Forums

Biafra Republic has nicely setup a subversion server for MEKA.
The address is: svn://tprinteractive.com/meka
Subversion is a source control system (ala CVS, SourceSafe). Search Google for clients (standard command-line 'svn' client or TortoiseSVN client for Win32, for exemple) and information.

Right now the server has sources for 0.71.
I haven't really used it yet but I'll try to. If I manage to get used to that thing, I'm all for using it so that changes are logged and made available immediately. Regular contributors (if any) could then get write access to it in the future, etc.

I'm off, trying to catch up with my broken MEKA sources tree. =)

Reading thru TortoiseSVN help file, "5.2.1. Repository Layout" suggests the following:
There are some standard, recommended ways to organize a repository. Most people create a trunk directory to hold the "main line" of development, a branches directory to contain branch copies, and a tags directory to contain tag copies.

trunk/
branches/
tags/

'tags' would hold milestone, as in, official release version (0.70, 0.71, etc).

tags/0.70/
tags/0.71/

However, I'm not sure to understand what branches are for.

Case exemple: I'm now in the process of switching to hi-color mode and removing paletted mode usage from MEKA. This is huge work and requires lots of change in GUI related code. My working code has trillion of modifications, (including some improvements not related to the hi-color switch), but is overally broken (games themselves are not even showing proper colors yet).

Should I:
- Upload this in the 'trunk', meaning the trunk is the official main copy in development (knowing that it's kind of broken now)
- Create a 'branch', called for exemple 'hicolor_transition' and work on this branch. Then overwrite/merge the trunk with the branch once it's done?
- Work offline and submit in the trunk once everything's done?

I've been living in an igloo during the past decade, so I'm more interested in knowing what are standard/popular techniques. When should a branch be created?

Anyone has advices/tricks regarding SVN usage? I am walking alone into a dark unsafe corridor?

I'm using SVN at work now, with TortoiseSVN. It's not vastly different to CVS, just a bit more able to handle some things.

The trunk should always be working. If you break something, don't commit it until it's fixed.

If you're doing something big, which will cause a lot of breakage and it'd be useful to have versioning (so you can undo some changes, for example) and possibly collaboration (so others can help with it), make a branch. The hi-color transition is a good case for that. When it's done, you merge it back with the trunk, so the trunk is generally not broken.

Generally, you'd create a branch before you start on the big branch project, so it'd start off identical to the working trunk.

SVN is generally useful, although less well-supported than CVS in tools/IDEs etc. It's a great feature to start using, and avoids everyone having to figure out how to merge code changes, do manual diffs, etc. The only thing I've found is that generally, people are unable to understand how TortoiseSVN's conflict resolving diff tool works :)

Sorry for unburying the old thread, but here are my observations. I have looked closer at the kadu project, the IM program.
trunk is the active development line (0.5.0), with new features being added
branches (now only 0.4 is present) are open only for regression and documentation fixes
tags contain official releases

I have noticed that kde svn repository is organised in a similar way.

Hello. This is just a quick little message from the person behind the Subversion server...

A WebSVN Interface has been set up whick will allow users to...

http://www.tprinteractive.com/websvn/listing.php?repname=MEKA&path=%2F&sc=0

More Updates from the Subversion maintainer...

Bad news: The WebSVN Auto-Tarball and RSS feed features were disabled due to problems with the cache files not being properly removed. The RSS feed will not likely be enabled again...

Good news: A cron job to generate a nightly source tarball from the trunk has been finished and installed. Nightly tarballs are generated at midnight server time (US Eastern), from the sources in the trunk directory of the MEKA repository, so it might compile for you correctly, or turn your computer into a neutron bomb (j/k but you know what I mean...).

You can download the nightly tarballs here:
http://www.tprinteractive.com/goldeneye/meka/meka-src-nightly-trunk.tar.bz2

Author	Message
Bock Site Admin Joined: 08 Jul 2001 Posts: 8661 Location: Paris, France	Subversion server for MEKA Posted: Sun Aug 14, 2005 12:27 am
	Biafra Republic has nicely setup a subversion server for MEKA. The address is: svn://tprinteractive.com/meka Subversion is a source control system (ala CVS, SourceSafe). Search Google for clients (standard command-line 'svn' client or TortoiseSVN client for Win32, for exemple) and information. Right now the server has sources for 0.71. I haven't really used it yet but I'll try to. If I manage to get used to that thing, I'm all for using it so that changes are logged and made available immediately. Regular contributors (if any) could then get write access to it in the future, etc. I'm off, trying to catch up with my broken MEKA sources tree. =)

