subscription

[Andreas Kress]

subscribe



--

Mit freundlichen Grüßen / With kind regards

Andreas Kress

HOOD Group
Office München
Keltenring 7
D-82041 Oberhaching
Germany
phone: 0049 89 45 12 53 0
fax: 0049 89 45 12 53 19
mobile: 0049 176 16 61 16 24
andreas.kr...@hood-group.com
http://www.HOOD-Group.com

More than 20 years experience in RM&E -
Requirements Management & Engineering

Events and trainings
Visit our conference websites - www.REConf.de | www.REConf.ch | www.SubConf.de
See our training dates - http://www.HOOD-Group.com/en/training/dates

The information in this e-mail (which may include files transmitted with it) is 
confidential and is intended for the addressee(s) named above only. If you are 
not named as an addressee you are to maintain confidentiality and must notify 
the author immediately, destroy any copies and delete the e-mail from your 
computer. The information in this e-mail is not to be relied upon by any person 
other than the addressee(s) except with prior written approval. If no such 
approval is given the author will not accept any liability (in negligence or 
otherwise) arising from any third party acting on such information. HOOD GmbH | 
Amtsgericht München | HRB 139 127 | Geschäftsführer: Colin Hood, Rupert Wiebel

Re: subscription

[Lorenz]

Andreas Kress wrote:
>subscribe

that would be users-subscr...@subversion.apache.org
-- 

Lorenz

RE: svn status: filter on file extension

[Niel De Clerk]

Hi
Thanks for the reply.

BTW, a good alternative to grep in Windows is findstr

The following works well (to only list *.kb files)

Svn status --verbose --show-updates |findstr /E .kb

Cheers
Niel

-Original Message-
From: C. Michael Pilato [mailto:cmpil...@collab.net] 
Sent: 19 January 2010 16:55
To: Niel De Clerk
Cc: users@subversion.apache.org
Subject: Re: svn status: filter on file extension

Niel De Clerk wrote:
> Good Day
> 
> I'd like to know if there is a way to make a SVN STATUS command 
> recursive while filtering the files using a wildcard?
> 
> I need this purely for performance reasons.  I'm guessing that the 
> amount of data that the repository need to pass back (via a rather 
> slow
> network) will be less, and also that I'll save time when processing 
> the results since I don't have to filter out lots of unneeded
information.

Actually, the amount of data currently passed over the network for 'svn
status -u' is proportional to the amount of *change* in the tree since
the working copy was last updated, not to the size of the tree itself,
and is optimized for directory comparisons as opposed to comparisons of
individual file revisions.h

> However, I'm often only interested in the files with a specific 
> extension (*.kb).  So I've tried to use
> 
> svn status --verbose --show-updates *.kb
> 
> But this only returns the .kb files found in the current directory
(i.e.
> non-recursive)
> 
> Your assistance will be greatly appreciated.

Yeah, this is pretty much where 'grep' comes into the picture.  :-)

--
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Re: I think I'm in big trouble! (committed changes error)

[Rolf Marsh]

Let me see if I can clarify this better... my (supposedly one and only) 
repository is on my desktop at home.  I have one (1) working copy, which 
is on my flash drive.  I normally do a commit after making changes when 
the flash drive is attached to the home desktop.


This time, I did a commit before leaving for work while the flash drive 
was connected to the home desktop.  I went to work, made some changes to 
the working copy, and foolishly did a commit while the flash drive was 
connected to my laptop.  I don't believe there is a repository on the 
laptop, at least that I knowingly created.


So, my plan now is to somehow replace everything in the repository (on 
the desktop) with the working copy, since I know it has the latest 
version.  The only problem is I don't see anywhere in the docs where it 
explains how to do that.


Thanks for the help so far... I appreciate it...

Regards,
Rolf


On 1/19/2010 10:59 PM, Ryan Schmidt wrote:

On Jan 19, 2010, at 19:03, Rolf Marsh wrote:

   

On 1/19/2010 5:01 PM, Ryan Schmidt wrote:

 

On Jan 19, 2010, at 18:58, Rolf Marsh wrote:
   
 


   

This morning, I took my working copy, which is on a flash drive, and committed 
the changes to the hard drive on my desktop.  I then went to work, made some 
more changes to the project on the flash drive (now attached to my laptop), and 
when I wasn't thinking, committed those changes.  I assume they went to the 
laptop?

So, the long and short of it is, I think my flash drive is out of synch with 
the desktop at home.  How do I verify this and fix it?
 

In both cases, you committed to whatever repository the working copy is 
associated with.
   
 
   

OK... that's what I thought... now how do I fix it, so the main (and supposedly 
only repository) reflects the latest changes?
 

I'm still unclear what you did, or what the problem is. A working copy is only linked to a single 
repository. You can have as many working copies of that repository as you like, stored in as many 
different places as you like. "svn commit" in any of them commits the changes in that 
working copy to the repository. "svn update" in any of the working copies retrieves the 
latest changes from the repository. Further information on how to use Subversion is in the book:

http://svnbook.org

I am assuming you have two working copies, one on your desktop hard drive, and 
one on your flash drive, and that your repository is located in neither of 
those places but on some server somewhere. But it's hard to tell from your 
description.


On Jan 19, 2010, at 19:08, Rolf Marsh wrote:

   

"svn info" on the working copy path will tell you the address of the repository.
   

Hi Ryan... I don't see a "svn info" (attached a screen shot)


 


The screenshot looked like a Windows directory listing of your working copy, which isn't 
relevant. You need to use a Subversion client to ask the working copy for information. 
For example, in a terminal window, change to directory of your working copy and type 
"svn info" to see the repository URL with which the working copy is associated. 
That is where your changes got committed to.


P.S: Remember to Reply All so your replies go to the list too, not just to me.


