Slashdot Mirror

Have you checked out current? We just rolled it out a few weeks ago in our lab and it is working great. If you have a big investment in 8.0 you could roll your own RPMs once support dies at the end of the year. A big pain compared to the old way, but you should be able to tweak the sources or upgrade the few packages you need from the current redhat version for a few years (i.e., when Redhat releases a security update, grab thier SRPM and rebuild it on 8.0, or use it to patch the 8.0 version. This can get kinda tricky, but I have done similar things before

[root@odin root]# up2date nmap-frontend
Installing...
1:nmap [100%]
2:nmap-frontend [100%]
The following packages were added to your selection to satisfy dependencies:

nmap

For example, for the C program that I'm writing right now, I decided to use GLib -- the base utillity library used by GTK.

I initially chose it for portability reasons, but soon discovered it had a wealth of cool stuff in it. In addition to providing the standard data structures (trees, hashes, linked lists), it also has a string type ( GString, ) which handles a lot of the string issues that C programmers get bogged down with.

A lot of the gotchas (buffer overflows, et. al.) mentioned in this article have to do with these string issues, and using GLib's GString data type has enabled me to avoid those.

There is another library similar to GLib, The Apache Portable Runtime, used in the Apache webserver, and also in Subversion.

In addition to all this, I'm using XML as the storage format for my program, mostly because libxml takes care of the file parsing issues so I don't have to.

Bottom line, choose your libraries carefully, they can make a world of difference.

Thomas

The command set does have differences, which would be natural since SVN offers quite a few enhancements over CVS.

Speaking from someone who has used SVN, CVS and perforce, I must way that the branch mechanism of CVS and the whole idea of 'tags' is the most annoying, even more annoying than non-atomic changesets.

The most popular free software source management tool (CVS) is a complete p-o-s in many respects and unsuitable for large projects and for those with automated builds.

A proper comparison would be with Subversion. CVS, while certainly not a POS, is showing its age. Subversion is supposedly "CVS brought up to date".

Why did Linus go to BitKeeper in the first place?

Often that layer is a DB - database. I suggest you to try ZODB, database in Zope, it's very good to handle files as documents - with many unified metadata about files.

Another good example to study is Subversion, which is revisionining/versioning metadata-management layer on a top of a regular FS.

You may research and find some software implementing a layer (on a top of a regular FS) specially designed to handle MP3 playlists. But again, that would be a layer on a top of FS, not a filesystem by itself.

One thing that always kept me from taking time to do things the right way is that it takes a long time to figure out where to start. Doing it right actually does not take a super long time when you know exactly what you need to do.

Doing things the right way myself is one thing. But getting other people on the team to do it the right way take can a LOT of explaining, and we end up arguing little points that are a total waste of time.

This project on freshmeat a few days back. It's HTML templates for test plans, use cases, project risks, etc. That could help do things the right way a lot quicker.

Can you think of even one Java application that you use on your desktop and like?

Actually, yes. Jedit.
And yes.
And yes.
And so on.

I like Python too, and choose it over Java if it's really applicable. But not understanding 'the allure of Java' is a little overinflated don't you think? The apps above are good reasons to use Java, and they're not even Javas prime homeground. Maybe you haven't done any serious server side programming or tight client-server stuff, but be ashured that there are very good reasons to use Java. Especially with a big team and an emphasis on solid OOAD.

Version control definitely belongs in the filesystem. Clearcase may have a lot of implementation uglies, but as a concept it clearly works. Filesystems manage files, and that should include managing file versions.

One place to look for (sort of) file system versioning support is Subversion, an Apache-licensed replacement for CVS. You can mount this via webdav_fs, and you get automatic versioning that works with just about any application, although right now you need to use a separate client to get access to older versions of the files. Still in alpha, but very stable and worth checking out.

--Paul

We can also run "up2date packagename" and it will grab it... as long as it was something shipped with the distro, that is. And up2date does dependencies for you, which is nice. Unfortunately, up2date doesn't grab programs that weren't in the distrobution. :(

You can easily set up your own up2date server, using current. Apart from letting you add your own packages, it will also save you a bunch of bandwidth, if you run multiple RedHat systems.

