Distributed Code Help

Lonley Driver · Post by **Lonley Driver** » Mon Mar 10, 2003 2:45 am

I have to agree CVS is probably the best solution for what you're trying to do.

1. It's free.

2. It's already coded for you so you don't have to waste time developing (and it's open source so if you can futz around with it if you need to)

3. It has some really neat Front Ends.

4. Your programmers don't have to change the way they do things too much, all they need to know is how diffs work, checkout, and commit.

I'm not really sure what you'd be able to do with a PHP/MYSQL system, that you wouldn't be able to do with CVS. You'd basically be writing a new CVS.

-David

kmv · Post by **kmv** » Tue Mar 11, 2003 4:41 pm

database change tracking
ok.. thats the kicker.. i have no idea how to pull this off in cvs. but again, could be done via code sumissions as diffs into a acceptance queue.

How big would the test db be? They are just files after all...

Lonley Driver · Post by **Lonley Driver** » Tue Mar 11, 2003 8:41 pm

Hmm... database tracking changes... You could write a little front end in PHP where you can execute SQL and comments on it, and then you just take the SQL and the comments and throw it into another table. It'd probably take someone a whole 10 minutes to write.

This would allow you to keep a record of every single change to the database, with coder comments.

Use something like that and CVS, and you have what you need.

-David

dj-ohki · Post by **dj-ohki** » Tue Mar 11, 2003 8:48 pm

Lonley Driver wrote:This would allow you to keep a record of every single change to the database, with coder comments.

cvs does this natively IIRC. i was just refering to a sumission queue to make life easier on the admin side insted of having everyone with commit access.

Org Profile · Post by **Phade** » Tue Mar 11, 2003 9:01 pm

Hey,

It's more complicated than that. We need to track which pages use which database tables (hopefully down to the specific columns used), database changes, obselecense, and other database/page relationships, all of which CVS is incapable of tracking (as far as I know). I also want user management to be stored in MySQL so that can be easily integrated into the current user management system.

Phade.

dj-ohki · Post by **dj-ohki** » Wed Mar 12, 2003 2:43 am

Phade wrote:Hey,

It's more complicated than that. We need to track which pages use which database tables (hopefully down to the specific columns used)

custom database... that stores page documentation. i dont know of any tool that does this.

database changes, obselecense, and other database/page relationships,

audit trail? this is looking like custom backup build up with built in audits and all kinds of crazy crap. and is sounding like a wee bit of overengineering in some aspects.

all of which CVS is incapable of tracking (as far as I know).

true, CVS isnt the Right Tool For The Job for this. like i said, looks like this is gonna need a custom back end job before you even get started into anything serious project wise. im up a for a challange though.

also, you could write a parser that searches for dependency in sql statements.. that shouldnt be hard at all, but poses its own issues.

I also want user management to be stored in MySQL so that can be easily integrated into the current user management system.

easily done if you build up your own backend, which, for what you're wanting, is no trival task, but can be done.

kmv · Post by **kmv** » Wed Mar 12, 2003 5:33 am

Phade wrote:It's more complicated than that. We need to track which pages use which database tables (hopefully down to the specific columns used), database changes, obselecense, and other database/page relationships

I think you are attempting to address a different problem here - or perhaps this is what you meant from the beginning and I missed it.

The type of relationship tracking you are describing is provided by modeling tools (or CASE tools) not version control systems. In practice you need both: the modeling tool to model your relationships (tables, pages, etc.) and the version control tool to version your model files.

If the modeling tool includes direct support for the version control system then you can have multiple people working on site design at the same time.

I use Rose, and it can certainly do what you want but frankly it is probably extreme overkill for you - and it is expensive. I haven't used it, but I hear that GDPro is also pretty good. Off the top of my head I can not think of an open source modeling tool.

Google turned this up for me:
http://www.webreview.com/2001/05_18/dev ... ex01.shtml

It has a good introduction to modeling web pages, and it even recommends some tools, including a free one called "Objecteering UML"