   

Re: I think I'm in big trouble! (committed changes error)

[C. Michael Pilato]

Rolf Marsh wrote:
> Let me see if I can clarify this better... my (supposedly one and only)
> repository is on my desktop at home.  I have one (1) working copy, which
> is on my flash drive.  I normally do a commit after making changes when
> the flash drive is attached to the home desktop.
> 
> This time, I did a commit before leaving for work while the flash drive
> was connected to the home desktop.  I went to work, made some changes to
> the working copy, and foolishly did a commit while the flash drive was
> connected to my laptop.  I don't believe there is a repository on the
> laptop, at least that I knowingly created.

Did the commit succeed?  If not, no big deal.  Just take the drive home and
commit again to the desktop repository.

If it did succeed, then you've some explaining to do.  Subversion won't
likely let you commit changes (which are expressed as deltas against some
known base, complete with checksums to verify file contents, etc.) easily
into the wrong repository.  Is it possible that your repository -- which you
thought lived on the desktop box -- actually rests on the same flash drive?
 What does 'svn info' in your working copy reveal about the Repository Root?

-- 
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand



signature.asc
Description: OpenPGP digital signature

Re: I think I'm in big trouble! (committed changes error)

[Rolf Marsh]

I believe the commit did succeed, because all of the AnkhSVN glyphs were 
gone... here is the current "svn info" on the flash drive; the F drive 
is on the desktop.  I know the flash drive is the latest.. should I do 
another commit while it's connected to the desktop?



L:\Inventory Program - Working Copy>svn info
Path: .
URL: file:///F:/Repository/trunk/Prager%20Book%20Maintenance
Repository Root: file:///F:/Repository
Repository UUID: 13b279d9-a03c-d346-aaa8-def0303d8a0c
Revision: 10
Node Kind: directory
Schedule: normal
Last Changed Author: Rolf
Last Changed Rev: 6
Last Changed Date: 2010-01-07 15:55:58 -0800 (Thu, 07 Jan 2010)


On 1/20/2010 6:50 AM, C. Michael Pilato wrote:

Rolf Marsh wrote:
   

Let me see if I can clarify this better... my (supposedly one and only)
repository is on my desktop at home.  I have one (1) working copy, which
is on my flash drive.  I normally do a commit after making changes when
the flash drive is attached to the home desktop.

This time, I did a commit before leaving for work while the flash drive
was connected to the home desktop.  I went to work, made some changes to
the working copy, and foolishly did a commit while the flash drive was
connected to my laptop.  I don't believe there is a repository on the
laptop, at least that I knowingly created.
 

Did the commit succeed?  If not, no big deal.  Just take the drive home and
commit again to the desktop repository.

If it did succeed, then you've some explaining to do.  Subversion won't
likely let you commit changes (which are expressed as deltas against some
known base, complete with checksums to verify file contents, etc.) easily
into the wrong repository.  Is it possible that your repository -- which you
thought lived on the desktop box -- actually rests on the same flash drive?
  What does 'svn info' in your working copy reveal about the Repository Root?

   

SVN Server Logs

[Ilan Yaniv]

Hi,

 

I have an old SVN 1.4 and I want to view the SVNSERVE.exe logs. Where
can I find them in the file system. How can I view them

 

Thanks

Ilan Yaniv

 

Re: SVN Server Logs

[Andy Levy]

On Wed, Jan 20, 2010 at 08:49, Ilan Yaniv  wrote:
> Hi,
>
>
>
> I have an old SVN 1.4 and I want to view the SVNSERVE.exe logs. Where can I
> find them in the file system. How can I view them
>

There are none. Logging didn't come to svnserve until 1.6.

Re: I think I'm in big trouble! (committed changes error)

[Ulrich Eckhardt]

On Wednesday 20 January 2010, Rolf Marsh wrote:
> Let me see if I can clarify this better... my (supposedly one and only)
> repository is on my desktop at home.  I have one (1) working copy, which
> is on my flash drive.  I normally do a commit after making changes when
> the flash drive is attached to the home desktop.
>
> This time, I did a commit before leaving for work while the flash drive
> was connected to the home desktop.  I went to work, made some changes to
> the working copy, and foolishly did a commit while the flash drive was
> connected to my laptop.  I don't believe there is a repository on the
> laptop, at least that I knowingly created.

On your flashdrive is a WC. Inside that WC is (in the .svn dirs) the URL of 
the repository, e.g. "file://C:/home/rolf/svn-repository" 
or "http:///home-desktop.example.local/svn";.

If you now try to commit, SVN creates a connection to the repository using the 
URL. If it's a file path, and it finds the repository there, it will commit 
to the repository, but I don't think that could have happened without error, 
because it is unlikely that at the exact same location on your work machine 
is also a repository that has the exact same content.

If you commit using an http URL and your home machine is found even though 
you're in a different network, the commit went to the right repository. 
Otherwise you would have gotten an error, too.

The only thing I could imagine is the repository being on the same flash disk 
as your working copy, but in that case, too, you wouldn't have any problem.


Summary: You should have gotten an error for that commit or you would have 
committed to the correct repository. It's still unclear what exactly you did. 
Also, the requested 'svn info' output is still missing. If you're using 
TortoiseSVN, you can select "Properties" from the Explorer, you should have a 
tab "Subversion" there which contains the relevant info.

Uli

-- 
ML: http://subversion.tigris.org/mailing-list-guidelines.html
FAQ: http://subversion.tigris.org/faq.html
Docs: http://svnbook.red-bean.com/

**
Sator Laser GmbH, Fangdieckstraße 75a, 22547 Hamburg, Deutschland
Geschäftsführer: Thorsten Föcking, Amtsgericht Hamburg HR B62 932
**
   Visit our website at 
**
Diese E-Mail einschließlich sämtlicher Anhänge ist nur für den Adressaten 
bestimmt und kann vertrauliche Informationen enthalten. Bitte benachrichtigen 
Sie den Absender umgehend, falls Sie nicht der beabsichtigte Empfänger sein 
sollten. Die E-Mail ist in diesem Fall zu löschen und darf weder gelesen, 
weitergeleitet, veröffentlicht oder anderweitig benutzt werden.
E-Mails können durch Dritte gelesen werden und Viren sowie nichtautorisierte 
Änderungen enthalten. Sator Laser GmbH ist für diese Folgen nicht 
verantwortlich.
**

Re: I think I'm in big trouble! (committed changes error)

[C. Michael Pilato]

I can't speak to AnkhSVN's behavior -- don't use it myself.