Of course there is no perfect solution, but a really good repository would help developers reuse free software.

If I could go to a single place and search a large database of open software components easily, I would be more inclined to reuse fee software. Sourceforge or the Gnu free software directory do not do this well. I need much more than just a name and short description and I don't have time to spend weeks trying different halfway completed projects every time I want to reuse something.

I would like to see something that is actually useful. For example, the repository should include design and implementation information for each component, as well as complete test results and profiling results. And I should be able to search based on all that and more. Basically, it should make the search as painless and accurate as possible. It should also be managed so that the only correct information is allowed.

I am working on a project called Hoot that aims at centralizing and standardizing open software process and engineering. A repository like this is one of the goals. Not much has been accomplished yet, mostly due to poor leadership ;). If you want to help join the dev list.

My approach to these things is to use a version control system. This is great for several reasons:

• It makes it really easy to work from different machines, which I do all the time. I've got my main desktop, my laptop, my firewall (which stores its configuration in version control), my dedicated server, plus my desktop and Linux machine at work. Oh and another firewall at my parents' house. Just "svn up" on a given machine and I've got the latest version of everything.
• I can see all the history of everything without ever wondering if "finalfinal02_B.*" is really final, to use your example. Whenever I make a change, I make a log message. Usually a pretty short one (for text-ish files, the diff is often self-explanatory) but it's enough. I can pull any version easily, see what changed between two, etc.
• I have lots and lots of data safety without much effort checking it in uploads it to my server, which is backed up every night to tape by my ISP. And I have working copies on many different machines, so at least the latest version of everything gets sent to another machine whenever I type "svn up" from it. It would take an act of God for me to lose my data.
• It's easy! I tend to use the commandline tools, but there are good GUIs available, too. At work we use TortoiseCVS, which integrates with the Windows shell. There's also TortoiseSVN (based on Subversion), which is more advanced but a little more bleeding edge.

I actually use Subversion for not only code but also configuration files, school projects, anything really. The only thing I can think of that it probably wouldn't be suitable for is huge media files, where the storage space of the repository history might be prohibitive. You mentioned archiving to DVD, so you might fall into this category, but most people should give this method a try.

In particular, their Windows development tools and office suite still completely outclass the OS equivalents.

I have to call bullshit on this one. Microsoft development tools (i.e., those sold by Microsoft) are, contrary to popular religion, not the cat's hind end. I'm a developer and I've found that Visual Studio .NET and Source Safe and nearly every other MS "product" I try is the most ridiculously bloated and cumbersome tool imaginable. I really can't understand how people use them to do real work.

The GNU/OpenSource/POSIX LIDE or Loosely Integrated Development Environment (e.g., autoconf, automake, make, gcc, etc.), however, is standardized for nearly every platform now (yes, including Mac OS X and Windows) and Just Makes Sense(tm). By this I mean these are good, simple and extremely flexible tools which are configurable to work the way you do (not work you the way its developers want to). If I see another god-damned fscking "IDE" which "tightly" (meaning inflexibly) integrates to a broken source control system like CVS or VSS and doesn't easily allow me to drop-in my own replacement like Subversion or Perforce, I'm going to vomit.

By the way, these are also the same tools which are used to teach computer science and programming courses in Universities across the world (at least for now). Graduate with a BS in computer Science and chances are, unless you came from a crack-pot or sell-out operation, you are already trained in their use and can develop software for nearly every popular platform on the planet.

As far as office products are concerned, I'm not a publisher, but I do write a fair amount of documentation and give a number of presentations annually, and I find that OpenOffice 1.0.1 (distributed with RedHat 8.0) is more than sufficient for my needs and is being improved at an amazing rate. In fact, our entire office (business and technical staff alike) uses it without complaint.

How that product will turn out remains to be seen, but were I thinking of using CVS for any source control, I'd look into SVN first.

I just got a job. Yeah for me!!!! And we're using Perforce. I can't make an intelligent judgement about it yet. I also want to try BitKeeper. I'll probably switch over to it at home once I'm knowledgeable about Perforce. Then I'll be able to make a pretty good comparison.

