notes-computer-sourceforge

Around 09-18-2008, sourceforge switched their webhosting setup so that the persistent website directory, which used to be in /tmp/persistent/$group, is now in ~/persistent. They didn't move your old data into the new directory, you have to do that manually. The old data resides in /home/old-persistent, which can only be reached via sftp. Here's a command to copy it (into a subdirectory of the group directory called "old-persistent"):

group=interwiki; user=bshanks; cd /tmp; lftp -e "open sftp:web.sf.net; login $user; cd /home/old-persistent; mirror $group; exit"; scp -r $group $user,$group@web.sf.net:old-persistent

---

I would like to use donations to establish bounties to encourage other developers to implement specific features for this project.

Please contact me if you'd like to donate money and have it be used in a specific way (bounty or otherwise).

Legal disclaimer to avoid getting sued:

However, I can make no guarantees as to the usage of the money. For example, although it is unlikely, perhaps I will suddenly totally lose interest in open source and not spend the time to set up the bounties. Perhaps no one will take the bounties. Perhaps my girlfriend will get cancer and I'll spend the money on her treatment. I don't have the resources to set up a separate account to make sure that the money stays separate from my personal money. In other words, although I'll most likely spend the donations on bounties, I wish to make it clear that no matter what happens, you cannot sue me for misusing the money.

---

btw, i finally started using sourceforge cvs for some of my projects. The CVS (and the rest of SF) was easier to setup than i had thought that it would be; I didn't set up CVS for awhile because I assumed it would take forever, but it didn't.

If you're holding back b/c you are afraid of setting up SF/SFcvs, I suggest that you go ahead and do it, it won't be that bad.

All you have to do is:

(1) ssh cvs.sourceforge.net (after you login, you'll be immediately logged out -- but this initializes your user account on the CVS machine)

(2) execute "export CVS_RSH=ssh" (or the setenv equivalent) in your shell each session before using CVS

(3) whenever you use the command "cvs", use the argument "-d:ext:username@cvs.sf.net:/cvsroot/PROJECTNAME" in addition to whatever other args you have

(to take care of (2) and (3), I have a bash shell script for each project on my PC called "sf.sh", containing

	export CVS_RSH=ssh
	alias sfcvs='cvs -d:ext:bshanks@cvs.sf.net:/cvsroot/PROJECTNAME'

(with PROJECTNAME different for each project's script).

I execute the script at the beginning of each session, and then I use the command "sfcvs" instead of "cvs".

(4) On SF, you don't have to do the "cvs init" command. So, the only thing left to do to start out is to put the first revision into the repository. You can read about this at

http://sourceforge.net/docman/display_doc.php?docid=14033&group_id=1#commandimport

but, basically, it comes down to "cd"ing to INSIDE of your project directory (not to just above it!), and then issuing a command like:

cvs -d:ext:username@cvs.sf.net:/cvsroot/PROJECTNAME import -m "Initial import (log message goes here)." MODULENAME VENDOR start

(MODULENAME is for if your project eventually contains different "modules", I've just been making it the same as PROJECTNAME so far. I don't think that VENDOR really matters -- I've been using "bshanks". For that matter, I think "start" is arbitrary too-- it's just a "release tag", which is just a word to use to reference this revision, i think)

I might have missed a step, but I think that's it. Just three commands, really;

ssh cvs.sourceforge.net

export CVS_RSH=ssh

cvs -d:ext:username@cvs.sf.net:/cvsroot/PROJECTNAME import -m "Initial import (log message goes here)." MODULENAME VENDOR start

list of things to do when making a new sf (sourceforge) project:

ln -s /home/groups/____ rmdir ____/* cp -r skelaton/* ___

chmod -R a+rw ____/usemodDb

change group ID, title, cookie title in usemodDb/config change groupID and URL in htdocs/index.php, usemodDb/config (LogoUrl?) change cgi-bin/wiki.pl to point to correct datadir

publicity info/categorize in trove

upload release

turn off unused trackers, forums, lists, tasks, cvs, docs, news. set preferred response mechanism.

put wiki's RSS into your aggregator

--- also, put up 2 line CVS shell script sf.sh

ssh -t bshanks,PROJECTUNIXNAME@shell.sf.net create cd /home/scm_hg/P/PR/PROJECTUNIXNAME hg init REPONAME chmod -R g+w REPONAME shutdown ^D hg clone ssh:bshanks@PROJECTNAME.hg.sourceforge.net/hgroot/PROJECTNAME/REPONAME

rsync -e ssh file.zip jsmith,fooproject@frs.sourceforge.net:/home/frs/project/f/fo/fooproject/Rel_1/