What does 'svn status' in that working copy show?  Are you local
modifications still sitting there waiting to be committed?


Rolf Marsh wrote:
> I believe the commit did succeed, because all of the AnkhSVN glyphs were
> gone... here is the current "svn info" on the flash drive; the F drive
> is on the desktop.  I know the flash drive is the latest.. should I do
> another commit while it's connected to the desktop?
> 
> 
> L:\Inventory Program - Working Copy>svn info
> Path: .
> URL: file:///F:/Repository/trunk/Prager%20Book%20Maintenance
> Repository Root: file:///F:/Repository
> Repository UUID: 13b279d9-a03c-d346-aaa8-def0303d8a0c
> Revision: 10
> Node Kind: directory
> Schedule: normal
> Last Changed Author: Rolf
> Last Changed Rev: 6
> Last Changed Date: 2010-01-07 15:55:58 -0800 (Thu, 07 Jan 2010)
> 
> 
> On 1/20/2010 6:50 AM, C. Michael Pilato wrote:
>> Rolf Marsh wrote:
>>   
>>> Let me see if I can clarify this better... my (supposedly one and only)
>>> repository is on my desktop at home.  I have one (1) working copy, which
>>> is on my flash drive.  I normally do a commit after making changes when
>>> the flash drive is attached to the home desktop.
>>>
>>> This time, I did a commit before leaving for work while the flash drive
>>> was connected to the home desktop.  I went to work, made some changes to
>>> the working copy, and foolishly did a commit while the flash drive was
>>> connected to my laptop.  I don't believe there is a repository on the
>>> laptop, at least that I knowingly created.
>>> 
>> Did the commit succeed?  If not, no big deal.  Just take the drive home and
>> commit again to the desktop repository.
>>
>> If it did succeed, then you've some explaining to do.  Subversion won't
>> likely let you commit changes (which are expressed as deltas against some
>> known base, complete with checksums to verify file contents, etc.) easily
>> into the wrong repository.  Is it possible that your repository -- which you
>> thought lived on the desktop box -- actually rests on the same flash drive?
>>  What does 'svn info' in your working copy reveal about the Repository Root?
>>
>>   


-- 
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand



signature.asc
Description: OpenPGP digital signature

Re: svnmerge.py vs. build-in merge support since version 1.5

[Stein Somers]

I haven't used svnmerge.py

On 20/01/2010 8:04, Dieter Oberkofler wrote:

1) svnmerge.py did generate an informative commit message listing all the
logs of the merged revisions is generated in a text file, as a suggestion
for a good commit message.


That seems like a workaround for the lack of support by tools. Now that 
mergeinfo is standard, if you use for instance TSVN, you can 
(optionally) drill down to the log messages of merged changesets. So the 
log message of a merge can be to the point: it's a merge. If you want 
detail of what was merged, let the SVN browser get it for you.



2) When suing svnmerge.py for development branches it had a "merge
everything" command (does everything automatically) that also kept track of
the directory the branch has been made of. This does no longer seem to be
available.


I would certainly like that: the default merge source could be the 
branch that the working copy's URL is a copy of. So "svn merge" would 
suffice. But then there are situation where you merge from a different 
branch, or where the merge source has been renamed (bad practice, I know).


--
Stein

Re: I think I'm in big trouble! (committed changes error)

[Rolf Marsh]

Most of the files are marked "under version control but are missing or 
incomplete", while a few are marked as "modified" which is what I 
expected.  Is there a way to do a "commit" of everything?  Or, is there 
a way to replace all of the files in the repository with those on the 
flash drive?


R

On 1/20/2010 7:19 AM, C. Michael Pilato wrote:

I can't speak to AnkhSVN's behavior -- don't use it myself.

What does 'svn status' in that working copy show?  Are you local
modifications still sitting there waiting to be committed?


Rolf Marsh wrote:
   

I believe the commit did succeed, because all of the AnkhSVN glyphs were
gone... here is the current "svn info" on the flash drive; the F drive
is on the desktop.  I know the flash drive is the latest.. should I do
another commit while it's connected to the desktop?


L:\Inventory Program - Working Copy>svn info
Path: .
URL: file:///F:/Repository/trunk/Prager%20Book%20Maintenance
Repository Root: file:///F:/Repository
Repository UUID: 13b279d9-a03c-d346-aaa8-def0303d8a0c
Revision: 10
Node Kind: directory
Schedule: normal
Last Changed Author: Rolf
Last Changed Rev: 6
Last Changed Date: 2010-01-07 15:55:58 -0800 (Thu, 07 Jan 2010)


On 1/20/2010 6:50 AM, C. Michael Pilato wrote:
 

Rolf Marsh wrote:

   

Let me see if I can clarify this better... my (supposedly one and only)
repository is on my desktop at home.  I have one (1) working copy, which
is on my flash drive.  I normally do a commit after making changes when
the flash drive is attached to the home desktop.

This time, I did a commit before leaving for work while the flash drive
was connected to the home desktop.  I went to work, made some changes to
the working copy, and foolishly did a commit while the flash drive was
connected to my laptop.  I don't believe there is a repository on the
laptop, at least that I knowingly created.

 

Did the commit succeed?  If not, no big deal.  Just take the drive home and
commit again to the desktop repository.

If it did succeed, then you've some explaining to do.  Subversion won't
likely let you commit changes (which are expressed as deltas against some
known base, complete with checksums to verify file contents, etc.) easily
into the wrong repository.  Is it possible that your repository -- which you
thought lived on the desktop box -- actually rests on the same flash drive?
  What does 'svn info' in your working copy reveal about the Repository Root?


