Galeon Developer Interview, July 2003
Due to his own curiosity and the apparent curiosity (or vocal ignorance) of folks around the net, Topher The Web Guy asked some of the Galeon developers a few questions. If your curiosity is not satiated, there'll be a form at the bottom to ask your own questions.
| ric || Ricardo Fernández Pascual |
| yaneti || Yanko Kaneti |
| philipl || Philip Langdale |
| tko || Tommi Komulainen |
How "healthy" is the galeon project?
Stable but serious.
We're operating at a fairly low level right now but we've gained a lot more focus recently and Crispin has been a great help. Having the whole gnome farce behind us is also a relief. ximian deciding to ship galeon as their primary browser is a big boost, especially given redhat has thrown us out (no surprises) and slackware too sniff. We still have problems dealing with the bad image we have of 1.3 as a featureless POS; most people don't realise how far we've come since October 2002. But we're getting there, slowly but steadily. Galeon isn't going away.
Pretty healthy all things considered. Not being "the official GNOME browser". Excluded from Red Hat rawhide. Dissed for all the wrong reasons by uninformed people. - Yet people still seem to be interested and most importantly "external" patches seem to have picked up recently, which is just great. Many thanks.
We're progressing nicely, if a bit slowly. We've come a long way after hitting rock-bottom and more people are starting to realize that, so the bad image we got last year should be a thing in the past.
In an interesting twist of events, the Epiphany team has started developing an extension system, and the first extension is mouse gestures - the sort of feature the whole disagreement that eventually lead to Epiphany's birth was all about. After a fashion they're sharing our view on the 'advanced' features after all, and I'm guessing by GNOME 2.6 it'll get where we are now. I only wonder why we couldn't start that last October when I was suggesting it, and skip all the fla^Wfriendly discussions. Oh well...
It is surprisingly healthy if you think that it has not had a stable release in a lot of time and that it no longer considered "the" gnome browser. I think that there is an important niche for galeon as the usable and useful gnome browser.
 even if the released versions are only development versions and not officially stable, they are quite stable actually. I use CVS builds always and it does not crash easily. discussions. Oh well...
How much work is left before Galeon 1.3 becomes Galeon 2.0?
I've just updated the TODO list.
Mainly, we want to:
Unfortunately this means we do have a big architectural step left which is exorcising bonoboui and switching to the egg library. The other stuff is comparatively straightforward.
- Dump the albatross called bonoboui
- Reorganise the prefs dialog to actually be useful.
- Make the bookmark editor not suck big fat rocks
- get the stylesheet chooser back. that is really missed.
- polish! polish! polish!
Thats a kind of managerial question which would be best left to philip. I count whats left by looking at the number of bugs on the 2.0 milestone, which might be entirely bogus because its only me who puts them there
A fair amount. While we have a good set of features and a few more coming, they're not presented well to users. For example, mouse gestures have been implemented since last September, but still people keep asking about them. We need to go through the preferences and decide which ones to show in the dialog, which ones to keep semi-hidden in GConf only, and which ones to remove.
Another large task to do is to update the documentation, although I'm not sure anyone reads it anyway I hope we can get someone who's both willing and capable of writing good documentation.
And we need a lot of polish. There are little things all around that are perfectly usable, but with small modifications could be even better.
There are also some technical changes we need to do, most notably getting rid of the pain called bonoboui. While quite nice for static UI, it's painful for dynamic menus and toolbars.
Galeon 2.0 should be released once we have feature (and stability) parity with 1.2.x. I don't miss any of the missing features from 1.2.x now, but there are some users still missing small things.
It's hard to give a time estimation, but I expect a 2.0 release before the end of the summer.
Are more developers starting to come on board?
Well, one at least. Crispin Flowerday has been doing a brilliant job; he's highly motivated, like we used to be and brings a fresh mind to old problems. We also have a volunteer for the stylesheet menu, but tko will be in a better position to talk about that.
