Gentoo Archives: gentoo-amd64

From: Duncan <1i5t5.duncan@×××.net>
To: gentoo-amd64@l.g.o
Subject: [gentoo-amd64] Re: Heads-up: KDEers: Particularly kde3-ers,
Date: Thu, 27 Aug 2009 07:51:03
Message-Id: pan.2009.08.27.12.59.53@cox.net
In Reply to: Re: [gentoo-amd64] Re: Heads-up: KDEers: Particularly kde3-ers, by Volker Armin Hemmann
1 Volker Armin Hemmann posted on Tue, 25 Aug 2009 01:33:23 +0200 as
2 excerpted:
3
4 > On Montag 24 August 2009, Sebastian Beßler wrote:
5
6 >> 3) Screen-setup: I have to monitores connected to my pc static in xorg
7 >> as two seperated displays :0.0 and :0.1. If I start kde <4.4-svn I have
8 >> both screens on my main monitor and the second is just as good as dead.
9 >> There is a patch for this in svn (or is it git now?) and as much as I
10 >> can say by testing a few live-builds this point will be gone by next
11 >> release.
12 >
13 > ok, sounds like a genuine stupid bug. Can't even fixed with krandrtray?
14
15 This one is a known kde4 issue, and earlier, they weren't planning to fix
16 kde4 to do multiple X sessions like kde3 did. From what I've read it was
17 purely an accident that kde3 handled multiple sessions that way.
18
19 It's possible that may change, given the requests they've gotten, but I
20 don't expect the devs to prioritize it, which practically speaking, means
21 they'll take patches...
22
23 Meanwhile, kde4.3.0 still has issues with xinerama (what SB calls "big
24 screen") mode in some cases (including mine), too. It detects two
25 monitors, but only sees one CRTC, apparently, and thus forces them into
26 clone mode. The systemsettings display applet preview shows just one
27 screen, with the label for the second monitor overwriting that of the
28 first, so the result is an almost unreadable combination of the two
29 labels. And the settings for each monitor are all there except for the
30 position control that would allow you to set on-top-off, to-the-right-of,
31 etc. (I didn't even know what it was supposed to look like until someone
32 else with the problem posted a screenshot to one of the kde lists I'm
33 following, and another poster followed up with a screenshot of how it /
34 should/ look.)
35
36 But I've seen several mentions that the problem has been fixed in trunk,
37 for 4.4, and the fix may or may not make it into 4.3.x.
38
39 Meanwhile, since kde3's display applet never worked right either for me,
40 I've developed my own scripts to manage resolution changes here, calling
41 xrander to do it. As long as I don't open the kde display settings
42 applet or run krandrtray and screw with it, I'm fine, but if I even so
43 much as open the display settings applet, kde screws things up royally,
44 and just running the xrander script doesn't fix that. I have to screw
45 around with plasma to get it back the way it's supposed to be, as well.
46
47 So for now, I just don't go there any more, and I'm fine. But I'm
48 looking forward to 4.4.0 or 4.3.x when that fix gets into the release,
49 hoping I can actually manage displays using kde then, for the first time
50 ever! =:^)
51
52 >> 4) Design: I have a very small design with two small bars at the bottom
53 >> and on the right side. I can't create a design in kde4.x so far that
54 >> consumes as few space on my monitor as i have it in kde 3.
55 >
56 > but you know that you can make the plasma bar very small - and add a
57 > couple of them to the desktop?
58
59 4.3 might actually be able to do this now. Each 4.X release has VASTLY
60 improved plasma, so far, and in this area 4.2 was finally becoming semi-
61 usable and 4.3 is actually very reasonable, now.
62
63 If small is what you want, 4.3 really does allow small now. I have one
64 panel set as small as it'll allow, looks like 12 px high (horizonal
65 panel). IIRC the minimum in kicker was 24 px?
66
67 But I still find the panel settings widget clumsy and inefficient,
68 compared to how kicker worked.