   


   

[ANN] VisualSVN Server 2.1 Released

[Danil Shopyrin]

Hello All,

We're delighted to announce VisualSVN Server 2.1. This release builds
on the reliability and functionality of previous versions with the addition of
Integrated Windows Authentication in Enterprise Edition.

VisualSVN Server is a package that contains everything you need to
install, configure and manage Subversion server on the Windows platform.
You may find more information about VisualSVN Server at:
http://www.visualsvn.com/server/

With Integrated Windows Authentication users gain access to VisualSVN
Server without being prompted for username and password. The current
Windows credentials on the client computer are supplied through a cryptographic
exchange involving hashing with the VisualSVN Server. Either NTLM or Kerberos
V5 authentication protocols can be used.

Integrated Windows Authentication is available if you use VisualSVN Server 2.1
Enterprise Edition and any of the standard Subversion clients such as
Subversion binaries (i.e. svn.exe), TortoiseSVN and VisualSVN.

All of this made Subversion into a first class citizen on the Windows
platform. Integrated Windows Authentication brings clear benefits:

Optimal security
---
Authentication is always performed in a secure manner and credentials are never
cached on disk.

Single sign-on
-
Users' Windows credentials are automatically used to access VisualSVN Server,
saving them the hassle of additional authentication steps.

Two-factor authentication

Compatibility with two-factor authentication systems (e.g. smart cards) ensures
that VisualSVN Server can be easily integrated to a secure enterprise
environment.

Integrated Windows Authentication is available in VisualSVN Server Enterprise
Edition only. But VisualSVN Server 2.1 brings some significant fixes for
Standard Edition too. See the full list of changes:
http://www.visualsvn.com/server/changes/#v2.1

We recommend all users upgrade to version 2.1. The upgrade is free to Standard
Edition users and existing Enterprise Edition customers. Upgrading is very quick
and easy: simply run the installer for VisualSVN Server 2.1 and follow the
installation wizard. It is not necessary or recommended to uninstall
older versions of VisualSVN Server.

Download VisualSVN Server 2.1 at the download page:
http://www.visualsvn.com/server/download/

--
With best regards,
Danil Shopyrin
VisualSVN Team

Re: I think I'm in big trouble! (committed changes error)

[Andrey Repin]

Greetings, Ulrich Eckhardt!

>> Let me see if I can clarify this better... my (supposedly one and only)
>> repository is on my desktop at home.  I have one (1) working copy, which
>> is on my flash drive.  I normally do a commit after making changes when
>> the flash drive is attached to the home desktop.
>>
>> This time, I did a commit before leaving for work while the flash drive
>> was connected to the home desktop.  I went to work, made some changes to
>> the working copy, and foolishly did a commit while the flash drive was
>> connected to my laptop.  I don't believe there is a repository on the
>> laptop, at least that I knowingly created.

> On your flashdrive is a WC. Inside that WC is (in the .svn dirs) the URL of 
> the repository, e.g. "file://C:/home/rolf/svn-repository" 
> or "http:///home-desktop.example.local/svn";.

> If you now try to commit, SVN creates a connection to the repository using 
> the 
> URL. If it's a file path, and it finds the repository there, it will commit 
> to the repository, but I don't think that could have happened without error, 
> because it is unlikely that at the exact same location on your work machine 
> is also a repository that has the exact same content.

More specifically, it's highly unlikely that repository at that location has
the same UUID() as the one at original location.
For that, you need to take certain steps to replicate *repository* at laptop.
And it falls into "you're enemy of yourself" category.

> If you commit using an http URL and your home machine is found even though 
> you're in a different network, the commit went to the right repository. 
> Otherwise you would have gotten an error, too.

Yep.


--
WBR,
 Andrey Repin (anrdae...@freemail.ru) 20.01.2010, <19:24>

Sorry for my terrible english...

Re: I think I'm in big trouble! (committed changes error)

[Henrik Sundberg]

On Wed, Jan 20, 2010 at 3:58 PM, Rolf Marsh  wrote:
> I believe the commit did succeed, because all of the AnkhSVN glyphs were
> gone... here is the current "svn info" on the flash drive; the F drive is on
> the desktop.  I know the flash drive is the latest.. should I do another
> commit while it's connected to the desktop?
>
>
> L:\Inventory Program - Working Copy>svn info
> Path: .
> URL: file:///F:/Repository/trunk/Prager%20Book%20Maintenance
> Repository Root: file:///F:/Repository
> Repository UUID: 13b279d9-a03c-d346-aaa8-def0303d8a0c
> Revision: 10
> Node Kind: directory
> Schedule: normal
> Last Changed Author: Rolf
> Last Changed Rev: 6
> Last Changed Date: 2010-01-07 15:55:58 -0800 (Thu, 07 Jan 2010)


Is the flash drive mounted as F: at both computers?
In that case you do have the repository on the flash drive.

/$

Re: I think I'm in big trouble! (committed changes error)

[C. Michael Pilato]

Rolf Marsh wrote:
> Most of the files are marked "under version control but are missing or
> incomplete", while a few are marked as "modified" which is what I
> expected.  Is there a way to do a "commit" of everything?  Or, is there
> a way to replace all of the files in the repository with those on the
> flash drive?

That means you didn't really commit them.  Just put the flash drive back
onto your home computer and re-issue the commit either via AnkhSVN, or using
"svn commit" at the command-line.

-- 
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand



signature.asc
Description: OpenPGP digital signature

Re: I think I'm in big trouble! (committed changes error)

[Rolf Marsh]

No, the flash drive is mounted as drive L on the desktop and drive F on 
the laptop...did it create a repository automatically on the flash drive 
when it was attached to the laptop and I did the commit?


