Using Subversion and MODx for Team Development

Quote from: smashingred at Jan 27, 2008, 12:12 PM

This is one of the mysteries for me svn:ignore. I can’t figure out how to make the /manager and /install ignored. If the are in the repo they will end up becoming part of the working copy won’t they?

You can’t turn folders in a working copy into ignored files can you?

I have been able to copy the files into the working dir and make them ignored but then when I do a commit some files seem to not be placed back in the repo.

I am at a loss.

It always takes me several tries to get svn:ignore right. First, there is the global ignore list (usually including .svn and/or _svn at the least). Then there are file and folder properties which I can never remember how to set in TSVN.

I don’t think there is any way to ignore files that are already under version control. If you set them to be ignored beforehand, though, as far as SVN is concerned, they don’t exist. And I’m fairly certain that you can ignore folders as well.

Hope this helps,

Bob

You can set/change the svn:ignore property (and most svn: properties) at any time. For svn:ignore’d files/folders that exist, they will not be included in updates or commits or other operations made to a working copy (I can’t recall if svn:ignore affects export/checkout for resources that do exist in SVN; will have to look it up). Setting the properties ahead of time will prevent those resources from ever being seen as changes in your working copy. It’s useful say, to set on cache directories that exist in SVN, so the cache folder/files are ignored on svn updates. You can commit changes to the properties, or just set them in your working environment, depending on the usage.

Try this link http://svn.haxx.se/users/archive-2005-02/1557.shtml

I think they arrive to the same conclusion as you did, but they have two or three ways to
implement it.

Hope this helps...

This seems more confusing than when I started learning OOP. It seems that most version control systems are not suited for PHP/MySQL web development. I have read threads and threads on SVN with MySQL applications that depend on data as part of function. It seems that there is no simple solution. I have found MySQLdiff and I am working my way through this article on Evolutionary DB Design.

I am going to have to play with this but in PHP/MySQL application design there is application dependent data in the DB (especially in the pre 097 versions of MODx) that can change as part of the development process. As far as I see SVN (and most other Version Control Systems) have no way to automate the management of a db.

On another note, I have been thinking about whether I should build my repository around a finished install or around the full MODx package. The issue is that there are common snippets, templates and such that I would like to have in the application.

I guess I could build out the basic site with the stuff in the db and the other team members would have to overwrite the db data with and SQL Dump as well as with the versioned files.

Suggestions on this part are welcomed.

Cheers,

Jay

I’m not sure how it’s not well suited to PHP/MySQL dev, but you are correct subversion is for simple version control of file systems and does not help you do anything specific to databases. Data-driven applications have always provided difficulties in change management that require some additional tools and process. You might consider using something like Ant (if your team uses Eclipse) or Phing to wrap around separate tasks to dump the database (in a very specific way) and update the file system.

Jason,
On doing a search of the subversion mail list archive I found that very respectable OS people (including Bes Sussman) don’t have a simple-ish workflow for managing application dependent sql for web applications. Sure SVN is great for templating and scripting but web applications as well as other types of files (Lifehacker has a great article on using Subversion for document managment) usually have some tables that are not user data but common config. MODx has more than most CMSs in the DB and that makes an extra challenge so I may look at using Eclipse (I have it with Subclipse and PDT and QuantumDB installed but I haven’t really started using it as there is a bit of a learning curve compared to my current application/IDE e-texteditor which is a windows port of TextMate. I like the look of Eclipse but I would also not necessarily force my team to work in a specific environment (unless I am building a web application). One of my team members uses DW as his application of choice and he seems to be proficient (he is adept at creating clean markup.)
The

On my previous question, I’ve still been toying with the whole repository thought. Now I am thinking that I should just be including the assets folder in the repo. I may use a seperate repo for a custom version of MODx sans the assets folder for custom installs that would ensure that I can distribute the files with tracked changes.

I am going to run some tests and see what I like and I will let you all know what I come up with.

I may also hit up the Subverison mail list as well as the SitePoint forums.

Thanks and more suggestions are welcome.

I think somebody already mentioned this, but for group development I think using includes and having as much as possible in the file system would be helpful. With a simple include snippet even the template can be included from the file system. Although I don’t like to think what this might do to a production site’s responsiveness, but then you can always put templates and some other things back in the database when finished with development.

Susan,

I thought about using includes but that seems like a pain with a whole new set of problems.

If I look at my typical workflow:

• Data Analysis
• Information Structure
• Template Design (Graphic)
• Template Markup and Styling
• Place Templates into MODx
• Integrate Snippets, Chunks and Modules
• CSS revisions
• Site Testing
• Revisions
• Launch
• Manage/Bugfix

Only the first two steps would be done outsite of VC. I think I am going to map out a way to do the whole project under VC from Template Design all the way to deployment and then for management and updates. I don’t think that I will use includes to do what MODx already does even if it means more hassle versioning the DB. The issues with using includes with all the added complexity of training a Team to do something that feels unnatural in Modx. Granted, if I deployed a base Db with the includes already in the templates in the DB this would be less difficult.

I’ll test this over a few projects and sample workflow scenarios. It has to be easy to do even if it requires discipline over automation.

Thanks again.

Jay

I found www.daversy.org and think that is exactly what you/we are looking for,
but unfortunatelly it does not have a "database provider" (read it driver) for MySQL. It only
has Oracle and SQLite.

I contacted the developers and they indicate that they translated/ported the program from C++
to Python and one of them is working in the Postgres driver. They mention that it should not
be very hard to write a MySQL driver for somebody that knows Python and are willing to help
with advice and help understanding the code... anybody out there is a Python guru?

On checking diffs of mysql dbs I have found an application that looks promising but I have yet to test it: http://www.mysqldiff.org/

It doesn’t look as fancy as DAVERSY but it may be good enough for a small shop.

Cheers,

Jay