Am 28.02.2011 um 20:41 schrieb Alexander Strange: > The deadline for applying to Google Summer of Code as an organization is > March 11. Has anyone thought of applying for Fink? > > I can think of several projects that could make package maintenance easier > and more reliable, but are missing manpower at the moment; e.g.: > > - automatic BuildDepends > - updating dpkg and co. > - fixing "Could not resolve inconsistent dependencies" > > We would still need people to be project mentors; in my experience GSOC > projects tend to fail if you don't stay in close contact with the students.
I think it would be great if Fink participated in GSoC, and I think there is an abundance of possible projects for students; the main problem, IMHO, will be to find people who have the time and knowledge required to act as mentor. I was myself project org and mentor for ScummVM in 2007-2010, and while it's a great fun & rewarding endeavor, it also requires quite some work and dedication from all involved parties. IMHO. As to projects: Here are some ideas that might serve as basis for full tasks. Some might be merged together into a "improve many things in Fink" task. Whatever :). For ScummVM, collecting all tasks on a Wiki page was very helpful in the past, see here: <http://wiki.scummvm.org/index.php/OpenTasks> * Look through our Feature request tracker: <https://sourceforge.net/tracker/?group_id=17203&atid=367203> Some selected examples: ** Detect firewall settings directly from the system prefs; if ProxyFTP / ProxyHTTP is set to "auto", then the system prefs are used. See e.g. here <https://sourceforge.net/tracker/?func=detail&aid=788801&group_id=17203&atid=367203> ** check whether a package can be safely moved to stable: <https://sourceforge.net/tracker/?func=detail&aid=1667947&group_id=17203&atid=367203> ** Use sandboxing for builds: <https://sourceforge.net/tracker/?func=detail&aid=3195416&group_id=17203&atid=367203> * Improve the PDB further. Many things spring to mind, I believe. We also have some old FRs on this: <https://sourceforge.net/tracker/?func=detail&aid=544279&group_id=17203&atid=367203> <https://sourceforge.net/tracker/?func=detail&aid=721482&group_id=17203&atid=367203> * Improve the .info format further E.g. it would be nice if one could annotate the dependency fields with comments. This would allow adding an explanation to every dependency, where it comes from. That should be quite easy to code, actually. * Add launchd support? Alternatively, improve deamonic to support launchd (I don't know what our status in this regard is, actually) * Port Fink to use aptitude ? (might be too big, and maybe not what we want) Or more generally, catch up with some of the improvements made to the Debian tools in the past decade, e.g. - the ability to use "~" in versions (requires new dpkg and apt-get, the Perl code require is trivial and done) - tracking why a package was installed: Explicitly by the user, or as a dependency; this affects auto-removal of packages which are no longer needed, and affects how dependencies are affected when upgrading a package - ... tons more ... * Often one wants to install package "A", and is prompted to install a ton of other packages, which are only indirect dependencies of the original package. AT least for package maintainers, it would be very helpful to quickly track down where all those indirect dependencies come from. E.g. I just wanted to install "gmime24", and it asked me to install several dozen, sometimes big, packages. Since gmime24 itself is a relatively small and trivial thing, this confused me. In this case, I quickly discovered that mono / gtk-sharp were the culprits (they are deps of gmime24, used to build some C# language bindings. Coincidentally, I think it would be nice if the gmime24 package was split, moving those language bindings into a separate package, but I digress ;) One would have to determine what exaxtly the command should do, how the output might look, etc.; part of that would be our job (creating the task description), part would be the student's (helped by the mentor(s)). * Oftentimes, students have surprising ideas of their own... so be open :) As I see it, ideas for improving Fink, even in student suitable chunks, is not the problem; the problem is describing these well, and then mentoring the student. Cheers, Max ------------------------------------------------------------------------------ Free Software Download: Index, Search & Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev _______________________________________________ Fink-devel mailing list [email protected] http://news.gmane.org/gmane.os.apple.fink.devel Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-devel