However, when I do a "svn info" on the laptop, "svn" is not recognized 
because I didn't install subversion on the laptop.. I only installed 
TortoiseSVN and ankhSVN on the laptop.  Since I saw the "gliphs", I 
didn't think I needed Subversion on the laptop.  Was that a bad assumption?




R


On 1/20/2010 8:49 AM, Henrik Sundberg wrote:

On Wed, Jan 20, 2010 at 3:58 PM, Rolf Marsh  wrote:
   

I believe the commit did succeed, because all of the AnkhSVN glyphs were
gone... here is the current "svn info" on the flash drive; the F drive is on
the desktop.  I know the flash drive is the latest.. should I do another
commit while it's connected to the desktop?


L:\Inventory Program - Working Copy>svn info
Path: .
URL: file:///F:/Repository/trunk/Prager%20Book%20Maintenance
Repository Root: file:///F:/Repository
Repository UUID: 13b279d9-a03c-d346-aaa8-def0303d8a0c
Revision: 10
Node Kind: directory
Schedule: normal
Last Changed Author: Rolf
Last Changed Rev: 6
Last Changed Date: 2010-01-07 15:55:58 -0800 (Thu, 07 Jan 2010)
 


Is the flash drive mounted as F: at both computers?
In that case you do have the repository on the flash drive.

/$
   

Re: I think I'm in big trouble! (committed changes error)

[Rolf Marsh]

I did just that, and it appears all is OK... of course I won't know 
until something dies, but at least I'm not worrying about it anymore.


Thank you everybody who took the time to get me through this... and by 
the way Michael, I just bought the latest version of your book... now, 
just to find the time to read it! :-)


Thanks again...

R

On 1/20/2010 8:56 AM, C. Michael Pilato wrote:

Rolf Marsh wrote:
   

Most of the files are marked "under version control but are missing or
incomplete", while a few are marked as "modified" which is what I
expected.  Is there a way to do a "commit" of everything?  Or, is there
a way to replace all of the files in the repository with those on the
flash drive?
 

That means you didn't really commit them.  Just put the flash drive back
onto your home computer and re-issue the commit either via AnkhSVN, or using
"svn commit" at the command-line.

   

Re: I think I'm in big trouble! (committed changes error)

[C. Michael Pilato]

My kids' college fund thanks you.

Rolf Marsh wrote:
> I did just that, and it appears all is OK... of course I won't know
> until something dies, but at least I'm not worrying about it anymore.
> 
> Thank you everybody who took the time to get me through this... and by
> the way Michael, I just bought the latest version of your book... now,
> just to find the time to read it!  :-)
> 
> Thanks again...
> 
> R
> 
> On 1/20/2010 8:56 AM, C. Michael Pilato wrote:
>> Rolf Marsh wrote:
>>   
>>> Most of the files are marked "under version control but are missing or
>>> incomplete", while a few are marked as "modified" which is what I
>>> expected.  Is there a way to do a "commit" of everything?  Or, is there
>>> a way to replace all of the files in the repository with those on the
>>> flash drive?
>>> 
>> That means you didn't really commit them.  Just put the flash drive back
>> onto your home computer and re-issue the commit either via AnkhSVN, or using
>> "svn commit" at the command-line.
>>
>>   


-- 
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand



signature.asc
Description: OpenPGP digital signature

Svn red-bean book...

[Bob Archer]

I was wondering if this page should be updated?

http://svnbook.red-bean.com/

There are links for the 1.5 version of the book and then for the nightly which 
says for version 1.6?

Shouldn't there be direct links to the 1.6 version and then the nightly be for 
1.7?

I'm not sure who maintains this page so I just thought I would mention it here.

Is this book being updated for 1.7? This book is basically the documentation 
for svn right?

BOb

Re: Svn red-bean book...

[C. Michael Pilato]

Bob Archer wrote:
> I was wondering if this page should be updated?
> 
> http://svnbook.red-bean.com/
> 
> There are links for the 1.5 version of the book and then for the nightly
> which says for version 1.6?
> 
> Shouldn't there be direct links to the 1.6 version and then the nightly
> be for 1.7?

There is no tagged 1.6 version.  The folks who contribute to that
documentation invested alot of energy to get the text up-to-date with
Subversion 1.5, and never quite recovered from the resulting burn-out.

> I'm not sure who maintains this page so I just thought I would mention
> it here.

Well, the page does say, "For feedback on the book or this website, contact
svnbook-...@red-bean.com".  :-)

> Is this book being updated for 1.7? This book is basically the
> documentation for svn right?

It is an unofficial set of documentation for Subversion, albeit the only
known open-source one.

-- 
C. Michael Pilato 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand



signature.asc
Description: OpenPGP digital signature

RE: Svn red-bean book...

[Bob Archer]

> Bob Archer wrote:
> > I was wondering if this page should be updated?
> >
> > http://svnbook.red-bean.com/
> >
> > There are links for the 1.5 version of the book and then for the
> > nightly which says for version 1.6?
> >
> > Shouldn't there be direct links to the 1.6 version and then the
> > nightly be for 1.7?
> 
> There is no tagged 1.6 version.  The folks who contribute to that
> documentation invested alot of energy to get the text up-to-date with
> Subversion 1.5, and never quite recovered from the resulting burn-out.


Well, if there is anything I can do to help let me know. I would expect that 
there is some revision in trunk where a 1.6 tag could be created. I know that 
the 1.5 version is somewhat out of date but an always concerned when looking at 
the nightly version if I am reading stuff that is only relevant to 1.7 and 
doesn't apply to 1.6.

> > I'm not sure who maintains this page so I just thought I would mention
> > it here.
> 
> Well, the page does say, "For feedback on the book or this website,
> contact svnbook-...@red-bean.com".  :-)


Sorry I missed that.

> > Is this book being updated for 1.7? This book is basically the
> > documentation for svn right?
> 
> It is an unofficial set of documentation for Subversion, albeit the only
> known open-source one.

