1)      I created a folder called D:\svn_repository\repo

2)      Under this repo folder, I created 3 folders called Trunk, Branch and Branch2

3)      I made all these 3 as repositories by changing directory to repo and keying in the following: -

3.1) Svnadmin create –fs-type fsfs branch

3.2) svnadmin create –fs-type fsfs branch2

3.3) svnadmin create –fs-type fsfs trunk

4)      I went into the ‘conf’ sub folders of all these 3 and edited the svnserve.conf files and uncommented the anon-access’, ‘auth-access’ and ‘password-db’ sections. In each of these sub folders I also edited the ‘passwd’ files and uncommented the default passwords (harry, harryssecret).

5)      I created 3 batch files (file1.bat, file2.bat and file3.bat) with contents like this :-

5.1) svnserve.exe –d –r  D:/svn_repository/repos/trunk  -–listen-port 3691

5.2) svnserve.exe –d –r  D:/svn_repository/repos/branch  -–listen-port 3692

5.3) svnserve.exe –d –r  D:/svn_repository/repos/branch2  -–listen-port 3693

I ran these 3 batch files so that 3 repositories were running

6)      I imported some files into the ‘Trunk’ repository by right clicking that folder , choosing ‘import’ and using the URL ‘svn://localhost:3691/Repos/Trunk’

7)      Then created some client folder like ‘F:/SVNMergeTrialRepo/Trunk’

8)      I checked out into this folder with the above URL

9)      Then created 2 more client folders like ‘F:/SVNMergeTrialRepo/Branch’ and ‘F:/SVNMergeTrialRepo/Branch2’

10)  Now I checked out into ‘Branch’ from URL ‘svn://localhost:3691/Repos/Branch’

11)  Then I checked out into Branch2 from URL ‘svn://localhost:3691/Repos/Branch2’

12)  Now I created a sub-folder F1 with a FileOne.Text in it within Branch client folder and added and committed it (revision 4)

13)  Now I created a sub-folder F2 with a FileTwo.Text in it within Branch2 client folder and added and committed it (revision 5)

14)  Then I right clicked ‘Branch’  folder and selected Merge. I selected the ‘From branch’ as  svn://localhost:3691/Repos/Branch2’ and selected revision number 5

15)  Then I committed the ‘Branch’ folder

16)  Now when I see the contents of ‘Branch’ folder, it has the ‘FileTwo’ which was added in revision 5 under ‘Branch2’

So something done in a specific revision in one branch was replicated in another branch using merge.

•Create a folder as instructed by the TL (maybe D:/SVNClient)

•Right click on this new folder and choose ‘SVN Checkout’

•Click on the ‘SVN Checkout’ option in the pop up menu

•A Checkout dialog box will pop up

•Key in the URL of the Subversion repository (svn://chakravarthy:3691/myWebApp)

•Ensure checkout directory is correct .Click ‘Ok’

•The authentication dialog box will open up

•Key in your ‘Username’ and ‘password’ Press ‘Ok’

•The checkout dialog box will list the files it has checked out and it’s revision number

  •The SVNClient folder now has a green tick mark to say that the client working folder (called working copy) is in synch with the repository

  •After some time you save your work .Now notice the red tick mark on the file you just now edited

  •You can see what are the changes you did just now . Right click on the file with the red tick, choose ‘TortoiseSVN’ and choose ‘Diff’.  You get a difference between the working copy and the version you had recently updated from server

  •Now that you are satisfied with the change you made, you can ‘commit’ the change to the server •Right click on the file, or any of its parent folders, and choose ‘SVN Commit’

  •The log message box opens for you to select the files you wish to commit. Tick the files you wish to commit •Only commit files that you are editing, or files which are needed to rebuild the system . Do not commit outputs of compilation •Key in a suitable message for this commit

  •If someone else has already committed the very same file in the meantime, SVN will ask you to first update your working copy

To   Update the working copy ,   Right click the file or the parent folders and choose the ‘SVN Update’ option

  •You will now get a ‘merged’ message box which means that your change and the change in the server have got merged in the working copy •

  My first conflict !

