Windows file:/// URL format for svnsync

[David Goldsmith]

Hi!  I can't seem to get the formatting for my source repository name--which is a Windows directory--correct for svnsync: I've tried forward slashes and backslashes, quotes and no quotes, relative path and absolute path--nada.  My repository, in Windows syntax, is C:\MWDM--how do I specify this as part of the source argument to svnsync?  Thanks,

David Goldsmith

Olympia, WA

[Ben Reser]

file:///C:/MWDM

Note that there are 3 forward slashes before the path because you want a blank
host entry.

http://svnbook.red-bean.com/en/1.7/svn.basic.in-action.html#svn.advanced.reposurls

[David Goldsmith]

Hi, Ben, and thanks for your reply.  Unfortunately, I think that was the first (or perhaps the second) thing I tried, to no avail (also unfortunately, I'm away from my work computer for the rest of the day so I can't check my command history or the error message it generated).  

As I think about my sitch, I realize that the folder I thought was the root of the repository probably isn't, because it's the root of the directory tree in which reside all the files that I edit day-to-day, and that's supposed to be a working copy, not the repository itself, correct? Assuming that's correct, my ultimate goal is to "relocate" my project, with history, to a new, empty Google code project (already created and reset)--how should I proceed: should I continue to try to svnsync my new Google project to my existing repository (to which I'll never again have access after tomorrow), and if so, how do I find my repository from knowing where a working copy is ('cause, clearly, I've forgotten)?  Or should I just upload my working copy from its root, and then check that out to any place else I want to be able to work on it--would such an upload include the history, and would Google Code recognize it?  Please advise/help!

Thanks,

DG

YOU!...are Big Data.

[Bert Huijben]

‘svn info WORKINGCOPY’ will tell you the url in the repository and the repository root.

 

Assuming that you have access to the entire repository you probably want to sync from the repository root to an empty repository to have a local backup. (Look in http://svnbook.red-bean.com for details on how to setup the right hooks, etc)

 

And with a bit less time pressure you can then sync that to google code.

 

                Bert

[Bob Archer]

Good point… our just use svnrdump to dump the repo… then you can take that file, load it into a repo on another machine and then sync that to google.

[Bob Archer]

Not sure how we got off list. But, I just used svnsync to move a repo to google code.. it was pretty easy. Just follow the instructions on their wiki page:

 

How do I import an existing Subversion repository?

To upload the history of an existing Subversion repository, use the svnsync tool that ships with Subversion 1.4 or later. Run svnsync help to read more about this tool.

Note that your Google Code repository must be reset to revision 0 for this to work. Your project's Source tab will display instructions on how to reset the repository yourself. (Note: you must be a project owner to reset your own repository, and also to push code up with svnsync.)

Here's a sample transcript that demonstrates how you can push history from an existing repository (located at file:///my/repos) to your repository on Google Code:

  $ svnsync init --username YOURUSERNAME https://YOURPROJECT.googlecode.com/svn file:///path/to/localrepos
  Copied properties for revision 0.
  $ svnsync sync --username YOURUSERNAME https://YOURPROJECT.googlecode.com/svn
  Committed revision 1.
  Copied properties for revision 1.
  Committed revision 2.
  Copied properties for revision 2.
  [...]

When prompted for your password, use your googlecode.com password, which can be found on the settings tab of your profile page.

Running svnsync on a large repository will take a significant amount of time. If you are disconnected during the process, you may see the error message "svnsync: Couldn't get lock on destination repos after 10 attempts". If this happens, you can remove the lock yourself; see the "Locks" section of svnsync.txt.

 

BOb

 

 

From: David Goldsmith [mailto:eulergau...@gmail.com]
Sent: Thursday, October 31, 2013 3:25 PM
To: Bob Archer
Subject: Re: Windows file:/// URL format for svnsync

 

They used to have restrictions--for security--on what kind of executable code you could upload, but I think that was (is?) just for Google docs; AFA Google Code is concerned, seeing as how it's intended as an open source code hub, they allow just about anything (at least, anything text-based)--if there are restrictions, they don't feature them prominently, so I'm not aware of them.

YOU!...are Big Data.

 

On Thu, Oct 31, 2013 at 11:23 AM, Bob Archer <Bob.A...@amsi.com> wrote:

I’ve never used google reports. Do they let you upload and import a dump file?

 

From: David Goldsmith [mailto:eulergau...@gmail.com]
Sent: Thursday, October 31, 2013 2:05 PM
To: Bob Archer

Subject: Re: Windows file:/// URL format for svnsync

 

Why wouldn't I just upload it to Google?  I guess what I'm missing is: how does SVN "recognize" history?  By path and filename?  So, if I preserve that, shouldn't that be enough to preserve the history?  Thanks again!

DG

YOU!...are Big Data.

[David Goldsmith]

OK, now we're going in circles: what you cut and paste is what I've been trying, over and over again, to no avail (as I said in my OP, it didn't like the way I'm specifying my source folder.)  I'm sure there's some subtlety that--as a user, not an admin--I'm missing, that those overly-simplified instructions aren't including, but I can't seem to communicate what it is.  I'll try posting a transcript of my efforts.

DG

YOU!...are Big Data.