Gotcha.

BOb

Some relative svn externals takes very long time

[Joakim Tjernlund]

Hi

We have a repo with lots of svn:externals that looks like this
 ^tags/x/y z
Each time one executes a svn up or svn status it takes a very long time
to complete. The reason is that for each such external a new svnserve if
started on the server.
The ^tags/x/y externals are relative to the repo so I wonder
if it is possible to follow these externals within the same
svnserve process instead?

 Jocke

Re: Python & RA

[Роман Донченко]

Daniel Eggert  писал в своём письме Wed, 20 Jan  
2010 19:54:40 +0300:


I got svn.ra.open2() and svn.ra.get_dir() working, but I'm unable to  
make the receiver/baton part of
	l = svn.ra.get_log2(session, None, 1, 4, 0, False, False, [],  
receiver, None)
work. The receiver is supposed to be a svn_log_entry_receiver_t. I  
tried to do

def receiver(baton, entry):
print entry
but I get
TypeError: argument number 8:


The immediate cause of this is that you forgot the  
include_merged_revisions argument. 8=] But were you to include it, your  
code would still be wrong, because...



How does swig wrap svn_log_entry_receiver_t?


A callable Python object (which can be a function, or a class instance)  
that you pass becomes both the C function and the C baton. Thus, your  
call should look like:


svn.ra.get_log2(
 session,   # session
 None,  # paths 
 1, # start
 4, # end
 0, # limit
 False, # discover_changed_paths
 False, # strict_node_history
 False, # include_merged_revisions
 [],# revprops
 receiver   # receiver & receiver_baton
# pool (is omitted)
)

And the receiver should look like:

def receiver(entry, pool):
   print entry


I have not been able to find any python sample code that uses svn.ra.


Look at the bindings' testsuite at  
.


Thanks. This is very helpful.

Some of those crash for me, too:


root = svn.ra.get_repos_root(session)
Assertion failed: (is_canonical(path, strlen(path))), function  
svn_path_is_empty, file  
/SourceCache/subversion/subversion-35/subversion/subversion/libsvn_subr/path.c,  
line 414.

Abort

It turns out that this was caused by my repository URL having a trailing  
"/". Not very elegant for the entire thing to simply crash, and not very  
easy to figure out what was causing this.


It would be great if this would be handled more gracefully.

I'm currently using version 1.6.5 (r38866).


This is a constraint of the C API. When you pass paths and URIs to  
Subversion functions, they must be in the canonical internal form. Unless  
the URI you use is an output of another Subversion function, you must use  
core.svn_path_internal_style to prepare it.


Roman.

Ask for confirmation when a File is updated

[Hernan Castagnola]

Hi all,

Is there any way (e.g editing the svn-properties) to make svn
give an Alert, or ask for confirmation when a sensitive file is updated.
  I have been searching and I didn't find anything.

*Example *

Config.groovy is a sensitive file

so on a Server X

you set Config.groovy as a sensitive file

when you run

svn up  and Config.groovy has been changed on the server, it tells you

Config.groovy has been changed on ther server.
Yes, No, Diff, etc
(as when you have a conflict)


I am asking this, because sometime files are committed by mistake. And maybe
is a good way to catch them before going to production.

Thanks in advance for your help

RE: Ask for confirmation when a File is updated

[Bob Archer]

No, but you could probably do some type of pre-commit hook that will request a 
verification log message or something. They have to type some random string. 
you would have to save the string some where or something like:

Commit failed, sensitive file included in change set. To verify that you want 
to update this file please enter "ABGHY5" in your commit message.

So, in the pre-commit hook you check to see if you have a sensitive file. Then 
check to see if you have a confirm string saved for the file in question. You 
might use a convention on your server like... filename.ext.confirm. If that 
file exists read the confirm string and see if it is contained in the change 
log. You will probably want to delete .confirm files that are older than say 5 
minutes or something.

You could also have some type of password that always allows the commit to go 
through... of course, that is dangerous... admins can make mistakes too.

Not very elegant but it should work.

I think a better idea is to not automatically promote stuff that has been 
checked into svn to production. You should have a staging server were stuff 
goes to first where you run your test suite... have a human decide if the 
staging should be prompted to production.

BOb


From: Hernan Castagnola [mailto:castagno...@gmail.com]
Sent: Wednesday, January 20, 2010 3:57 PM
To: users@subversion.apache.org
Subject: Ask for confirmation when a File is updated

Hi all,

Is there any way (e.g editing the svn-properties) to make svn give 
an Alert, or ask for confirmation when a sensitive file is updated.
  I have been searching and I didn't find anything.

Example

Config.groovy is a sensitive file

so on a Server X

you set Config.groovy as a sensitive file

when you run

svn up  and Config.groovy has been changed on the server, it tells you

Config.groovy has been changed on ther server.
Yes, No, Diff, etc
(as when you have a conflict)


I am asking this, because sometime files are committed by mistake. And maybe is 
a good way to catch them before going to production.

Thanks in advance for your help

Re: Ask for confirmation when a File is updated

[Ryan Schmidt]

On Jan 20, 2010, at 14:57, Hernan Castagnola wrote:

> Is there any way (e.g editing the svn-properties) to make svn 
> give an Alert, or ask for confirmation when a sensitive file is updated. 
>   I have been searching and I didn't find anything. 
> 
> Example 
> 
> Config.groovy is a sensitive file 
> 
> so on a Server X
> 
> you set Config.groovy as a sensitive file
> 
> when you run
> 
> svn up  and Config.groovy has been changed on the server, it tells you
> 
> Config.groovy has been changed on ther server. 
> Yes, No, Diff, etc
> (as when you have a conflict)
> 
> 
> I am asking this, because sometime files are committed by mistake. And maybe 
> is a good way to catch them before going to production.

No, there's nothing like that built-in, and I think you're trying to solve the 
wrong problem. It sounds like you are directly deploying your trunk to 
production with no testing, review or QA process; I would suggest that adding 
testing, review and/or QA would be the solution.

