/browse/CONF-11546 has more info about this.)

Note that this problem happens in some compiled macros as noted, and also in many user macros. With user macros, I believe that when "don't render body" is checked, the links aren't updated.

I am baffled as to why other people do not seem to be running into this. We've run into this, and it's a MAJOR problem!

It is one of the key features of Confluence that links automatically get updated when pages are moved or renamed. If some macros shield links contained in their macro bodies from this BASIC feature ... and there is no way to predict which macros do this and which do not ... then people should be running into mysteriously broken links all the time.