How can I tag content in a confluence wiki OR store cofluence wiki in SVN

[Caroline Gordon]

I'm reading the Jez Humble book about continuous delivery and love the idea that EVERYTHING should be in source control.

However, I see a disconnect here, using confluence to build a wiki with all your requirements documents. How can I easily tag a version of the whole wiki so I know what source code it relates to?

Why would I want to do this? For all the reasons outlined in Jez's book, putting everthing in source control ensures it is tracked, controlled and bundled together. I will always know which wireframe / version of requirements that label went with.

I really like the pairing of Jira / Greenhopper with Confluence, there is some nice integration but there is still this disconnect from the artifact that really matters - the source code.

[Alejandro Conde Carrillo]

Hi Caroline.

I still do not fully understand why you want to put Confluence in SVN but may be I should just read that book.

I have never implemented anything like that so my advice is only an idea. You could use WebDAV to get a copy of Confluence in your file system and then commit that to your SVN. Everytime you want to take a snapshot you can just download and commit again.

Please note that Confluence already keeps it own versions of all pages and attachments so you get a bit of source control build in.

[Caroline Gordon]

In essence the idea of putting all your artificats, including documentation, in source control is so that you have one consistent way of understanding / seeing what your system does and how it does it.

Often in software projects this information is dispersed across a number of different systems. So, if you want a snapshot of the code and to grab all doco related to it then you have to traverse a number of systems and relate different information. In fact you will proabably find it difficult to relate things like build numbers, doc version numbers, confluence page versions, wireframe versions etc and you will probably give up.

I have solved this before by having release notes with manually filled out version numbers of specifications, when tracability is very important this kind of thing is neccessary.

Anyway ... How nice would it be to check out and get the code and ALL associated documentation artifacts that are associated with that code in one operation. If you check out a tagged / labelled verison you get the doco assocaited with that particular tag.

I had a quick read of WebDAV and I guess it's not really what I want.

One idea I had is to pdf the confluence pages each sprint and store that, but it relies on someone doing this manually which is not so great. It sounds like webdav would do a similar thing, and in fact maybe we could setup Jenkins to do that.

BTW - yes you should read the book too - it's awesome...mind you I didn't actually finish it yet ;-)