[David Goldsmith]

C:\MWDM>svnsync init --username eulergaussriemann https://ecy2.googlecode.com/sv
n file:///C:/MWDM
svnsync: warning: W200007: Target server does not support atomic revision proper
ty edits; consider upgrading it to 1.7 or using an external locking program
Authentication realm: <https://ecy2.googlecode.com:443> Google Code Subversion R
epository
Password for 'eulergaussriemann':
svnsync: E180001: Unable to connect to a repository at URL 'file:///C:/MWDM'
svnsync: E180001: Unable to open an ra_local session to URL
svnsync: E180001: Unable to open repository 'file:///C:/MWDM'

YOU!...are Big Data.

[David Goldsmith]

Yup, that was it: I wasn't trying to sync my repo, I was trying to sync my wc; luckily, I'm cleaning out as much of my stuff as I can find and I found a folder on my C: drive called SVN, and inside that, a folder called Repo--when I used that for file:/// everything worked; of course, now it's uploading the whole 1210 commit history one commit at a time!  Ouch...I now see why Bert said "w/ a bit little less time pressure."  Well, I have a meeting at two: hopefully it'll run long, and if it doesn't finish by the end of that, I'll just leave my computer running as I run out the door! :-)

Thanks all!!!

DG

YOU!...are Big Data.

[Dave Huang]

On Oct 31, 2013, at 15:17, David Goldsmith <eulergau...@gmail.com> wrote:

> OK, now we're going in circles: what you cut and paste is what I've been trying, over and over again, to no avail (as I said in my OP, it didn't like the way I'm specifying my source folder.) I'm sure there's some subtlety that--as a user, not an admin--I'm missing, that those overly-simplified instructions aren't including, but I can't seem to communicate what it is. I'll try posting a transcript of my efforts.

Well, earlier, you said:

> As I think about my sitch, I realize that the folder I thought was the root of the repository probably isn't, because it's the root of the directory tree in which reside all the files that I edit day-to-day, and that's supposed to be a working copy, not the repository itself, correct?

So I take that to mean C:\MWDM is the working copy, not the repository. You need the repository URL. In response to your message, Bert Huijben gave instructions on how to find the repository URL:

> ‘svn info WORKINGCOPY’ will tell you the url in the repository and the repository root.

However, it seems like you're still using file:///C:/MWDM as the repo URL, rather than the what svn info gave.

[Branko Čibej]

On 31.10.2013 21:21, David Goldsmith wrote:
> C:\MWDM>svnsync init --username eulergaussriemann
> https://ecy2.googlecode.com/sv
> n file:///C:/MWDM
> svnsync: warning: W200007: Target server does not support atomic
> revision proper
> ty edits; consider upgrading it to 1.7 or using an external locking
> program
> Authentication realm: <https://ecy2.googlecode.com:443> Google Code
> Subversion R
> epository
> Password for 'eulergaussriemann':
> svnsync: E180001: Unable to connect to a repository at URL
> 'file:///C:/MWDM'
> svnsync: E180001: Unable to open an ra_local session to URL
> svnsync: E180001: Unable to open repository 'file:///C:/MWDM'

This looks a pretty straight-forward error message; what does

dir C:\MWDM

show you?

-- Brane

--
Branko Čibej | Director of Subversion
WANdisco // Non-Stop Data
e. br...@wandisco.com

[Bob Archer]

Are you sure you are specifying a REPOSITORY and not a working copy???

[David Goldsmith]

I'm pretty sure I am now...however, while I was away, I got:

 

Committed revision 821.
Copied properties for revision 821.

svnsync: E175002: MKACTIVITY request on '/svn/!svn/act/d9a29da4-969c-9043-8d2b-1
fe43a60c2c2' failed: 500 Internal Server Error
svnsync: E175002: Additional errors:
svnsync: E175002: Error running context
svnsync: E175002: RA layer request failed

 

If I just run the command again, is it smart enough to resume where it left off?  Is there an (undocumented) command swich I can use to resume where it left off?  Is this even the kind of error I might hope to resume where it left off?  Thanks!

 

DG

 

YOU!...are Big Data.

[Branko Čibej]

On 31.10.2013 22:57, David Goldsmith wrote:
> I'm pretty sure I am now...however, while I was away, I got:
>
> Committed revision 821.
> Copied properties for revision 821.
> svnsync: E175002: MKACTIVITY request on
> '/svn/!svn/act/d9a29da4-969c-9043-8d2b-1
> fe43a60c2c2' failed: 500 Internal Server Error
> svnsync: E175002: Additional errors:
> svnsync: E175002: Error running context
> svnsync: E175002: RA layer request failed
>
> If I just run the command again, is it smart enough to resume where it
> left off? Is there an (undocumented) command swich I can use to
> resume where it left off? Is this even the kind of error I might hope
> to resume where it left off? Thanks!

Yes, svnsync should continue from the last revision recorded in the
revprops on revision 0 on the target. From your log, that would appear
to be revision 821; but I do suggest you verify that it is, and that
it's properly recorded in the r0 revprops:

svn info <mirror-url> | grep ^Revision:
svn propget --revprop svn:sync-last-merged-rev -r0 <mirror-url>
svn propget --revprop svn:sync-currently-copying -r0 <mirror-url>
svn propget --revprop svn:sync-lock -r0 <mirror-url>

Make sure that the revision reported by "svn info" and the one in the
svn:sync-last-merged-rev property are the same. Depending on the state
in which the command failed, you may have to clean up the
svn:sync-currently-copying and svn:sync-lock properties.