  •If the file I am trying to commit is out of date, SVN asks me to update my copy first •So I update it •However, if the same lines have been edited or touched in my working copy as well as the Repository copy, then I get a conflict error message ….. •The file and folders now have a yellow warning sign

  When there is a conflict   there are 2 additional files beside the file which has the conflict •These files have names like ‘filename.mine’, ‘filename.revisionx’ and ‘filename.revisiony’ •Open the conflicting file in an editor

  In case of Conflict   •The changes made by you and the changes made by the other SVN user are marked by arrows <<<<<<< •Your change has a caption ‘.mine’ and the other change has a caption ‘.r(n)’ where n is the revision number

  •Retain the change which is correct, and remove the unwanted change •Save the file

  •Having cleaned up the conflicting file, now right click the file or the parent folders and click the ‘Resolved’ option

  •Conflict has been resolved , unwanted revision files removed •Now we need to commit

  How to add a file to the repository

  •If you have added a file in your working copy which needs to be added to repository, right click the file and click ‘Add’ •The ‘file’ that has been added will have a blue plus sign – but it is not yet in repository !

  Deleting a file in Server

  •If you choose a folder / file , right click , choose ‘Delete’ and then commit, the files / folders will get deleted from the repository •If you only want to delete the client copy and get a fresh one from repository, use a normal file delete and Update your root folder

Now the file ‘A new file’ has disappeared from the folder since I used the Tortoise Delete option against the file

  •After deleting a file using the Tortoise client, the parent folder needs to be committed for the file to be deleted from the repository •Otherwise the client file would have been deleted but the server repository still has the file

  •If you wish to rename a file which is under SVN, use the Tortoise rename, not the Windows file rename

  •The tortoise rename does 2 steps – it first copies the file to another file with the new name and it deletes the earlier file

 View the history of all changes  To view who did what , right click the folder, or file and choose ‘Show Log’.  Right click and choose ‘Check for modifications’

  To see contents of Repository 

  Right click any item in SVN client, choose ‘Repo browser’

  How to tag

  •svn copy comes to the rescue. If you want to create a snapshot of /calc/trunk exactly as it looks in the HEAD revision, then make a copy of it: •$ svn copy http://svn.example.com/repos/calc/trunk \ http://svn.example.com/repos/calc/tags/release-1.0 \ -m “Tagging the 1.0 release of the ‘calc’ project.” Committed revision 351. This example assumes that a /calc/tags directory already exists. (If it doesn’t, see svn mkdir).

  Authorization

  •SVN can authorize a person to read, or read and write , or not read any particular folder •So the SVN Admin will be assigning folder rights to different roles so that each user of a certain role only gets the file rights that is required for him

  Glossary of common terms

– BASE revision

  •This is the revision the file or folder was in, when the last checkout, update or commit was run .

 - Branch

  You can create a branch off the main development line so as to develop a new feature without rendering the main line unstable. Or you can branch a stable release to which you make only bugfixes, while new developments take place on the unstable trunk. In Subversion a branch is implemented as a “cheap copy”.

 - Blame

  •This command is for text files only, and it annotates every line to show the repository revision in which it was last changed, and the author who made that change. Our GUI implementation is called TortoiseBlame and it also shows the commit date/time and the log message when you hover the mouse of the revision number.

- Commit

  •This Subversion command is used to pass the changes in your local working copy back into the repository, creating a new repository revision.

  – BDB

  Berkeley DB. A well tested database backend for repositories, that cannot be used on network shares. Default for pre 1.2 repositories

- Copy

  In a Subversion repository you can create a copy of a single file or an entire tree. These are implemented as “cheap copies” which act a bit like a link to the original in that they take up almost no space. Making a copy preserves the history of the item in the copy, so you can trace changes made before the copy was made. To ‘tag’ a version , create a SVN folder and use ‘Copy’ to copy the snapshot to the new folder

- Export

  •This command produces a copy of a versioned folder, just like a working copy, but without the local .svn folders.