Bock Site Admin Joined: 08 Jul 2001 Posts: 8661 Location: Paris, France	Repository layout, suggestions? Posted: Sun Aug 14, 2005 1:26 pm Last edited by Bock on Sun Aug 14, 2005 2:24 pm; edited 1 time in total
	Reading thru TortoiseSVN help file, "5.2.1. Repository Layout" suggests the following: There are some standard, recommended ways to organize a repository. Most people create a trunk directory to hold the "main line" of development, a branches directory to contain branch copies, and a tags directory to contain tag copies. trunk/ branches/ tags/ 'tags' would hold milestone, as in, official release version (0.70, 0.71, etc). tags/0.70/ tags/0.71/ However, I'm not sure to understand what branches are for. Case exemple: I'm now in the process of switching to hi-color mode and removing paletted mode usage from MEKA. This is huge work and requires lots of change in GUI related code. My working code has trillion of modifications, (including some improvements not related to the hi-color switch), but is overally broken (games themselves are not even showing proper colors yet). Should I: - Upload this in the 'trunk', meaning the trunk is the official main copy in development (knowing that it's kind of broken now) - Create a 'branch', called for exemple 'hicolor_transition' and work on this branch. Then overwrite/merge the trunk with the branch once it's done? - Work offline and submit in the trunk once everything's done? I've been living in an igloo during the past decade, so I'm more interested in knowing what are standard/popular techniques. When should a branch be created? Anyone has advices/tricks regarding SVN usage? I am walking alone into a dark unsafe corridor?

Maxim Site Admin Joined: 19 Oct 1999 Posts: 14750 Location: London	Posted: Sun Aug 14, 2005 2:15 pm
	I'm using SVN at work now, with TortoiseSVN. It's not vastly different to CVS, just a bit more able to handle some things. The trunk should always be working. If you break something, don't commit it until it's fixed. If you're doing something big, which will cause a lot of breakage and it'd be useful to have versioning (so you can undo some changes, for example) and possibly collaboration (so others can help with it), make a branch. The hi-color transition is a good case for that. When it's done, you merge it back with the trunk, so the trunk is generally not broken. Generally, you'd create a branch before you start on the big branch project, so it'd start off identical to the working trunk. SVN is generally useful, although less well-supported than CVS in tools/IDEs etc. It's a great feature to start using, and avoids everyone having to figure out how to merge code changes, do manual diffs, etc. The only thing I've found is that generally, people are unable to understand how TortoiseSVN's conflict resolving diff tool works :)

Belegdol Joined: 27 Apr 2005 Posts: 50 Location: Poland	Posted: Thu Oct 13, 2005 7:04 am
Belegdol Joined: 27 Apr 2005 Posts: 50 Location: Poland	Sorry for unburying the old thread, but here are my observations. I have looked closer at the kadu project, the IM program. trunk is the active development line (0.5.0), with new features being added branches (now only 0.4 is present) are open only for regression and documentation fixes tags contain official releases I have noticed that kde svn repository is organised in a similar way.

BiafraRepublic Joined: 16 Mar 2005 Posts: 25 Location: Houston metro area, TX, USA	Posted: Sun Mar 26, 2006 3:00 am
	Hello. This is just a quick little message from the person behind the Subversion server... A WebSVN Interface has been set up whick will allow users to... Look at recent changes/revisions View the repository over the Web (for those having issues with your svn clients, etc.) Some other stuff I will probably recall later... http://www.tprinteractive.com/websvn/listing.php?repname=MEKA&path=%2F&sc=0

BiafraRepublic Joined: 16 Mar 2005 Posts: 25 Location: Houston metro area, TX, USA	Posted: Mon Sep 11, 2006 6:38 pm
	More Updates from the Subversion maintainer... Bad news: The WebSVN Auto-Tarball and RSS feed features were disabled due to problems with the cache files not being properly removed. The RSS feed will not likely be enabled again... Good news: A cron job to generate a nightly source tarball from the trunk has been finished and installed. Nightly tarballs are generated at midnight server time (US Eastern), from the sources in the trunk directory of the MEKA repository, so it might compile for you correctly, or turn your computer into a neutron bomb (j/k but you know what I mean...). You can download the nightly tarballs here: http://www.tprinteractive.com/goldeneye/meka/meka-src-nightly-trunk.tar.bz2

Forums

View topic - Subversion server for MEKA