ArgoUML is a Java based (and biased) UML development toolkit - well worth investigating

Tigris itself contains many other open source software engineering tools for design, construction, testing, deployment and more


Synopsis:
Tigris.org Community Scope
Tigris.org is a mid-sized open source community focused on building better tools for collaborative software development.
You will not find thousands of unrelated projects here: every project fits into the Tigris vision.
You will not find dead projects here: every project is welcomed into the community with a commitment to see it through and active developers cycle among related projects.
Tigris.org is hosted by CollabNet, but the Tigris vision is one for the entire open source movement and one that has attracted senior open source developers from many organizations.

ArgoUML is a Java based (and biased) UML development toolkit - well worth investigating

Tigris itself contains many other open source software engineering tools for design, construction, testing, deployment and more


Synopsis:
Tigris.org Community Scope
Tigris.org is a mid-sized open source community focused on building better tools for collaborative software development.
You will not find thousands of unrelated projects here: every project fits into the Tigris vision.
You will not find dead projects here: every project is welcomed into the community with a commitment to see it through and active developers cycle among related projects.
Tigris.org is hosted by CollabNet, but the Tigris vision is one for the entire open source movement and one that has attracted senior open source developers from many organizations.

I can strongly suggest using Dia in conjunction with tedia2sql. You can do the graphical design in Dia and then convert the result to SQL for various databases (Oracle/MySQL/PostgreSQL/...).
It is really fast and stable, it supports tables with indexes, primary keys, foreign keys, ...

However, those are largely things we don't control. The controllable factors of success are more interesting to me. I guess it's because there are lessons on software engineering here. Cool projects can be run into the ground, and tiny niche projects can do well if they're well-run.

Hands down, the best nuts-n-bolts coverage I've ever seen on important issues to successfully developing open source is in a book by Karl Fogel, Open Source Development with CVS. Fogel's one of the developers behind CVS and it's planned successor, Subversion

The book is an interesting paradox: it has 1/2 the chapters GPL'ed. When I started working with CVS, they were useful enough that I bought our development team two copies of the book. Then I read the rest of it... and those are the chapters I'm talking about. Absolutely, they're the best summary of what it takes to successfully run a GPL-ish project. (Ironically, they've GPL'd the technical detail chapters and you have to buy the book to read the parts that talk about things critical to the success of an open source project).

Success is helped by things like doing lots of releases (seeing progress gets others to buy in, and not seeing progress leads to people quitting in frustration) and only adding features you need (let someone else add the features they need). There's a lot more here, but I'm not about to steal Fogel's thunder. Many of these are ideas that are effective in regular development, especially on custom coding projects within big companies.

The focus on GPL code is not the same as on shrinkwrapped products: you're not trying to add features just to add selling points. You're trying to get more people to use the project.

I hate VSS with a passion. We've been using it as our Source Control system for many many years where I work. Lately, we get corrupt databases about every week or so. How does it happen? Usually people are connecting to VSS from home through a VPN, and the tunnel breaks in the middle of a checkin to the database. Since VSS's database is managed by the client, if it fails halfway through it is easy for the database to get corrupted. Additionally, the entire VSS database is writable for our entire development staff, hence no access control at all.

It is easy to fix a lightly corrupted VSS database with their dumb tool. Who the hell wants to do that, though?

So once upon a time, the database got corrupted and wouldn't magically fix. A couple of us showed off CVS to our manager and convinced him, eventually, to give CVS a try. We had it integrated on every developer's box through Jalandi Igloo, and moved all the projects over to a CVS server.

The experiment failed, and we're back on VSS now. Why did it fail? The developers found that Jalandi Igloo was about an order of magnitude slower than VSS's visual studio integration. Personally, I didn't have any problem with it. Occasionally Igloo would hang for a few minutes during a simple operation like a commit or a get. It was a pain, I'll admit, but we got many things out of it: A source control database that was text based and easily correctable, a source control system with sane branching, and the happiness of moving away from VSS. But the developers complained, some of them saying that utter frustration at CVS' speed was causing them to loose 100% of their productive time because they were so frustrated. [What dicks, eh? How can a slow source control system keep you from coding at all?] Others complained of more realistic productive time losses.