Yes, Crispin Flowerday has been wonderful in fixing and adding and patching up things we've overlooked or not bothered doing ourselves yet. If he's the sample, we'll take a dozen
Robert Clark is working on the links/related documents toolbar button, and he also showed interest in restoring the stylesheet selector in the menu. Unfortunately I'm slow at reviewing the patches so it takes time.
Besides work, life, and other interests there's only so much time left for Galeon; it would be a big boost if we were actually paid for doing this. One can always dream, right?
Yes, and it's great to see new contributors submitting patches to the list.
Most times, the easiest way to get something done is to submit a patch. Even if it is wrong it does motivate a lot to fix the problem.
What are your top 3 bugs that need fixing?
- Stupid fscking inability of the toolbar to respond to pref changes correctly.
- Printing crashes (if they're true, not yet properly confirmed)
- Can't escape out of typeaheadfind (this seems fairly minor but it really gets me because it reflects the suckiness of bonoboui. bonoboui eats the shortcut (escape) and refuse to let anyone see it.
Just one I can think of atm: Printing seems to be crashing in some cases, not even sure its a galeon problem but it just comes up a lot - see bug 114869.
We have a pretty big amount of open bug reports many of which UNCONFIRMED and many others just not relevant to the 1.3 codebase anymore. The former are certainly partly my fault, sometimes I just can't catch everything thats coming, or I fail to sort it in the right place. More eyes on these would be highly welcome. The latter perhaps just can be closed someday but I can't really decide when it would be good time to do that.
Although, I'm not sure if it's in our power to fix these :-/
- inactive tabs steal keyboard focus when done loading
- cookie confirmation dialog occasionally ends up in wrong workspace
- a dialog popping up while performing gestures blocks X
I have lots of bugs in my bugzilla folder that should be fixed. I can't think of a top 3... the problem is that I have got used to the bugs so I ignore them when using galeon.
What are the top three features you want added?
You'll notice these are 1.2.x features. I haven't had a chance to think beyond getting back to where we started.
- Feature complete Tab menu (restoring move-to menu items)
- depends on ripping out bonoboui. Dynamic menu generation in bonoboui is a horrible messy bloated process that I refuse to touch.
- Stylesheet menu.
- Page info dialog.
Can't think of anything major. Many small additions bringing it to the 1.2.x usefulness level would be nice. Most of these are tied to the menu and toolbar code which unfortunately seems to be hard to deal with currently. A port to a better system , perhaps the new egg/gtk code would be very welcome.
Sharing bookmarks, cookies, history, passwords, and preferences with all and any browsers that might be installed on the system; at the moment if I want to try another browser, say epiphany, I have to reconfigure all settings from scratch and import or export bookmarks. You can basically forget about cookies and history. After browsing for a while I'd have to re-import the bookmarks back to Galeon, assuming the other browser didn't completely destroy the bookmarks structure or some metadata (smart bookmarks, for example.) Next day, try Konqueror and do everything all over again.
I'm painfully aware the above is not exclusive to browsers (I've lost count how many times I've needed to configure mail folders somewhere.) As a user I see no real reason why I'd need configure something more than once, unless I'm overriding some specific setting.
It's a huge task and progress is slowly being made in that direction. Unfortunately I, like probably everyone else, find it more satisfying to work on something where I can see the results immediately.
Oh, you meant in this lifetime? Well, let's see...
- icons in notebook tabs
- related documents navigation
- smart bookmarks in the context menu
- Getting rid of libbonoboui. I hate libbonoboui. Using it has caused a lot of harm to galeon. I'm glad that gnome is getting rid of it now using EggToolbar?, but it's a shame that we had to write a lot of code to workaround that horrible api. Unfortunately, I think that we are not going to be able to do this before 2.0.
- Drag and drop and context menus everywhere, like in galeon 1. This is dependent from the previous one.
- Saving backward/forward history in session.