Commit | Line | Data |
5dbb47c8 |
1 | == Todo for 2.0400 |
85910734 |
2 | |
1b1eb1ca |
3 | === RT Tickets |
4 | |
c78a59d7 |
5 | RT#59478/RT#63000 - 0+ overload causes NV conversion on == on perls before |
6 | 5.14 - this causes comparisons to fail when the number can't fit in an NV |
7 | without precision loss. I'd like to fix this in a more general way (forcing |
8 | anyone else who might be using == on tc objects to do weird things isn't very |
9 | good), although it's hard to test to see what actually works. |
10 | |
b927b1d4 |
11 | == Todo for 2.0600 |
12 | |
85910734 |
13 | === Revise MetaRole API to reunify class/role metaroles: |
14 | |
15 | apply_metaroles( |
16 | for => $meta, |
17 | roles => { |
18 | attribute => [...], |
19 | class => [...], |
20 | role_attribute => [ ... ], |
21 | } |
22 | ); |
23 | |
24 | If the $meta is a class, we apply the roles to the class. If it's a role, we |
25 | hold onto them and apply them as part of applying the role to a class. |
26 | |
27 | To make this all work nicely, we'll probably want to track the original role |
28 | where a method was defined, just like we do with attributes currently. We'll |
29 | also need to store method modifiers with their original role, which may mean |
30 | adding some sort of Moose::Meta::Role::MethodModifier class. |
31 | |
32 | For each role-specific thing (methods, attributes, etc.) we should allow a |
33 | role_attribute, role_method, etc. key. The common case will be that the |
34 | metaroles are intended for the consuming class, but we should allow for |
35 | metaroles on the role's metaobjects as well. |
36 | |
ef0801e7 |
37 | === Deprecate old-style Moose extensions |
38 | |
39 | Moose extensions that work by calling Moose->init_meta(metaclass => |
40 | 'Some::Custom::Metaclass', ...) during their own init_meta should be |
41 | deprecated, so they can be removed later (this should fix the issues with |
42 | init_meta generation in Moose::Exporter, see RT51561) |
43 | |
d8fed7e3 |
44 | This needs to wait until the previous fix gets in, since it will hopefully eliminate the need to write custom init_meta methods entirely. |
45 | |
24b9aca7 |
46 | === Attributes in roles need to be able to participate in role composition |
47 | |
48 | Right now, this fails with no decent workaround: |
49 | |
50 | package R1; |
51 | use Moose::Role; |
52 | has foo => (is => 'ro'); |
53 | |
54 | package R2; |
55 | use Moose::Role; |
56 | with 'R1'; |
57 | requires 'foo'; |
58 | |
59 | package C; |
60 | use Moose; |
61 | with 'R2'; |
62 | |
63 | Role attributes really need to be able to participate in role-role combination. |
64 | This should also fix "with 'Role1', 'Role2'" being broken when Role1 implements |
65 | a method as an accessor and Role2 requires that method, but at least in that |
66 | case you can split it into two 'with' statements with minimal loss of |
67 | functionality. |
68 | |
69 | === Method modifiers in roles should silently add 'requires' for them |
70 | |
71 | This shouldn't be a functionality change, just a better error message (and |
72 | better introspectability). This shouldn't happen if the role already contains a |
73 | method by that name, so it'll depend on the previous fix going in (so "has foo |
74 | => (is => 'ro'); around foo => sub { }" doesn't produce a 'requires' entry). |
75 | |
76 | === Add overloading support |
77 | |
78 | or at least, don't break existing overloading support |
79 | |
80 | This shouldn't treat the overloading stuff as actual methods, since that's just |
81 | an implementation detail, but we should provide an API for add_overload, |
82 | get_overload, get_overload_list, etc. In particular, this would allow |
83 | namespace::autoclean to not break things. |
84 | |
85 | Also, MooseX::Role::WithOverloading should probably be cored. |
86 | |
d8fed7e3 |
87 | This should probably also wait for the metarole unification fix, to avoid the |
88 | ::WithOverloading stuff being too insane. |
89 | |
24b9aca7 |
90 | |
91 | == Todo for later |
92 | |
3091d476 |
93 | === has +foo in roles |
94 | |
95 | There's no actual reason for this not to work, and it gets asked often enough |
96 | that we really should just do it at some point. |
97 | |
24b9aca7 |
98 | === Actual API for metaclass extensions |
99 | |
100 | Right now, the only way to bundle multiple metaclass traits is via |
101 | Moose::Exporter. This is unhelpful if you want to apply the extension to a |
102 | metaclass object rather than a class you're actually writing. We should come up |
103 | with an API for doing this. |
104 | |
105 | === MooseX::NonMoose in core |
106 | |
107 | I think all of the actual issues are solved at this point. The only issue is |
108 | the (necessary) implementation weirdness - it sets up multiple inheritance |
109 | between the non-Moose class and Moose::Object, and it installs a custom |
110 | constructor method at 'extends' time (although perhaps this could be solved by |
111 | moving some of the logic back into Moose::Object::new?). Other than that, it |
112 | handles everything transparently as far as I can tell. |
113 | |
114 | === Fix attribute and method metaclass compatibility |
115 | |
