PureBasic Forums - English

Just wondering how other people do things..

I've never done anything beyond playing with bits and pieces of code, and my Text RPG will be my first "big" project to work on.. I'm curious when people start attaching Version Numbers to their work? Do you do it from inception? Do you do it on the first compile to test a particular system? Or do you wait until you have what resembles an actual program (sans features) before slapping a number on it?

To a lesser extent I am wondering about CVS (Source Code Management). While I don't work within a group of programmers (Just Me, Myself, and I!) Certain aspect of CVS may or may not appeal to me as I work on development. I really like the idea of having a central DB where I store my Source Code, that archives previous versions. I always worry that I will make some changes to a file in haste and later regret it because I broke something, or made a series of changes that precipitated more changes that just royally screw something up, etc..

So it would be nice to "roll back" if I need to... without having to manually archive all my work. I don't really get the concept of "Branches" or "Streams" or the various terms like that some CVS systems use as their paradigm of operation, but I'd still like to look into it..

Does anyone here use CVS for solo work? Are there any good free CVSes that are easy to work/learn?

Or should I just create a fancy Shell Script that can create/sync an archive every time I run it

Even if you are the only developer on a project always use some form of version control. I currently use mercurial and it's great, free and fast.

P.s. Don't use cvs it's old and there's better available.

Also have a look at Fossil for SCM:
http://www.fossil-scm.org/index.html/do ... index.wiki

Although I am not using SCM at the moment it's one of the SCM solutions that appealed to me.

Fossil is written by the author of SQLite.
It is free and easy and allows for a simple web server view of your timelines and changes, etc.
There is no GUI per se, but you will easily create your own custom one with PB code.
Your repository(all the files & stuff you want tracked) is housed in a single SQLite database file!
Very elegant.

You can version ANYTHING anytime... no rules or laws about it!

I use CVS constantly... I code the base of the program. When that is working... I start adding stuff I want to add... so I save the BASE coed as PROJECTNAME-000.pb (obviously I replace PROJECTNAME with the name of the project!) Then I add a feature or two. If the code runs I SAVE AS: PROJECTNAME-001.pb and on and on... I also make a backup folder in my development directory INSIDE the project folder and frequently unclutter the primary folder by moving past versions (up to a point) into it...

That way, if an error or bad code I can attempt to run earlier version and compare and try to figure out what happened!

EXTREMELY important in real world application development.

ALSO: In developing mY POS I actually created EACH SCREEN seperately in the VD and cut and pasted the code into the main program just so I could make changes to each window more easily. SAVED HOURS!!!

Good luck whatever you do: be consistent... SAVE OFTEN! and have phun!!!

I use git or subversion myself, locally hosted and mirrored on my web development machines. Dead handy for looking after all my programming needs.

I run a virtualmachine on my dev machine which runs virtualmin (virtualmin is a server control panel). Out of the box virtualmin offers a control panel to manage SVN datastores. So when I start a project, I just make a new SVN database. I then installed Trac (though that requires command line knowledge so it's not particularly simple) which gives me a nice place to put notes and documentation and stuff. But using just SVN is dead easy.

I use tortoise svn on my client machines to work with the SVN server.

About once a week I shutdown the server and make a backup of the virtual drive and put it on a rotating backup on a different machine.

Thanks for the tips, they helped at lot.

Mercurial (via TortoiseHG), or if you don't mind the 30 day version cut off: Dropbox (every single change is recorded though!)

Well, I'm a big fan of SQLite (It will be the backbone of my Text RPG's data storage needs), so I am going to read up on Fossil and see if it will appeal to me, then give it a shot.

We are using TortoiseSVN, a subversion system for our project.

I've been using version numbers since, euh, the stone age I guess. No fancy management system here.

vx.yyz

v = v
x = main version, typically a major rewrite
yy = sub version, version up to 99, then I'll just be inconsistent and increase x
x - working version
a - alpha release, early release, I used this in the past when writing commercial stuff, not so much anymore
b - beta release, aka release candidate, not planning to do much on it anymore

Typical progression:

v0.01x
v1.00a
v1.00b
v1.00

I also use numbered backups using CodeCaddy, ie. my backup folder contains hundreds of previous working versions, allowing me to retrace my steps.

Then again, I'm a one man team and few people use my crap, so there's little need for sharing documentation and bugs online

blueznl wrote:I've been using version numbers since, euh, the stone age I guess. No fancy management system here.

Then again, I'm a one man team and few people use my crap, so there's little need for sharing documentation and bugs online

Ha!
I was the same way for almost as long...
But really, Fossil kicks A$$!
I do not have to deal with compression and archival and versioning crapola anymore.
It is not just for team development.
Once you grasp the few simple commands involved, I guarantee you will never ever go back to manual methods.

It's definitely all "new" to me... I was kind of fretting because Fossil isn't GUI driven or anything... I do like my GUIs, but I think I should get the hang of it, when I need to start using it.