Anyways, we're back on VSS now. So now, when I go to edit a file that isn't checked out, sometimes it takes 5 minutes to check it out through visual studio. It just depends on who else is accessing the database at that moment, and how it gets locked. *le sigh*.

On a related note, I highly recommend Subversion as a source control system rather than CVS these days. The Subversion website could really use some work on becoming focused towards people who are interested in the software, rather than their developing of it, though. But I don't know of any visual studio integration worth mentioning, yet.

We started using cvs when OSX came out and had nothing but success with it. It has its limitation but it is straight foward and I even think it is Open Source (correct me if I am wrong) so if you don't like something, just change it!

True, but it's an open-source dinosaur. The code has been so badly hacked over the years that some influential guys decided it needed a huge rewrite and that's subversion. When you get *really* large scale CVS use (like the GCC project) that it uses rcs and file locks becomes quite a pain, and some of the operations had bad O() orders on file count. (There's even duplication of the core code for network client and file repository access.)

That's not to say that CVS is stable and very usable for small-scale. But if you want to enhance it, give your time to subversion.

I personally like CVS very much. But I think that I am going to like Subversion
Subversion
even more, in a few versions time. If you can wait a bit more than that you might want to look at an IBM research project called Stellation Version Control
Stellation

CVS is quite nicely integrated in that super IDE, Eclipse. And the two others have integration underway.

CVS may be the worst version control software in wide use.

Those of you who use and hate CVS may wish to check out SubVersion, a CVS replacement which leverages WebDAV via an Apache module. It can be used locally (i.e., sans server) as well, just like CVS. It's still in beta, but we use it and it's served us a lot better than CVS has. The only problem is some dev tools have (shortsighted) tight integration with one particular VCS or another (like CVS or MSVSS), so using an alternative breaks the "convenience" offered in these tools.

But I say fuck it. Tight integration is for wussies. Pipes and environment variables are the only integration I need.... ;o)

Or at least i think so. If I understand this correctly then sourceforge is just like sourceforge.net - a hideously large database of projects. I don't think that's what you are looking for. CVS is tested and good. But it has some problems. Go check out subversion too - subversion.tigris.org, it's a cvs-like tool that has fixed many of cvs' inconveniences.
and BitMovers BitKeeper is used for the Linux-kernel, so it has proven to be very reliable and extendable too.

I've been running RedHat 9 (ISO's downloaded via BitTorrent) all week on my desktop machine (10:32:32 up 4 days, 19:58, 21 users, load average: 0.66, 0.44, 0.86) and it seems fine, the first thing I did was install apt-get to make updates nice and easy, the next thing I discovered was that there is no WindowMaker, which is a real shame, the reason stated by RH is "developer resource constraints". I hope they put it back in, anyhow, this is how I get it working with RH 8 SRPMS:

1. Install the WindowMaker SRPM from RedHat 8.0 updates.
rpm -ivh WindowMaker-0.80.1-5.src.rpm

2. Edit the SPEC file:
vi /usr/src/redhat/SPECS/WindowMaker.spec
Change this:
%define autoconf_ver -2.53
to:
%define autoconf_ver -2.57
Change this:
autoheader%{autoconf_ver}
to:
autoheader
Change this:
autoconf%{autoconf_ver}
to:
autoconf
(I don't know what the implications of this is -- I just played with the SPEC file till it worked...)

3. Rebuild:
rpm-build -bb /usr/src/redhat/SPECS/WindowMaker.spec

4. Install:
rpm -ivh /usr/src/redhat/RPMS/i386/WindowMaker-0.80.1-5.i38 6.rpm \
WindowMaker-libs-0.80.1-5.i386.rpm

I also got wmapm, wmclock and wmix SRPMS from RedHat 8.0 and rebuilt those.

There has been some discussion on the wm-user list about doing some RPMS for RedHat 9.