 -FSFS

  •FS File system. A proprietary Subversion file system backend for repositories. Can be used on network shares. Default for 1.2 and newer repositories.

  – Diff

  •Shorthand for “Show Differences”. Very useful when you want to see exactly what changes have been made.

 - HEAD revision

  The latest revision of a file or folder in the repository.

  – Patch

  If a working copy has changes to text files only, it is possible to use Subversion’s Diff command to generate a single file summary of those changes in Unified Diff format. A file of this type is often referred to as a “Patch”, and it can be emailed to someone else (or to a mailing list) and applied to another working copy. Someone without commit access can make changes and submit a patch file for an authorized committer to apply. Or if you are unsure about a change you can submit a patch for others to review.

  – Merge

  –The process by which changes from the repository are added to your working copy without disrupting any changes you have already made locally. Sometimes these changes cannot be reconciled automatically and the working copy is said to be in conflict. –Merging happens automatically when you update your working copy. You can also merge specific changes from another branch using TortoiseSVN’s Merge command. •

  - Switch

  •Just as “Update-to-revision” changes the time window of a working copy to look at a different point in history, so “Switch” changes the space window of a working copy so that it points to a different part of the repository. It is particularly useful when working on trunk and branches where only a few files differ. You can switch your working copy between the two and only the changed files will be transferred.

  – Update

  This Subversion command pulls down the latest changes from the repository into your working copy, merging any changes made by others with local changes in the working copy

  – Working Copy

  •This is your local “sandbox”, the area where you work on the versioned files, and it normally resides on your local hard disk. You create a working copy by doing a “Checkout” from a repository, and you feed your changes back into the repository using “Commit”. •

  – Property

  In addition to versioning your directories and files, Subversion allows you to add versioned metadata – referred to as “properties” to each of your versioned directories and files. Each property has a name and a value, rather like a registry key. Subversion has some special properties which it uses internally, such as svn:eol-style. TortoiseSVN has some too, such as tsvn:logminsize. You can add your own properties with any name and value you choose

  – Import

  •Subversion command to import an entire folder hierarchy into the repository in a single revision

  – Relocate

  •If your repository moves, perhaps because you have moved it to a different directory on your server, or the server domain name has changed, you need to “relocate” your working copy so that its repository URLs point to the new location. •Note: you should only use this command if your working copy is referring to the same location in the same repository, but the repository itself has moved. In any other circumstance you probably need the “Switch” command instead.

  - Revision

  •Every time you commit a set of changes, you create one new “revision” in the repository. Each revision represents the state of the repository tree at a certain point in its history. If you want to go back in time you can examine the repository as it was at revision N.

  – Lock

  When you take out a lock on a versioned item, you mark it in the repository as un-committable, except from the working copy where the lock was taken out.

  – Cleanup

  To quote from the Subversion book: “ Recursively clean up the working copy, removing locks and resuming unfinished operations. If you ever get a working copy locked error, run this command to remove stale locks and get your working copy into a usable state again. ” Note that in this context “lock” refers to local file system locking, not repository locking.

Notes :-

  Integration with Bug Tracking

  •To tag each commit to a Bug ID, you can set one these one of these properties while configuring Subversion :- •bugtraq:url : Set this property to the url of your bugtracking tool. It must be properly URI encoded and it has to contain %BUGID% •http://issues.tortoisesvn.net/?do=details&id=%BUGID%

  •bugtraq:warnifnoissue •Set this to true, if you want TortoiseSVN to warn you because of an empty issuenumber textfield. Valid values are true/false. If not defined, false is assumed. •

