Comment 21 for bug 1164016

Revision history for this message
Adam Dingle (adam-yorba) wrote :

Sebastien, thanks for your comment - it's nice to know that someone at Ubuntu is paying attention to this thread.

I don't think forking an old version of Nautilus is practical. The problem is that it would be a ton of work to keep such a fork compatible with new versions of GTK, GLib and other libraries. Instead I think the current version of Nautilus (3.8 in Saucy) should be patched to fix this.

Removing type-ahead find has been controversial upstream (i.e. at GNOME) - see, for example, the discussions at

https://bugzilla.gnome.org/show_bug.cgi?id=680118
https://bugzilla.gnome.org/show_bug.cgi?id=699087

Nevertheless I think it's very unlikely that the GNOME developers will fix this for 3.10, or possibly ever. Even if they were to fix this in, say, Nautilus 3.12 next spring, that version won't appear in Ubuntu until late 2014 (since Ubuntu lags GNOME by a release cycle). So if this going to be fixed in Ubuntu in the near future, I think Ubuntu will need to take responsibility for doing that. You may not like that - sure, in an ideal world this would be discussed rationally and fixed upstream - but this is how it is.

You wrote "not sure that's a problem that can easily by fixed by Ubuntu". I guess that depends on what you mean by "easily". I'm quite confident that a competent hacker could come up with a patch to the current Nautilus that restores type-ahead find in, say, 2-3 weeks. It might even be easier. I'd start by attempting to git revert the commit where this breakage was introduced:

https://git.gnome.org/browse/nautilus/commit/?id=469eb89117e199a450aec3411183ed2e9f10f893

That will yield merge conflicts, which must be studied and resolved in some nice way. This might not be solvable in a day, but I'd be very surprised if this took more than a month of work.

To those of us on the outside, it's baffling that Ubuntu would not want to invest that amount of effort to solve this, since this is a highly visible core program on the Ubuntu desktop and it seems pretty evident that the majority of users are harmed by this. Has someone at Ubuntu already spent several weeks working on this and concluded it was too difficult? Or does Ubuntu not feel it would be worth that amount of effort?