For example, use a post-commit hook to send an email to the team with any 
changes as they get made. This way the team can review all changes, and 
hopefully spot incorrect ones before they become a problem.

And instead of deploying direct from trunk, deploy from a tag. Presumably you 
have an individual who is responsible for deploying to production. This 
individual will test the trunk, and when satisfied, make a tag of it, and 
deploy the tag. There are post-commit scripts you can use to auto-deploy tags 
to production, so that the only step needed to deploy something is to make the 
tag in the repository. Now all you have to do is not create a tag when there is 
anything questionable or unverified going on in trunk.

Re: Python & RA

[Daniel Eggert]

On Jan 20, 2010, at 11:10 , Роман Донченко wrote:

> Daniel Eggert  писал в своём письме Wed, 20 Jan 2010 
> 19:54:40 +0300:
> 
 I got svn.ra.open2() and svn.ra.get_dir() working, but I'm unable to make 
 the receiver/baton part of
l = svn.ra.get_log2(session, None, 1, 4, 0, False, False, [], receiver, 
 None)
 work. The receiver is supposed to be a svn_log_entry_receiver_t. I tried 
 to do
def receiver(baton, entry):
print entry
 but I get
TypeError: argument number 8:
>>> 
>>> The immediate cause of this is that you forgot the include_merged_revisions 
>>> argument. 8=] But were you to include it, your code would still be wrong, 
>>> because...
>>> 
 How does swig wrap svn_log_entry_receiver_t?
>>> 
>>> A callable Python object (which can be a function, or a class instance) 
>>> that you pass becomes both the C function and the C baton. Thus, your call 
>>> should look like:
>>> 
>>> svn.ra.get_log2(
>>> session,# session
>>> None,   # paths 
>>> 1,  # start
>>> 4,  # end
>>> 0,  # limit
>>> False,  # discover_changed_paths
>>> False,  # strict_node_history
>>> False,  # include_merged_revisions
>>> [], # revprops
>>> receiver# receiver & receiver_baton
>>> # pool (is omitted)
>>> )
>>> 
>>> And the receiver should look like:
>>> 
>>> def receiver(entry, pool):
>>>  print entry
>>> 
 I have not been able to find any python sample code that uses svn.ra.
>>> 
>>> Look at the bindings' testsuite at 
>>> .
>> 
>> Thanks. This is very helpful.
>> 
>> Some of those crash for me, too:
>> 
> root = svn.ra.get_repos_root(session)
>> Assertion failed: (is_canonical(path, strlen(path))), function 
>> svn_path_is_empty, file 
>> /SourceCache/subversion/subversion-35/subversion/subversion/libsvn_subr/path.c,
>>  line 414.
>> Abort
>> 
>> It turns out that this was caused by my repository URL having a trailing 
>> "/". Not very elegant for the entire thing to simply crash, and not very 
>> easy to figure out what was causing this.
>> 
>> It would be great if this would be handled more gracefully.
>> 
>> I'm currently using version 1.6.5 (r38866).
> 
> This is a constraint of the C API. When you pass paths and URIs to Subversion 
> functions, they must be in the canonical internal form. Unless the URI you 
> use is an output of another Subversion function, you must use 
> core.svn_path_internal_style to prepare it.

Aha! Thanks a lot!

/Daniel

svnlook problem when fetching log by command svnlook

[Xia, Richard]

Hello,
I came across a svnlook problem, and try to get help.

Bellow is my environment:
Os: windows XP English version, but locale is set to Japan
Subversion:  Setup-Subversion-1.5.1.en-us.msi

I added a file in svn repository with comment in which mixed with English, 
Chinese and Japanse. When I use svnlook command to fetch the comment log, only 
Enlish and Japanese display correctly.
It seems that the  the svnlook command translate the log encoding into 
shift_jis.
However, using tortoiseSVN client ,the log displays correctly.
I guess there is something to do with the encoding. Is there any way to make 
svnlook using utf8 encoding?
Any help would be appreciated.

Trying to rebase branch with trunk results in the error svn: REPORT of '/repos/path/!svn/vcc/default': Could not read response body: Secure connection truncated (https://ourdomain)

[EwBoonTwo]

This is with subversion client 1.5.5.

I've been working with subversion for years on the same codebase and
haven't run into this issue before. I did some digging around and
found a few threads about the error above, but all were related to
checking-out repositories, not merges, and none had a solution that I
saw.

I've got a branch created from trunk that I'm trying to rebase with
trunk. This is the first rebase I've run since the branch was created.
Running the following command:
$ svn merge --non-interactive https://path/to/trunk

result in the following:
svn: REPORT of '/repos/path/!svn/vcc/default': Could not read response
body: Secure connection truncated (https://ourdomain.com)

If I specify revision numbers from the creation of the branch to
trunk's current revision:
$ svn merge -r 40584:40659 --non-interactive https://path/to/trunk

I get the same error, which is what I'd expect. If I back up a few
revisions, no error is thrown and the merge occurs:
$ svn merge -r 40584:40655 --non-interactive https://path/to/trunk

What can I do to track down what's going wrong in the problematic
revisions? Thanks!

Svn Bug? Cherrypicking a change causes incorrect mergeinfo recording later; breaks reintegrate.

[Rick Ballard]

Hello,

I'm hitting what appears to be a bug in subversion. The problem is that I get 
incorrect subtree mergeinfo after a merge to my branch; the subtree is not 
recording the merge, and this later prevents reintegrate from working (via 
"Reintegrate can only be used if revisions [X] through [Y] were previously 
merged from [URL]..."). It appears that this bug is triggered by a 
differentbranch->trunk cherrypicking merge earlier in the history of the 
subtree.

I've distilled this down to a simple test case that reproduces the problem. The 
attached "cherrypick-mergeinfo-fail-test.sh" script will run the steps 
necessary to demonstrate. I've also attached the output this script generates 
as "cherrypick-test-output.txt", so you can see the problem without running the 
script.