116 | So i got this wrong when rewriting it last year - right now, metaclass compat |
117 | checks the default attribute and method metaclasses, which is wrong. This means |
118 | that if a parent class does "use MooseX::FollowPBP", then attributes declared |
119 | in a subclass will get PBP-style accessors, which is quite surprising. |
120 | |
121 | On the other hand, sometimes metaclasses might need to be able to say "I'm |
122 | going to assume that all of my attributes at least inherit from this custom |
123 | class", so we might need to split it into "default specified by the user" and |
124 | "default specified by the metaclass" and only do compat checking on the second? |
125 | I'm not actually sure this is a valid use case though. |
126 | |
127 | Something that probably should be taken into account though is attributes and |
128 | methods that extend existing attributes or methods from a superclass should |
129 | inherit the metaclass of the existing one. Also not sure if this is correct, |
130 | but something to think about. |
131 | |
132 | === Rename a bunch of the public API methods |
133 | |
134 | Right now the public API is kind of a mess - we have things like get_method vs |
135 | find_method_by_name (you almost always want to use the latter), there being no |
136 | has_method equivalent that checks superclasses, get_method_list being public |
137 | but only returning method names, while _get_local_methods is private (returning |
138 | method objects), and yet neither of those looks at superclasses, and basically |
139 | none of this naming follows any kind of consistent pattern. |
140 | |
141 | What we really need is a consistent and easy to remember API where the method |
142 | that people would think to use first is the method that they actually mean. |
143 | Something like renaming find_method_by_name to find_method, and get_method to |
144 | find_local_method or something along those lines. |
145 | |
6ddfb832 |
146 | === use Sub::Identify instead of doing our own thing with get_code_info |
147 | |
148 | No idea why we stopped using Sub::Identify in the past, but there's no reason |
149 | not to do this. We have a bug fix in our version (the isGV_with_GP thing), so |
150 | this should be submitted to Sub::Identify first. |
151 | |
c762df57 |
152 | === Move method modifiers out to an external module |
153 | |
154 | Class::Method::Modifiers uses a different method for doing method modifiers, |
155 | which I'm not sure why we aren't using in Moose right now. Optionally using |
156 | Class::Method::Modifiers::Fast would be even better - it uses Data::Util to |
157 | implement XS method modifiers, which could help things a lot. |
158 | |
159 | === Move type constraints out to an external module |
160 | |
161 | There's nothing about our type constraint system that requires being tied to |
162 | Moose - it's conceptually an entirely separate system that Moose just happens |
163 | to use. Splitting it out into its own thing (that Moose could extend to add |
164 | things like role types) would make things conceptually a lot cleaner, and would |
165 | let people interested in just the type system have that. |
166 | |
167 | === Merge Class::MOP and Moose |
168 | |
169 | This is a long term goal, but would allow for a lot of things to be cleaned up. |
170 | There's a bunch of stuff that's duplicated, and other stuff that's not |
171 | implemented as well as it could be (Class::MOP::Method::Wrapped should be a |
172 | role, for instance). |
173 | |
24b9aca7 |
174 | |
175 | == Things to contemplate |
176 | |
177 | === Does applying metaroles really need to reinitialize the metaclass? |
178 | |
179 | Seems like the logic that's actually necessary is already contained in |
180 | rebless_instance, and not reinitializing means that existing attributes and |
181 | methods won't be blown away when metaroles are applied. |
182 | |
183 | === Do we want to core namespace::autoclean behavior somehow? |
184 | |
185 | This would add Variable::Magic as a required XS dep (not a huge deal at the |
186 | moment, since Sub::Name is also a required XS dep, but it'd be nice for Moose |
187 | to be able to be pure perl again at some point in the future, and I'm not sure |
188 | what the relative chances of Sub::Name vs Variable::Magic making it into core |
189 | are). If we enabled it by default, this would also break things for people who |
190 | have introduced Moose into legacy-ish systems where roles are faked using |
191 | exporters (since those imported methods would be cleaned). |
192 | |
193 | If we decide we want this, we may want to core it as an option first ("use |
194 | Moose -clean" or so), and move to making it the default later. |
195 | |
196 | === Fix the error system |
197 | |
198 | oh god it's terrible |
199 | |
200 | More specifically, we really want exception objects. |
201 | |
202 | === Should using -excludes with a role add 'requires' for excluded methods? |
203 | |
204 | It seems to make sense, since otherwise you're violating the role's API |
205 | contract. |
206 | |
207 | |
98fb200c |
208 | == TODO test summary |
209 | |
210 | Note that some of these are fairly old, and may not be things we actually want |
