Gentoo Archives: gentoo-dev

From: Peter Stuge <peter@×××××.se>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] Making systemd more accessible to "normal" users
Date: Wed, 01 May 2013 21:40:52
Message-Id: 20130501214043.10611.qmail@stuge.se
In Reply to: [gentoo-dev] Making systemd more accessible to "normal" users by Fabio Erculiani
1 Fabio, I think you're doing awesome work!
2
3 Steven, I think you can behave a lot better on the internet. kthx.
4
5
6 Steven J. Long wrote:
7 > > It looks like there is some consensus on the effort of making systemd
8 > > more accessible,
9 >
10 > Sure there is: there's also consensus that this approach is wrong for
11 > Gentoo.
12
13 In my world naysayers have no say and doers decide, as long as there
14 are no obvious negative consequences from doing.
15
16 In my world it is also an absolute no-brainer to try to make systemd
17 as accessible as possible in Gentoo.
18
19
20 > Switching init isn't done that often
21
22 That doesn't mean that it couldn't be, and it also doesn't mean that
23 it wouldn't be handy to use eselect to do so.
24
25
26 > and when it is a Gentoo user is expected to deal with configuration.
27
28 I don't know where such expectation could come from since, as you
29 write, switching init isn't done that often; so there can't be a lot
30 of user feedback from doing it, and it hasn't been discussed very
31 much by developers.
32
33 If you mean that *you* expect users to deal with configuration then
34 that's fine and valid, but I think that if we can find a neat way for
35 users *not* to have to deal with configuration when they want to
36 switch init then that would be really nice!
37
38
39 > In this case, it's a doddle to set the command-line to
40 > init=/sbin/fubar to try it
41
42 I think it's less about telling the kernel which binary will be
43 process 1 and more about what gets started by process 1 on next boot.
44
45
46 > If they can't handle the above, they shouldn't be on Gentoo imo.
47
48 You have all right to your opinion, but I for one don't share this
49 opinion. If we can make it easy to switch around I think that's
50 awesome.
51
52
53 > I don't see the use-case frankly.
54
55 I would say that it is to make migration easy.
56
57
58 > So I just don't see which Gentoo users this is helping.
59
60 Anyone who has a Gentoo system using OpenRC who would like to try out
61 systemd.
62
63
64 > Making it even more trivial to change init than it already is, is
65 > actually a bad thing in my eyes.
66 > It gives the impression that it can be undertaken lightly which is
67 > simply bad practice.
68
69 Sorry, I don't buy your argument. Consider how lightly I can
70 undertake deletion of all my data, which is also bad practise:
71
72 rm -rf ~
73
74
75 > AFAIK it's been policy for a while that systemd unit files should
76 > be installed by default, for all the reasons you've given. I can't
77 > see a maintainer being bothered by the systemd herd adding them
78 > when they have no interest: after all users can already set an
79 > INSTALL_MASK, and it makes binpkgs more useful.
80
81 Yep, +1 for all of this. I think Fabio shouldn't let unresponsiveness
82 of others create very long wait states when doing benign changes.
83
84
85 > > The final outcome will hopefully be:
86 > > - easier to migrate from/to systemd, at runtime, with NO recompilation
87 > > at all (just enable USE=systemd and switch the device manager from
88 > > *udev to systemd -- unless somebody wants to drop the udev part from
89 > > systemd, if at all possible)
90 >
91 > How is adding USE=systemd to a system with it switched off (ie:
92 > enabling it), *not* going to lead to recompilation?
93
94 Setting the USE flag doesn't lead to recompilation per se, but the
95 question is good - what will the USE flag actually mean when we
96 arrive at the final outcome? Will it do anything at all? Fabio?
97
98 In the end, maybe it would just control if baselayout DEPENDs on
99 openrc or systemd?
100
101
102 > > - we make possible to support new init systems in future, and even
103 > > specific init wrappers (bootchart anyone?)
104 >
105 > Which is possible already, so this is null.
106
107 Consider that Fabio and I are not native english speakers. I would
108 recommend spending more time seeking to understand what was intended
109 to be transmitted, rather than merely interpreting the words received.
110
111 Communication becomes significantly more effectively that way, which
112 you probably don't need me to bring up, if you're used to talking to
113 users on forums. :)
114
115
116 > > - we put back some fun into Gentoo
117 >
118 > Eh, I've been having much more fun since ..