I'm still on subversion 1.6.5 (r38866), but the 1.6.6 CHANGES file doesn't 
appear to address this. I've tried this on a standard Mac OS X SnowLeopard 
(10.6.2) system with the system install of subversion.

Is there something I've actually done wrong here that's triggering the problem? 
Or does this appear to be a bug?

Thanks very much,

- Rick Ballard



cherrypick-mergeinfo-fail-test.sh
Description: Binary data
 
Committed revision 1.
Checked out revision 1.
A trunk-wc/subtree
A trunk-wc/subtree/file.txt
Adding trunk-wc/subtree
Adding trunk-wc/subtree/file.txt
Transmitting file data .
Committed revision 2.

Committed revision 3.
Abranch1-wc/subtree
Abranch1-wc/subtree/file.txt
Checked out revision 3.
Sendingbranch1-wc/subtree/file.txt
Transmitting file data .
Committed revision 4.
At revision 4.
--- Merging r4 into 'trunk-wc/subtree':
Utrunk-wc/subtree/file.txt
Sendingtrunk-wc/subtree
Sendingtrunk-wc/subtree/file.txt
Transmitting file data .
Committed revision 5.
At revision 5.

Committed revision 6.

Committed revision 7.
Abranch2-wc/subtree
Abranch2-wc/subtree/file.txt
Checked out revision 7.
Abranch3-wc/subtree
Abranch3-wc/subtree/file.txt
Checked out revision 7.
Sendingbranch2-wc/subtree/file.txt
Transmitting file data .
Committed revision 8.
At revision 8.
--- Merging differences between repository URLs into 'trunk-wc':
Utrunk-wc/subtree/file.txt
Sendingtrunk-wc
Sendingtrunk-wc/subtree
Sendingtrunk-wc/subtree/file.txt
Transmitting file data .
Committed revision 9.
###
### Mergeinfo on branch3 pre-merge:
###
Properties on 'branch3-wc/subtree':
  svn:mergeinfo
/branches/branch1/subtree:4
###
--- Merging r7 through r9 into 'branch3-wc':
Ubranch3-wc/subtree/file.txt
 U   branch3-wc/subtree
 U   branch3-wc
Sendingbranch3-wc
Sendingbranch3-wc/subtree
Sendingbranch3-wc/subtree/file.txt
Transmitting file data .
Committed revision 10.
###
### Mergeinfo on branch3 subtree was not set correctly by merge: Should include 
trunk revisions just merged.
###
Properties on 'branch3-wc':
  svn:mergeinfo
/branches/branch2:6-8
/trunk:7-9
Properties on 'branch3-wc/subtree':
  svn:mergeinfo
/branches/branch1/subtree:4
/branches/branch2/subtree:6-8
###
### This will cause the next command (a reintegrate) to fail.
###
At revision 10.
svn: Reintegrate can only be used if revisions 7 through 10 were previously 
merged from file:///Users/rick/Desktop/repos/trunk to the reintegrate source, 
but this is not the case:
  branches/branch3/subtree
Missing ranges: /trunk/subtree:7-9

Re: svnmerge.py vs. build-in merge support since version 1.5

[Blair Zajac]

Dieter Oberkofler wrote:

I have been using svnmerge.py for quite a while and although I was looking
forward for the build-in functionality it worked quite well. After now
upgrading to version 1.6 with the build-in functionality that replaces the
need for svnmerge.py I'm a little surprised that not all of the script
functionality seems to be available or I'm just unable to find it:
1) svnmerge.py did generate an informative commit message listing all the
logs of the merged revisions is generated in a text file, as a suggestion
for a good commit message. Is this no longer available in the build-in
support?
2) When suing svnmerge.py for development branches it had a "merge
everything" command (does everything automatically) that also kept track of
the directory the branch has been made of. This does no longer seem to be
available.
3) Is there any documentation that explains the differences between
svnmerge.py and the new build-in functionality?

I would love to understand if I just did not find the functionality, it is
not yet available or it has been omitted on purpose.


The main differences are that

- svnmerge.py does whole revision only merges, it doesn't support per-file 
merging
- built in svn merge doesn't have a revision blocking feature
- svnmerge.py does do the log message tracking, but the svn merge log can get 
this same information for you with -g


Regards,
Blair

Re: No such revision 0

[Ryan Schmidt]

On Jan 20, 2010, at 21:27, Kiko Man wrote:

> On Tue, Jan 19, 2010 at 11:37 PM, Ryan Schmidt wrote:
> 
>> And also in /var/svn/repos/db/revprops?
> 
> Yes, it's there as well:
> jj db # ls -alt /var/svn/repos/db/revprops/0
> -rw-rw 1 backuppc apache50 2009-10-30 19:35 0
> 
> Good to know about FSFS

Ok, then I'm not sure what's up. i'm sending this reply back to the mailing 
list so others can have a go at helping you.

Subversion in 2010 and Beyond Slides

[Elaine]

 http://bit.ly/svn2010slides

Interesting presentation by Subversion developers Hyrum Wright, Philip
Martin, Julian Foad and Dave Brown.

Highlights include:

HTTPv2.0
This update of the current WebDAV HTTP protocol will boost performance
while preserving the benefits of providing web access to Subversion.

SVNJ
A rich server-side JAVA API that leverages Subversion's established
code base. SVNJ will enable the Subversion community to reach out to
millions of JAVA developers to create the same rich functionality and
extensibility on the server that's available on the client.

Obliterate
A new feature that cleanly removes obsolete files and other data from
Subversion repositories. Obliterate will include comprehensive audit
and recovery capabilities to guarantee that history is always
available.

Working Copy
New features include centralized metadata storage and improved
extensibility. These enhancements will lay the groundwork for offline
commits and other capabilities normally associated with distributed
version control systems.