One thing that is great is the inclusion of Subversion -- I installed subversion from scratch on RH 8 and it took some time...

That looks like a nice, standard mis-informed Slashdot post to me.

For Windows, there is CVS NT which works great. Then you can use Igloo for IDE integration. Or the even cooler Tortoise CVS for explorer integration (an easier to use source control tool I have yet to find).

OR, there is always Subversion if you don't want CVS. Subversion also has a Tortoise port.

Or, back in the closed source world, Perforce and I'm sure a handful of others.

As for free software in java, try looking at all the serverside tools.

jakarta
jboss
(and about a million other projects...)

for client developer tools
tigris

I don't understand the distinction of development shops. If anything, developers are less tolerant of poorly performing tools than consumers, at least in my experience...

I saw it mentioned briefly before, but I think most people have missed it. I saw it mentioned some time ago on Linux.com:

Current is an open-source implementation of an up2date server.

I've used it, and it does work, however I'm waiting for multple channel functionality and some other features before I switch to using it instead of apt.

...is under development here.

There's ArgoUML, but I've used it very little. Haven't used Umbrello at all. Would be interested in a head-to-head comparison.

Subversion is written in C because it's an Apache module, and Apache is written in C. Check for yourself, version 0.17 has just been released.

Just a suggestion, based on similar experience with bug tracking software and getting management approval..

Part of the problem might be that its open-source software, part of the problem might be that it (if you consider Bugzilla) written in Perl, a language that your managment may not be comfortable with..

I'd suggest you test the waters with Scarab, implemented using servlets.. that might go down better..since you obviously seem to develop using Java (based on the other open source tools you cited)

If you are sure the choice of language isnt the issue, I would HIGHLY recommend Mantis, a simple, VERY useful bug tracker written in PHP.. I use this myself, and its very small and fast to configure (only problem I have with it right now is that it can only use MySQL as a source database)..

Sometimes, the easiest way to get bug tracking software in a large org. is to install it and start using it personally (ie: for bugs related to modules that you develop).. People will gradually come to appreciate exactly how much easier it is on your system as opposed to the other thing, and there will be a creep towards an easier to use system that cant be ignored..

Sometimes, the persons in charge of decisions like this simply dont know how much easier bug tracking software can be than your present software, which is what holds them back :D

Hope this helps and good luck

where have you seen libraries using open-source technologies (like Linux) to solve problems?

UC (University of California) San Diego has started moving towards open-source software. We were a Solaris/Sybase/Netscape shop a few years back. But the cost just isn't practical, especially for places with tight budgets like libraries. So we've been moving (albeit slowly) to commodity hardware, Linux, Apache, Tomcat, etc. for our server-side stuff. Some of the developers use Linux and/or MacOSX for their desktops, too. There are links on my homepage, or respond to this if you're interested in more info.

In terms of software written specifically for libraries, Greenstone and OSS4Lib that other posters mentioned are good. Also check out D-Space, NC State's MyLibrary, and if you want to handle MARC data in Java/XML, MARC4J. Of course, all the standard open-source software works for libraries, too.

-Esme

Subversion does look somewhat better and cleaner than CVS. But there are lots of add-on tools for CVS that will need to get ported (GUIs, servers, web interfaces, IDE integration, etc.). Just the retraining required to get people to use it in a multi-user environment is pretty daunting--CVS is used by many people who are not primarily developers, and the switch wouldn't be easy for them.

Good point, but this is also a big concern of the Subversion folks. This is why subversion looks so much like CVS. The commands and aliases are almost all the same, and a great part of the comportment the users see also is.

The ViewCVS scripts has already been ported to SVN, though it's not perfect yet, it does work. The GUI is pretty much in development indeed : RapidSVN is a working one, yet not complete either. An Emacs mode, similar to the CVS mode, shouldn't bee too hard to code I suppose, this is just a matter of time, will, and knowledge of elisp :)

There was talkings about using SVN as a backend for a wiki too, this could be fun and really nice. A first draft had been coded by Greg Stein (if I'm not mistaken), but it was mostly test stuff.