119 ..
120 > the latter is only possible because Unix is designed on a modular
121 > basis, and we can still swap components in and out on Linux (for now.)
122
123 You are implicitly communicating that systemd can not be fun because
124 it is not modular. Basic flaming. What's the point of that?
125
126
127 > Please note: I fully support your effort to make it easy to switch
128 > back and forth
129
130 I have no doubts that this was true in your original email, but you
131 should consider that the words you chose communicate something very
132 different.
133
134
135 > I just don't think that adding a fragile eselect module (along with
136 > "this needs investigation" as things come up) is the way to do it.
137
138 I think the point is to investigate exactly to ensure that the module
139 *is not* fragile.
140
141
142 > Nor should someone change init on a whim, without being ready to
143 > handle configuration.
144
145 I think it would be awesome if we can allow changing init on a whim,
146 without having to handle configuration.
147
148
149 > In fact, dumbing down Gentoo is dangerous imo.
150
151 I think you misinterpret the intention. Creating powerful tools to
152 make complex tasks appear simpler doesn't fit my understanding of
153 "dumbing down." (My understanding is to artificially restrict what
154 tasks can be done.)
155
156
157 > Gentoo, like Unix, doesn't stop you from doing stupid things, as
158 > that would stop you doing clever things.
159
160 Switching init can be wise. (You later use "clever" in a derogatory
161 fashion.) I nearly replaced init with supervise on my systems before
162 I started using systemd. supervise is one of a few components I would
163 have written myself if I hadn't already found that someone else had
164 done it.
165
166
167 Steven J. Long wrote:
168 > Ah I see: sorry you're taking my email in the wrong spirit.
169
170 That is quite understandable to me, since you made liberal use of
171 flammable wording next to the explicit, but brief, expression of
172 support of the effort.
173
174
175 > I thought I made it quite clear I'm not hostile to your intentions,
176 > but you appear to be hostile to everything I've written.
177
178 See above. I think you could have communicated your points a lot
179 better, so that they would not have been misunderstood.
180
181
182 > > Have you ever tried to fully migrate to systemd from openrc? Clearly not.
183 >
184 > OFC not, that's the point: it's why I'm asking you.
185
186 I guess you realize that it isn't so useful to first educate peers
187 (answering you) before moving on to actual discussion. If you haven't
188 got experience from the details of this topic, perhaps your time is
189 better spent on another topic..
190
191
192 > > > Again, what about people not using consolekit, nor logind, with
193 > > > no intention of ever installing systemd?
194
195 They might change their mind at some point, and I think it would be
196 cool to make it as easy as possible to switch both ways.
197
198
199 > > > I've got nothing against this so long as it is guaranteed not
200 > > > to break my pam setup. As-is I feel *very* wary of a change
201 > > > that unconditionally requires a 'pam_systemd.so'. Please note I
202 > > > am not hostile to your aims: I am merely seeking reassurance.
203 > >
204 > > Do you know how pam works? And did you understand the meaning of
205 > > my words?
206 >
207 > Again, you're not helping yourself with this attitude. Just a
208 > friendly warning.
209
210 Your words are far from friendly.
211
212 I for one did not understand the meaning of Fabio's words, it would
213 be cool if he can clarify the details about the pam_systemd.so file.
214
215
216 > > Do you know what optional means in this context?
217 >
218 > "Always enable the optional.." means "require the currently
219 > optional.." to me.
220
221 I think this is a misunderstanding, because it doesn't fit with the
222 general intention I receive in Fabio's mail. I can't explain what
223 Fabio meant exactly, I believe I also don't quite understand what
224 he meant. I hope he'll clarify a bit.
225
226
227 > I still don't see why you think it's a miracle openrc works with lvm,
228
229 I think it's valid to ask for more details about potential problems
230 with openrc+lvm, although such details are also not really on topic
231 for this thread. (Very good to do in another thread however, maybe
232 there is also some misunderstanding about how openrc+lvm is supposed
233 to work, which would allow a smoother user experience and perhaps
234 improved documentation.)
235
236
237 > unless you mean it was an effort for you.
238
239 I don't see the point of this ad hominem.
240
241
242 > > Please, write about something you actually manage to _know_.
243 > > I want to discuss about improving the systemd experience.
244 >
245 > Hmm.. no.
246
247 What no? No you will not write about matters where you have
248 experience, or no you do not agree that this discussion is
249 about improving the systemd experience?