69
70 And while the panel settings widget at least works, the entirely
71 automated panel plasmoid sizing is still simply broken in some regards.
72 There really needs to be a way to force a plasmoid to only take up so
73 much space in the direction of the orientation of the panel (width on a
74 horizontally oriented panel, height on a vertically oriented one). By
75 that I mean, have /plasma/ force it, regardless of what ideas the
76 plasmoid has about its size. Some of the plasmoids simply hog space, and
77 can't be forced smaller, without changing the size of the panel itself
78 (height on a horizontally oriented panel, width on a vertical oriented
79 panel).
80
81 If plasma could force it, the plasmoids would have to adapt. If they
82 didn't, they'd probably simply die out from disuse, as people found
83 solutions that actually worked.
84
85 I was hoping the spacers introduced with 4.3.0 would correct the problem,
86 and they would have if they could be arbitrarily placed, thereby limiting
87 the area the plasmoids on either side had to work with, but the hogging
88 plasmoid simply shoves over the spacer.
89
90 One way plasma could implement this while staying compatible with older
91 plasmoids, would be to lie to them about the panel size, if it had to,
92 thereby forcing them to adapt to what they believed was a smaller sized
93 panel.
94
95 Another problem still existing in 4.3.0, is that if the install new
96 widgets functionality is used (so the new widget appears in the list
97 available to add), there's no GUI method to remove what has been
98 installed (so it no longer appears in that list). One must actually find
99 the location in kde's configuration tree on-disk, and remove the offender
100 there. If it can be installed via GUI, it should be uninstallable via
101 GUI as well.
102
103 And one of my big irritations is that there's no ksysguard plasmoid to
104 replace the ksysguard kicker applet. It's bugged. IDR whether I filed
105 the bug or just CCed an existing bug, adding my own request, but either
106 way, there have been several folks add comments asking for it as well,
107 since I did whichever.
108
109 Of course, ksysguard4 still has some serious bugs of its own, including
110 the fact that it doesn't properly restore user set minimum and maximum
111 settings on the fancy-plotters, if the user has turned off auto-ranging.
112 That too is bugged.
113
114 Meanwhile, I've been working on setting up a superkaramba scheme that
115 does what I want, but that's still on my list, I've not been able to do
116 it yet. (The other last big issue I had was khotkeys multi-key, but I
117 solved that with my own script, as I described in a previous post.)
118 Superkaramba is much more flexible than ksysguard in layout, etc, so it
119 should be great when I get it setup, but it's a matter of actually having
120 the time to do it.
121
122 Until then, I have to reset a half-dozen plotters to their proper ranges
123 every time I restart kde or just ksysguard. That sort of repetitive work
124 is /exactly/ the sort of thing computers are supposed to be able to
125 handle for us humans, so it's rather ironic that ksysguard is making me
126 do it, instead of allowing me to tell it how I want it set and have it do
127 it.
128
129 >> The fifth point on my list is that many of the functions and (3rd
130 >> party) kde programms I use day by day aren't there yet at all or only
131 >> with lack of functions and mostly in late alpha-state: k3b, amarok,
132 >> kdevelop, kvirc to just name a few.
133 >
134 > I haven't found anything missing in k3b or amarok - and I don't use
135 > kdevelop or kvirc. What are you missing from k3b?
136
137 k3b I use in general, but haven't tried the kde4 version yet. (I have it
138 merged, and ran it briefly to see how it looked, tho, and nothing lept
139 out at me as lacking.)
140
141 But I finally got so fed up with amarok I found a different solution.
142 They killed all the functionality I actually used, while adding a bunch
143 of junk I'm not interested in.
144
145 Have they gotten the winamp/xmms skinnable mini-player back yet? Last
146 time I checked, they weren't interested in the idea. What about
147 visualizations? That actually may be back in the kde4 version now, I
148 don't know, but I had a lot more use for that then all the fancy main UI
149 changes they did. And I never /did/ use all that fancy scoring and etc.
