You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Jan Horák <ho...@gmail.com> on 2010/03/09 23:50:08 UTC

Repository structure

Hi, after a long pause I started to analyze the SQL backend again. I'm 
just studying the BDB and FSFS backends to collect all requirements 
needed to design the SQL scheme.

1) I have already taken a look at the SVN source to find out, which 
operations in the DAG (repository tree) are used mostly while an 
ordinary checkout is performing (or generally while the repository is 
reading). It seems like the most used operation is retrieving the 
ancestors in the tree, while reading a subtree is not used so often, am 
I right?

2) The aim of the second question is to imagine, how the repository 
looks inside (in BDB backend for example). I've prepared very simple 
snapshot (see the attachment, hope it will be attached) of the 
repository objects after two simple commits.
   * The first commit will add a directory '/b' and files 'a.txt' and 
'/b/c.txt' to the root in an empty repository.
   * The second commit will change the content of the file 'a.txt', add 
a file 'd.txt' and rename the directory '/b' to '/bb'.
The attached picture shows how I imagine the repository (very simple, no 
copies and changes table, many attributes are not shown). If anybody 
from developers has a second, please, could you take a look at the 
picture, if there is not any critical mistake?

Thanks for answers

Jan Horak

Re: Repository structure

Posted by Jan Horák <ho...@gmail.com>.
Attempt number 3.. with .txt extention...

Jan Horák
E-mail: horak.honza@gmail.com


Dne 10.3.2010 11:08, Stefan Sperling napsal(a):
> On Wed, Mar 10, 2010 at 12:52:02AM +0100, Jan Horák wrote:
>    
>> Sorry, the attachment forgotten :)
>>      
> It didn't reach the list. Usually a .txt extension helps.
>
> Stefan
>    

Re: Repository structure

Posted by Jan Horák <ho...@gmail.com>.
Attempt number 3.. with .txt extention...

Jan Horák
E-mail: horak.honza@gmail.com


Dne 10.3.2010 11:08, Stefan Sperling napsal(a):
> On Wed, Mar 10, 2010 at 12:52:02AM +0100, Jan Horák wrote:
>    
>> Sorry, the attachment forgotten :)
>>      
> It didn't reach the list. Usually a .txt extension helps.
>
> Stefan
>    

Re: Repository structure

Posted by Stefan Sperling <st...@elego.de>.
On Wed, Mar 10, 2010 at 12:52:02AM +0100, Jan Horák wrote:
> Sorry, the attachment forgotten :)

It didn't reach the list. Usually a .txt extension helps.

Stefan

Re: Repository structure

Posted by Stefan Sperling <st...@elego.de>.
On Wed, Mar 10, 2010 at 12:52:02AM +0100, Jan Horák wrote:
> Sorry, the attachment forgotten :)

It didn't reach the list. Usually a .txt extension helps.

Stefan

Re: Repository structure

Posted by Jan Horák <ho...@gmail.com>.
Sorry, the attachment forgotten :)

Dne 10.3.2010 00:50, Jan Horák napsal(a):
> Hi, after a long pause I started to analyze the SQL backend again. I'm 
> just studying the BDB and FSFS backends to collect all requirements 
> needed to design the SQL scheme.
>
> 1) I have already taken a look at the SVN source to find out, which 
> operations in the DAG (repository tree) are used mostly while an 
> ordinary checkout is performing (or generally while the repository is 
> reading). It seems like the most used operation is retrieving the 
> ancestors in the tree, while reading a subtree is not used so often, 
> am I right?
>
> 2) The aim of the second question is to imagine, how the repository 
> looks inside (in BDB backend for example). I've prepared very simple 
> snapshot (see the attachment, hope it will be attached) of the 
> repository objects after two simple commits.
>   * The first commit will add a directory '/b' and files 'a.txt' and 
> '/b/c.txt' to the root in an empty repository.
>   * The second commit will change the content of the file 'a.txt', add 
> a file 'd.txt' and rename the directory '/b' to '/bb'.
> The attached picture shows how I imagine the repository (very simple, 
> no copies and changes table, many attributes are not shown). If 
> anybody from developers has a second, please, could you take a look at 
> the picture, if there is not any critical mistake?
>
> Thanks for answers
>
> Jan Horak

Re: Repository structure

Posted by Jan Horák <ho...@gmail.com>.
Sorry, the attachment forgotten :)

Dne 10.3.2010 00:50, Jan Horák napsal(a):
> Hi, after a long pause I started to analyze the SQL backend again. I'm 
> just studying the BDB and FSFS backends to collect all requirements 
> needed to design the SQL scheme.
>
> 1) I have already taken a look at the SVN source to find out, which 
> operations in the DAG (repository tree) are used mostly while an 
> ordinary checkout is performing (or generally while the repository is 
> reading). It seems like the most used operation is retrieving the 
> ancestors in the tree, while reading a subtree is not used so often, 
> am I right?
>
> 2) The aim of the second question is to imagine, how the repository 
> looks inside (in BDB backend for example). I've prepared very simple 
> snapshot (see the attachment, hope it will be attached) of the 
> repository objects after two simple commits.
>   * The first commit will add a directory '/b' and files 'a.txt' and 
> '/b/c.txt' to the root in an empty repository.
>   * The second commit will change the content of the file 'a.txt', add 
> a file 'd.txt' and rename the directory '/b' to '/bb'.
> The attached picture shows how I imagine the repository (very simple, 
> no copies and changes table, many attributes are not shown). If 
> anybody from developers has a second, please, could you take a look at 
> the picture, if there is not any critical mistake?
>
> Thanks for answers
>
> Jan Horak