1 |
On czw, 2017-06-15 at 18:07 +0200, Alexis Ballier wrote: |
2 |
> On Thu, 15 Jun 2017 17:59:13 +0200 |
3 |
> Michał Górny <mgorny@g.o> wrote: |
4 |
> |
5 |
> > On śro, 2017-06-14 at 16:09 +0200, Alexis Ballier wrote: |
6 |
> > > On Wed, 14 Jun 2017 15:57:38 +0200 |
7 |
> > > Michał Górny <mgorny@g.o> wrote: |
8 |
> > > [...] |
9 |
> > > > > [...] |
10 |
> > > > > > > > > > [1]:https://wiki.gentoo.org/wiki/User:MGorny/GLEP:ReqUse |
11 |
> > > > > > > > > |
12 |
> > > > > > > > > I really don't like the reordering thing. Even the |
13 |
> > > > > > > > > restricted syntax does not fix the issue with '^^ ( a b |
14 |
> > > > > > > > > ) b? ( a )' already mentioned here. It'd be much better |
15 |
> > > > > > > > > and simpler for the spec just to assign a fixed value |
16 |
> > > > > > > > > and use the solving rules with those. |
17 |
> > > > > > > > |
18 |
> > > > > > > > You're not going to convince me by providing examples |
19 |
> > > > > > > > that are utterly broken by design and |
20 |
> > > > > > > > meaningless ;-). |
21 |
> > > > > > > |
22 |
> > > > > > > Well... if it's so obvious that the example is broken by |
23 |
> > > > > > > design that you don't even bother to explain why, I assume |
24 |
> > > > > > > you have an algorithm for that. Where is the code ? What |
25 |
> > > > > > > are the numbers ? How many ebuilds might fail after |
26 |
> > > > > > > reordering ? How can this be improved ? |
27 |
> > > > > > |
28 |
> > > > > > Are you arguing for the sake of arguing here? I just presumed |
29 |
> > > > > > that this example is so obviously broken there is no point |
30 |
> > > > > > wasting any more time on it. The code of nsolve clearly |
31 |
> > > > > > detects that, so I don't really understand what you're trying |
32 |
> > > > > > to prove here. |
33 |
> > > > > |
34 |
> > > > > Those are real questions. You should take breath, think a bit |
35 |
> > > > > about it, and try to run the 2 possible orderings of the ^^ |
36 |
> > > > > through nsolve or even solve.py. They both are very happy (and |
37 |
> > > > > are right to be) with the above ordering. You might want to |
38 |
> > > > > think a bit more about what is the relation between this broken |
39 |
> > > > > 10 chars example and the 10 lines python targets one below. |
40 |
> > > > > |
41 |
> > > > > You should also realize that all the above questions have |
42 |
> > > > > already been answered in length if you do as I suggest. |
43 |
> > > > |
44 |
> > > > No. I have already spent too much time on this. We're already long |
45 |
> > > > past all useful use cases, and now I feel like you're going to |
46 |
> > > > argue to death just to find a perfect algorithm that supports |
47 |
> > > > every absurd construct anyone can even write, if only to figure |
48 |
> > > > out the construct is completely useless. |
49 |
> > > |
50 |
> > > I'm not going to argue to death. It's already proven reordering is |
51 |
> > > broken. |
52 |
> > > |
53 |
> > > > If you want to play with it more, then please by all means do |
54 |
> > > > so. |
55 |
> > > |
56 |
> > > There is nothing to do for reordering. It's broken by design. |
57 |
> > > |
58 |
> > > > However, do not expect me to waste any more of my time on it. I've |
59 |
> > > > done my part, the code works for all reasonable use cases and |
60 |
> > > > solves all the problems I needed solving. If you want more, then |
61 |
> > > > it's your job to do it and solve the resulting issues. |
62 |
> > > |
63 |
> > > Like... writing code handling all the cases and describing how it |
64 |
> > > works ? We're past that. The only thing we're not past is that you |
65 |
> > > fail to understand it and attempt to block it. |
66 |
> > > |
67 |
> > |
68 |
> > Then please provide a single valid example that: |
69 |
> |
70 |
> app-text/wklej-0.2.1-r1 ^^ ( python_single_target_pypy |
71 |
> python_single_target_pypy3 python_single_target_python2_7 |
72 |
> python_single_target_python3_4 python_single_target_python3_5 |
73 |
> python_single_target_python3_6 ) python_single_target_pypy? |
74 |
> ( python_targets_pypy ) python_single_target_pypy3? |
75 |
> ( python_targets_pypy3 ) python_single_target_python2_7? |
76 |
> ( python_targets_python2_7 ) python_single_target_python3_4? |
77 |
> ( python_targets_python3_4 ) python_single_target_python3_5? |
78 |
> ( python_targets_python3_5 ) python_single_target_python3_6? |
79 |
> ( python_targets_python3_6 ) vim? ( ^^ ( python_single_target_python2_7 |
80 |
> ) ) |
81 |
> |
82 |
> |
83 |
> Simplified as: |
84 |
> ^^ ( a b ) c? ( b ) |
85 |
> |
86 |
> (see the pattern now ? :) ) |
87 |
> |
88 |
> > a. is completely 'correct' (that is, provides a valid, predictable |
89 |
> > and acceptable solution) with the default ordering O_a, |
90 |
> |
91 |
> c? ( b ) ^^ ( b a ) |
92 |
> |
93 |
> |
94 |
> > b. is not 'correct' with at least one reordering O_b (assuming only |
95 |
> > > > , ^^, ?? is subject to reordering), |
96 |
> |
97 |
> c? ( b ) ^^ ( a b ) |
98 |
> |
99 |
> > |
100 |
> > c. nsolve reports O_a as all good, and O_b as not good. |
101 |
> |
102 |
> I'll let you run this. It does. |
103 |
> |
104 |
> > The best way to convince me is through valid examples. |
105 |
> |
106 |
> |
107 |
> It is also easier to be convinced when you try to understand and ask |
108 |
> for clarifications instead of just rejecting without thinking :) |
109 |
> |
110 |
|
111 |
Ok, now I get your point. Not that I like it but I don't see any sane |
112 |
way around it. |
113 |
|
114 |
The question then is, how can you reliably ensure that developers will |
115 |
use the same ordering in cross-relevant packages? For example, consider |
116 |
the following: |
117 |
|
118 |
^^ ( provider_ssl_openssl provider_ssl_libressl ) |
119 |
|
120 |
Since those providers block each other, all packages will have to have |
121 |
either openssl or libressl consistently (or another provider). |
122 |
|
123 |
The reordering idea was mostly addressing this. However, it just |
124 |
occurred to me it only solved some the case when user selected both |
125 |
and not the one where neither was preferred over the other. |
126 |
|
127 |
Without reordering, I think we need to enforce the specific ordering |
128 |
consistently across the tree. Any idea how to achieve that? |
129 |
|
130 |
-- |
131 |
Best regards, |
132 |
Michał Górny |