150 functionality.
151
152 It's fine if they don't do it. It's just that our paths have obviously
153 separated (not that they were ever really the same, but it was convenient
154 to travel with them for a time). Like the last ISP I left, they're free
155 to develop the product as they wish, but I'm simply no longer interested
156 in going where they were headed. Thus, they can go their way and I'll go
157 mine.
158
159 Still, the thing that really had me fed up was somewhat different.
160 That's the way they handled the transition to the MySQL backend. Not
161 only is that entirely unnecessary for the features I'm interested in (tho
162 I'd have tolerated it if they had kept the features I actually used), but
163 they demonstrated a COMPETE disregard for their amd64 users when they
164 added it, as it was ENTIRELY broken on this arch at the time. The
165 library they were using wasn't designed for dynamic use, only static, and
166 thus simply wouldn't function as a dynamic library on amd64 (at least not
167 without affecting the performance of all the rest of the package,
168 including the mysql app itself), as dynamic libs here require -FPIC, and
169 at the time the only way to get that was to compile the entire package
170 with it.
171
172 That was the last straw for me. Not only were they dropping all the
173 useful stuff (from my perspective) from their kde4 version, but
174 demonstrating a total disregard for their amd64 users as they did, that
175 was more than I was willing to take, and I decided there were simply
176 better options available.
177
178 >> The last reason I stick with kde 3.5.10 for a while is that working
179 >> with kde 4.x just doesn't feel right. Switching to kde4 is like
180 >> switching to a completly different DE. KDE 4 isn't kde anymore, it is
181 >> something absolutly different that calls itself kde.
182 >
183 > people said the same when going from 1.1 to 2.0 and 2.2 to 3.0....
184
185 Indeed. I expected some adjustment time, and that's why I was running
186 the live version before 4.0 came out, with the idea that by the time it
187 was ready, I'd be ready too.
188
189 But I had no idea it would be 4.2 before it even got to the beta state
190 this early-adopter likes to jump in at, 4.3 before it got to rc, and
191 apparently 4.4 before it gets to actual normal X.0 release quality!
192
193 And /as/ such an early adopter, I had no idea I'd be feeling the clock
194 ticking on the end of support and removal from my distribution on the
195 actually generally usable version, before I could even do the usual beta
196 testing I normally do with such important products.
197
198 > The only two things that I really disliked about kde 4.X is the new
199 > 'systemsettings' - I liked kcontrol. Very much. And akonadi creeping
200 > into everything.
201
202 akonadi is frustrating, for sure. As is nepomuk, soprano, etc, at least
203 here.
204
205 And I agree on kcontrol vs. system-settings, as well. FWIW, you can
206 change the system-setting menu entry (using kmenuedit), renaming it
207 kcontrol, if you want, and you can use the qt --caption or the kde --
208 title option if you wish. But the title changes back to System Settings
209 as soon as you activate any of the applets.
210
211 Of course I'm using the tree view option they added back to it in 4.3,
212 that was the way kcontrol3.
213
214 But as with kde3, I put the kcontrol/systemsettings menu widget on a
215 panel (complete with its own keyboard shortcut, even), and seldom invoke
216 the full kcontrol/systemsettings any more. There's really no reason to,
217 once you figure out what applet a particular setting is in, and you have
218 your system setup in general the way you want it, so you're only doing a
219 single change or two in a particular place, here or there. But as with
220 kde3, I still access the settings enough that the menu's worth having.
221
222 --
223 Duncan - List replies preferred. No HTML msgs.
224 "Every nonfree program has a lord, a master --
225 and if you use the program, he is your master." Richard Stallman

Replies

Subject Author
Re: [gentoo-amd64] Re: Heads-up: KDEers: Particularly kde3-ers, "Sebastian Beßler" <sebastian@××××××××××××.de>