  •bugtraq:logregex •This property activates the Bugtracking System in Regex mode. It contains one or two regular expressions, separated by a newline •If only one expression is set, then the bare bug ID’s must be matched in the groups of the regex string. Example: [Ii]ssue(?:s)? #?(\d+)

  Tools to integrate with VS.NET

  •AnkhSVN •VisualSVN

  Resources

  •Please read the Tortoise SVN Help which will be in the client •You could also read the SVN Book  in the site http://svnbook.red-bean.com/

•Open a command prompt

•Navigate to drive on which you wish to create the Subversion repository (say D:\)

•Create an empty folder D:\SVNx

•Create another folder within it called ‘MyWebApp’

•Go to D:\SVNx and type ‘svnadmin create – -fs-type fsfs MyWebApp’ 

•Now go into the new directory ‘myWebApp’

•See the files created by Subversion (dir)

• Go into the sub folder ‘conf’

•Edit the file titled ‘svnserve.conf’

•Uncomment the lines ‘anon-access’, ‘auth-access’ and ‘password-db’. Save the file

•Edit the file ‘passwd’

•Uncomment the 2-user/password combinations for harry and sally

•Change the user name (harry and sally) to the user names you wish to assign for the project to the developers in the team

•Change the passwords as required

•Save the file

•Create a batch file called ‘StartSVN.bat’

•Key in the value ‘svnserve.exe –d –r  D:/SVNx/myWebApp -–listen-port 3691’

•Run this batch file

•Now the Subversion server is up and running

•Now go to the folder, which has the project to be ‘imported’ into the above repository

•Right click that folder and choose the ‘Import’ option

•Key in ‘svn://localhost:3691/myWebApp’ in the URL field of the dialog box (Replace local host with the name or IP address of the server)

•Folders will now get imported into the Repository

•   To take hot backups use Svnadmin –hotcopy SourceRepositoryFolder BackupFolder

1)     have 3 projects – One is the UI project (JMS.UI)

2)     One is the Business Layer project (JMS.BL)

3)     One is the Presentation layer project (JMS.PL)

4)     In the UI, have a class like ProductMaster.aspx

5)     In the Business Layer, add a class for each business object – like Product.cs , which will have all properties with get and set and Save / Fetch methods

6)     In the presentation layer , add a class called ProductMasterPresenter.cs (or ProductEditor.cs)

7)     In this ProductMasterPresenter.cs, reference the BL

8)     Create a public interface like IproductMasterView , with the properties of the business classes needed by the ProductMaster and the EVENT of the Product class (like Save). (event EventHandler<EventArgs> Save;)

9)     In the ProductMasterPresenter class, create an instance of this interface as mView

10) In the constructor of this ProductMasterPresenter class, accept the interface as a parameter (IProductMasterView view)

11) In the constructor of step 10, Set the local instance of the view (mView) to the view that was received in the constructor :-

this.mView = view;

12) Next in the same constructor, call the Initialize method, in which you wire the Save event to the event handler. this.mView.Save += new EventHandler<EventArgs>(mView_Save);

13) Implement the mView_Save(object sender, EventArgs e) handler by creating an instance of the Product class and setting the properties from the view and saving it .

product = new Product();          product.Description = this.mView.ProductDescription;               product.ID = this.mView.ProductID;                product.Save();

14) In the ProductMaster.aspx.cs, reference the BL and the PL

15)  In the page load event or page init event , create an instance of the Presenter class , passing itself , the form itself to the constructor : this.mPresenter = new ProductMasterPresenter(this);

16)  In the ProductMaster.aspx.cs, implement the Presenter interface defined in the Presenter class :- public partial class ProductEditor : System.Web.UI.Page, IproductMasterView

17) Implement the properties of the IproductMasterView interface, by setting the text box values to them – public int ProductID

    {

        get { return Convert.ToInt32(this.txtID.Text); }

    }

18)  In the btn_Save click event of the UI form, call this handler this.Save(this, EventArgs.Empty);

19) Declare an eventhanlder here

public event EventHandler<EventArgs> Save;

20) Since there is an eventhandler in the Presentation class for this Save event, it will be invoked , which will then call the business class save method. Since the view (UI) was already passed to the presenter in the Load event , and that UI’s save event was wired to the Save event handler in the presentation class, when the Save event occurs in the UI, the save event handler of the presentation class is invoked.