250
251
252 > I'm afraid you haven't shown that Gentoo users don't currently have
253 > a choice of init systems: so you're not some liberator endowing us
254 > with "rights" we didn't otherwise enjoy til you came along with
255 > your magic impl, I'm afraid.
256
257 I think you're behaving like an asshole, I don't see the point of that.
258
259 When studying systemd it becomes clear that there are potentially
260 interesting challenges in migration between process 1
261 implementations, and experience quickly confirms it. :)
262
263 I don't think Fabio has claimed to endow you with rights you didn't
264 have before, or enable new choice. I think he intends to make it
265 *easier* to effect one's choice. He points out several things which
266 would help accomplish this.
267
268
269 > As for this topic being solely about improving the systemd
270 > experience, that's a change. I could have sworn i read something
271 > about "improving the love between openrc and systemd" and making
272 > *both* work better. But since you're now stating this is just about
273 > systemd, I'll just point out that you're awfully territorial
274 > yourself.
275
276 I think "focused" is a better word. Since systemd is a new
277 alternative, and since it works differently in various ways, other
278 parts need to change to fit together with it. I think everyone agrees
279 that such changes should not have negative effects on already well-
280 established openrc.
281
282
283 > And your attitude of ignoring openrc people does not increase the
284 > love at all.
285
286 I for one don't see that attitude from Fabio. Can you be specific
287 about where openrc people (do you mean developers, users, or both?)
288 were ignored?
289
290
291 > > The topic here is to improve the systemd experience, if you are
292 > > an openrc user that doesn't care about systemd and other stuff,
293 > > you are off topic.
294 >
295 > No the above is,
296
297 Do you intend to say that Fabio is being off-topic in a thread he
298 himself started just two emails earlier?
299
300
301 > There is no consensus as you claim.
302
303 I think there is, it's just more local than I think you interpreted
304 Fabio to mean.
305
306
307 > > We should stop thinking about Gentoo like a guru-distro. Gentoo is
308 > > about choice, but choice != complexity. Making things easier is not
309 > > against our Manifesto.
310
311 Fabio makes excellent points here.
312
313
314 > The thing you're ignoring is that your setup is more complex,
315
316 Sorry, what do you mean? What setup is more complex than what alternative?
317
318
319 > and you clearly don't give a damn about, and have not considered,
320 > the effect on other downstreams.
321
322 That's not at all clear to me. What are some concrete negative
323 effects of Fabio's suggestions on "other downstreams," and which
324 downstreams do you have in mind?
325
326
327 > So we get more complexity and less choice overall,
328
329 I don't follow you. Please clarify? Tooling that simplifies switching
330 might end up complex, but only if the task itself actually requires
331 complexity. I don't understand the "less choice overall" part at all. :\
332
333
334 > as is usual with idiot-box approaches.
335
336 Another ad hominem, wheter against Fabio or Lennart this isn't a very
337 helpful comment in the discussion. It's clear that you aren't very
338 interested in making systemd work (easily) on Gentoo..
339
340
341 > And sorry, but a distro that doesn't hold your hand is a lot
342 > _easier_ to work with in the longer-term.
343
344 ..from this comment and others. You could have saved us all a lot of
345 time if you had simply written a brief email saying something like
346
347 "I disagree with the goal of making it easier to switch from and to systemd."
348
349 along with some to-the-point qualification.
350
351
352 > I give up trying to be polite in the face of such crap, it's more
353 > than I can stomach.
354
355 If you can't compose yourself in the face of someone who doesn't seem
356 to understand you then please think twice before entering into
357 discussions. Misunderstandings are frequent on the internets.
358
359
360 > > Implementing new stuff also means making things easier, especially
361 > > in the systemd case.
362 >
363 > LMAO. You go girl, strut that nonsense like it means something.
364
365 Something is obviously meant, but you didn't receive it. I also
366 haven't received it. I don't know exactly what "new stuff" Fabio
367 refers to, but I can certainly think of things that he might have
368 intended to communicate, which allows his sentence above to have
369 wise meaning. Please try to think of such things you too.
370
371
372 > > >> while there are problems with submitting bugs about new
373 > > >> systemd units of the sort that maintainers just_dont_answer(tm).
374 > > >> In this case, I am just giving 3 weeks grace period
375 > > >
376 > > > AFAIK it's been policy
377 > >
378 > > Thanks for reminding me a policy I am supposed to already know about.