Subversion still needs help and contributors. People keep whining about CVS not handling file renaming etc, and they also keep using complicated tricks to avoid those flaws. I know, I've done it too. The very same people look at subversion and say "mh, nice, but not mature yet, let's wait it grows up a little". I doubt it'll grow quickly on its own, it just needs some help from all these coders who *will* use it in a few months/years !

Believe me, once you've switched to svn, it just looks life is *so* easier. Try it, it won't bite, and you'll most likely love it !

Yes, try reading the FAQ.

http://subversion.tigris.org/project_faq.html#cv s2 svn.

So far, its "only" a Python script. Very much in beta, the usual warnings apply, but they claim it's working ok.

As an example, the whole first
year of Subversion's own history was converted from CVS into a
3000+ revision svn repository. It took about 30 minutes.

http://svn.collab.net/repos/svn/trunk/tools/cvs2 sv n/README

If you're interested in clients, check out the http://scm.tigris.org/ pages: There's a GTK and a cross-platform GUI in the works (And I've seen a few other GUIs on elsewhere (sourceforge?)

the free software world is severely lacking in UML diagramming tools

What, ArgoUML's BSD license isn't free enough for you?

Hmm. Have you tried ArgoUML?

Other than that, do you really find UML diagrams useful beyond sketches on a white board?

I don't know, I kind of like ArgoUML myself and think it has the potential to be a decent open source entry into the field. It's functional and can produce good work today.

Are there any other decent entries out there?

What about ArgoUML?

Well, ArgoUML can even do round-tripping, and it's Free. http://argouml.tigris.org/

do we need UML to design free software?

I guess the lack of UML-documented free software projects is not due to it not being neccessary, but to the lack of free UML tools. ArgoUML is quite a cool project, but unfortunatly it's a major PITA to actually draw models with it. Of course, there are several other UML tools at sourceforge, but, as usual, most of them are still unusable, and will always be.

There are many distributions and channels for open source projects to reach the end user, so how do users, especially non-technical ones, effectively submit bug reports to the right database? How do open source projects make it easier for users to submit bug reports and consolidate the bugs in a single database?

They all standardise on Bugzilla, and use Bugzilla's import and export (or move) features to move bugs between instances :-)

Other bug trackers (e.g. Scarab) also support import of Bugzilla's XML format for bugs.

Gerv

I should learn to preview. That's http://www.webdav.org

The scenario gets even better as Subversion moves forward.

Well, CVS is fine for small projects. Linux is anything BUT a small project;

KDE, GNOME and XFree86 are not precisely small projects. They all use CVS.

there's a lot of things that cvs either doesn't support, or supports poorly. Binary files and renaming files, for example

Well, subversion will support those two, and also be almost compatible with CVS (except when there is a strong reason not to). A compromise solution would be: use CVS and then switch to Subversion. They are going golden early next year it seems.

I for one plan to resist this bogus, anticompetitive license. As others have mentioned, this is like MS changing their EULA to exclude developers of competing operating systems. The best way to fight BK is to write a compelling replacement. My best wishes go out to those who are already doing such admirable work!

Cheers,
Fyodor

Concerned about your network security? Try the Free Nmap Security Scanner.

Maybe post-1.0 they'll offer features that would bring it up to the level of bitkeeper, but right now, that isn't their main goal.

Is Subversion stable enough for me to use for my own projects?

We believe that Subversion is stable and have confidence in our code, in fact, we've been self-hosting since September of 2001--eating our own caviar so to speak. We declared alpha because we're ready for the world to try Subversion.

Hope this helps

Yes, and there are others points :

- CVS cannot move files and keep a track of the log.
- CVS directory handling is quite horrible

Now, I use CVS as everybody else here does, it works, sure. But there are some problems that should be fixed (and cannot be because of the CVS base), that's why I'll probably switch to Subversion soon. It's still under heavy development, but it gets better from day to day.

Let's support some free software instead of proprietary ones !

I'm curious -- what type of maturity problems are holding you back? Is it the lack of ported third-party modules?

I'm using it at work for either serving static + cgi content, or for svn.