Hi Chakravarthy,

That’s because the tables are very small. If you examine the execution plans of the UPDATE statements, you will see that SQL Server decides to use a full scan of the clustered index (essentially a table scan), instead of using the nonclustered index to support the filter. To “fix” this, you can use a hint to force SQL Server to use the index:

– Connection 1

BEGIN TRAN

UPDATE dbo.T1 SET col2 = col2 + ‘A’

FROM dbo.T1 WITH(index = idx_col1)

WHERE col1 = 101;

– Connection 2

BEGIN TRAN

UPDATE dbo.T1 SET col2 = col2 + ‘A’

FROM dbo.T1 WITH(index = idx_col1)

WHERE col1 = 102;

Of course, when the tables are bigger, there’s no need to use the hint since SQL Server will use the index (assuming it’s selective enough).

Take care,

Itzik

www.SolidQualityLearning.com

www.insidetsql.com

From: Chakravarthy [mailto:chakra@eldora-software.com]
Sent: Wednesday, May 02, 2007 8:18 PM
To: Itzik Ben-Gan
Subject: Self deadlocks

Mr. Itzik,

I have been browsing through your 2 books on T-SQL Querying and Programming for SQL 2005, and appreciate their quality.

However, there is a subtle variation of the Self deadlock scenario that i wanted to present to you, with your feedback if possible :-

1) I first create the 2 tables as in Listing 9-1 of your SQL Programming book (Appendix A below )

2) I then run this in one Query screen :

BEGIN TRAN

UPDATE dbo.T1 SET col2 = col2 + ‘A’ WHERE col1 = 101;

3) I run this in ANOTHER query screen (same Table, different row) :-

BEGIN TRAN

UPDATE dbo.T1 SET col2 = col2 + ‘A’ WHERE col1 = 102;

THIS IS ENDLESSLY WAITING

4) I then ROLLBACK BOTH and apply indexes as you have suggested :-

CREATE INDEX idx_col1 ON dbo.T1(col1);

CREATE INDEX idx_col1 ON dbo.T2(col1);

GO

5) I AGAIN try steps 2 and 3 above, AND Step 3 continues to WAIT. Can you try this and tell me what you think ?

Regards,

Chakravarthy.

Appendix A

SET NOCOUNT ON;

IF DB_ID(‘testdb’) IS NULL

CREATE DATABASE testdb;

GO

USE testdb;

GO

IF OBJECT_ID(‘dbo.T1′) IS NOT NULL

DROP TABLE dbo.T1;

IF OBJECT_ID(‘dbo.T2′) IS NOT NULL

DROP TABLE dbo.T2;

GO

CREATE TABLE dbo.T1

(

keycol INT NOT NULL PRIMARY KEY,

col1 INT NOT NULL,

col2 VARCHAR(50) NOT NULL

);

INSERT INTO dbo.T1(keycol, col1, col2) VALUES (1,101,’A');

INSERT INTO dbo.T1(keycol, col1, col2) VALUES (2,102,’B');

INSERT INTO dbo.T1(keycol, col1, col2) VALUES (3,103,’C');

CREATE TABLE dbo.T2

(

keycol INT NOT NULL PRIMARY KEY,

col1 INT NOT NULL,

col2 VARCHAR(50) NOT NULL

);

INSERT INTO dbo.T2(keycol, col1, col2) VALUES (1,201,’X');

INSERT INTO dbo.T2(keycol, col1, col2) VALUES (2,202,’Y');

INSERT INTO dbo.T2(keycol, col1, col2) VALUES (3,203,’Z');

GO

Features by default configuration…..