How do I list all files ever committed to the repository?

How could I let SVN list all files which were ever committed to a repository (if possible, along with location and revision).

More exactly: I need all ever existed paths.

If someone knows how to do a fulltext search on the entire repository (including all revisions), this would help me even better.

Example

Let's say I commit a file SomeFileOne.txt, later rename it to SomeFileTwo.txt.

After performing the required task (for which solut开发者_如何学Goion I'm looking for), I should get not only SomeFileTwo.txt, but also SomeFileOne.txt.

...
/trunk/SomeProject/SomeFileOne.txt - revision 100
/trunk/SomeProject/SomeFileTwo.txt - revision 101
...

Found out something by playing around.

A slow-and-dirty solution might be:

svnadmin dump MY_REPO | grep Node-path

Output looks like:

* Dumped revision 3039.
Node-path: trunk/proj1/src/p/keywords/main/NoOp.java
Node-path: trunk/proj1/src/p/keywords/main/ScopeDecl.java
* Dumped revision 3040.
Node-path: trunk/myCommons/src/i/IConverter.java
Node-path: trunk/myCommons/src/i/ListUtils.java
* Dumped revision 3041.
Node-path: trunk/proj1/src/p/execution/impl/IterativeExecuter.java
Node-path: trunk/proj1/src/p/keywords/DeclConstant.java

Well, at least it solves my problem.

NOTE: Dumped revision ... is shown, because it's stderr.

For an improved, structured output to svn log -v, use the following bash script:

$ svn log -v | egrep '^   A' | cut -c 6- | sort | uniq

Explained:

• fetch the whole svn log of the given path
• select only lines starting with ' A', which is the svn log output for an added file.
• cut away the first 5 chars, which is the mentioned add info
• sort the output
• remove duplicate lines (e.g. if a file was readded later)

This also works for subpaths of the checkout, you don't even need a checkout at all, if you provide the repository path:

$ svn log -v https://github.com/schacon/kidgloves/ | egrep '^   A' | cut -c 6- | sort | uniq

Result:

/branches
/branches/testing (from /trunk:17)
/trunk
/trunk/README
/trunk/example.rb
/trunk/kidgloves.rb
/trunk/simple_rack.rb

Right, the file list may not be perfect for automated processing yet, but it goes a long way from just the log output, if I may say so. Enjoy.

To me, svn log -v would have been a more intuitive method - or have you already tried that ? Is it possible to share the need for requiring such an extensive report on all the file commits ever made in the repository?

If you just want to browse around you might get what you want with one of the svn web access tools. A roundup was written up over here.