379 >
380 > So why are you complaining about maintainers who are not interested
381 > in systemd, who ignore your bugs and don't add the unit files you
382 > want them to?
383 >
384 > Maybe they know the policy too.
385
386 Fabio is being polite to give a grace period and it would be polite
387 of maintainers to answer, even if only to point out that they are
388 fine with him making the changes immediately.
389
390 It would be polite of them because it would remove a lot of wait
391 states. If there would be critical mass then at some point no new
392 wait states would be created. It is quite clear from my rather brief
393 experience with Gentoo developers that no matter what policy you have
394 to back you up you can make someone upset enough to flame you by
395 doing something that they don't like.
396
397 The wait states introduced by Fabio giving a grace period is a
398 typical example of the chilling effect which is a quite natural
399 result from such attitude.
400
401 It is what it is. Lots of software developers simply suck at dealing
402 with other people, and unfortunately this affects the software we all
403 work on, because most significant (open source) software development
404 is too much for any one person. Sad, but a fact.
405
406
407 > > Please take more time reading about what's in my overlay before
408 > > jumping the gun.
409 >
410 > No way, sunshine. If you make what is effectively a marketing claim
411 > like "no recompilation" then don't add the qualifications later on.
412 > Be precise upfront, instead of typing so much noise. Or at very
413 > least be polite when someone queries it.
414
415 Your query was not particularly polite either. I think it is
416 reasonable to ask for you to review Fabio's work before rejecting it.
417
418
419 > If you post to a wider mailing-list like this, you should bear in
420 > mind that the audience is not simply Gentoo developers, by _design_.
421 > If you don't like that, too bad.
422
423 Do you mean to say that because someone receives an email they don't
424 have a moral responsibility to consider if a reply will contribute
425 something to the topic, and a duty to optimize their reply for
426 efficiency for the sake of readers? I disagree with that.
427
428
429 > Further, if you're posting to get feedback and buy-in from other
430 > people, you severely limit yourself when you suddenly state that
431 > only those who have already done the openrc -> systemd migration
432 > are qualified to discuss it.
433
434 Maybe you agree that it's a lot less useful to discuss solutions
435 with someone who hasn't experienced the problem? It's not impossible,
436 in particular it's already very useful to discuss based on experience
437 from using openrc and systemd independently, since that already
438 forces thinking about the problem in concrete terms.
439
440 It's possible to come to same conclusions through theory, but that
441 usually takes significantly more effort. Our time as contributors is
442 scarce, so we tend to prefer conclusions drawn efficiently.
443
444
445 > Doubly so when you're rude to someone who actually felt quite
446 > supportive of your effort, if not the design.
447
448 I think Fabio reacted quite composed to your words, which were
449 indistinguishable from verbal attacks in spite of your explicit
450 expression of support.
451
452
453 > Believe me, I don't now. I just think you're a loud-mouthed amateur
454
455 You're writing what you think about Fabio, when the topic is making
456 it easier to switch process 1 implementations in Gentoo. Please focus.
457
458
459 > I don't ever want to interact with you again.
460
461 This is behaving like an asshole, which is harmful not only for Fabio
462 (which I guess you intend, but which I certainly don't find good) but
463 also for this mailing list and indeed the Gentoo project as a whole.
464
465 Please behave better than that.
466
467
468 > And interacting with you is not fun at all.
469
470 Volunteer contributor collaboration across nations, cultures and
471 languages using seven bit text is rarely fun - most of the time it's
472 damned hard work and requires boatloads of patience, to arrive at
473 even halfway good things.
474
475
476 > Your designs sucks afaic
477
478 I don't know about that. I think it sounds pretty good from a
479 usability perspective, and the way I understand Fabio's intentions
480 it also seems to make sense from a technical perspective.
481
482
483 > Just so long as I can keep hard-masking
484
485 What to mask is always your choice.
486
487 > your rubbish
488
489 I don't know about that, since again I didn't review Fabio's work.
490 But again, from what he describes it doesn't sound like rubbish to me.
491
492
493 > I'll .. switch distros
494
495 This statement creates a pretty bad atmosphere for little reason.
496
497
498 //Peter

Replies

Subject Author
Re: [gentoo-dev] Making systemd more accessible to "normal" users Matt Turner <mattst88@g.o>