211 | to do anymore. |
212 | |
213 | === t/basics/basic_class_setup.t |
214 | |
215 | Imports aren't automatically cleaned. Need to think about bringing |
216 | namespace::autoclean functionality into core. |
217 | |
218 | === t/bugs/create_anon_recursion.t |
219 | |
220 | Loading Moose::Meta::Class (or probably a lot of other metaclasses) before |
221 | loading Moose or Class::MOP causes issues (the bootstrapping gets confused). |
222 | |
223 | === t/bugs/handles_foreign_class_bug.t |
224 | |
225 | There should be a warning when delegated methods override 'new' (and possibly |
226 | others?). |
227 | |
228 | === t/bugs/role_caller.t |
229 | |
230 | Role methods should be cloned into classes on composition so that using |
231 | caller(0) in a role method uses the class's package, not the role's. |
232 | |
233 | === t/cmop/metaclass_incompatibility.t |
234 | |
235 | If a child class is created before a parent class, metaclass compatibility |
236 | checks won't run on the child when the parent is created, and so the child |
237 | could end up with an incompatible metaclass. |
238 | |
239 | === t/cmop/modify_parent_method.t |
240 | |
241 | Modifying parent class methods after a child class has already wrapped them |
242 | with a method modifier will cause the child class method to retain the original |
243 | method that it wrapped, not the new one it was replaced with. |
244 | |
245 | === t/immutable/inline_close_over.t |
246 | |
247 | Initializers and custom error classes still close over metaobjects. |
248 | Initializers do it because the initializer has to be passed in the attribute |
249 | metaobject as a parameter, and custom error classes can't be automatically |
250 | inlined. |
251 | |
252 | === t/metaclasses/moose_exporter_trait_aliases.t |
253 | |
254 | Renamed imports aren't cleaned on unimport. For instance: |
255 | |
256 | package Foo; |
257 | use Moose has => { -as => 'my_has' }; |
258 | no Moose; |
259 | # Foo still contains my_has |
260 | |
261 | === t/metaclasses/reinitialize.t |
262 | |
263 | Special method types can't have method metaroles applied. Applying a method |
264 | metarole to a class doesn't apply that role to things like constructors, |
265 | accessors, etc. |
266 | |
267 | === t/roles/compose_overloading.t |
268 | |
269 | Overload methods aren't composed during role composition (this is detailed |
270 | above - Add overloading support). |
271 | |
272 | === t/roles/method_modifiers.t |
273 | |
274 | Method modifiers in roles don't support the regex form of method selection. |
275 | |
276 | === t/roles/role_compose_requires.t |
277 | |
278 | Accessors for attributes defined in roles don't satisfy role method |
279 | requirements (this is detailed above - Attributes in roles need to be able to |
280 | participate in role composition). |
281 | |
282 | === t/todo_tests/exception_reflects_failed_constraint.t |
283 | |
284 | Type constraint failures should indicate which ancestor constraint failed - |
285 | subtype 'Foo', as 'Str', where { length < 5 } should mention Str when passed an |
286 | arrayref, but not when passed the string "ArrayRef". |
287 | |
288 | === t/todo_tests/moose_and_threads.t |
289 | |
290 | On 5.8, the type constraint name parser isn't thread safe. |
291 | |
292 | === t/todo_tests/replacing_super_methods.t |
293 | |
294 | Modifying parent class methods after a child class has already wrapped them |
295 | with a override will cause 'super' in the child class to call the original |
296 | parent class method, not the one it was overridden with. |
297 | |
298 | === t/todo_tests/required_role_accessors.t |
299 | |
300 | Role attribute accessors don't satisfy requires from roles they consume. |
301 | |
302 | === t/todo_tests/role_insertion_order.t |
303 | |
304 | Roles don't preserve attribute insertion_order. |
305 | |
306 | === t/todo_tests/various_role_features.t |
307 | |
308 | * Role attribute accessors don't satisfy requires from roles they consume. |
309 | * Role combination should produce a conflict when one role has an actual method |
310 | and the other role has an accessor. |
311 | * Role attribute accessors should not override methods in the class the role is |
312 | applied to. |
313 | * Role attribute accessors should be delegated when a class does |
314 | handles => 'Role'. |
315 | * Delegating to a role doesn't make $class->does('Role') true. |
316 | * Method modifier in a role doesn't create a method requirement. |
317 | * Role->meta->has_method('attr_accessor') is false. |
318 | |
319 | === t/type_constraints/type_names.t |
320 | |
321 | Type constraint object constructors don't validate the type name provided. |
322 | |
323 | |
85910734 |
324 | == Old todo (does anyone look at this?) |
325 | |
8b59f8d6 |
326 | ------------------------------------------------------------------------------- |
7af2c1d2 |
327 | BUGS |
d03bd989 |
328 | ------------------------------------------------------------------------------- |
43d599e5 |
329 | |
330 | ------------------------------------------------------------------------------- |
7af2c1d2 |
331 | FEATURES |
8b59f8d6 |
332 | ------------------------------------------------------------------------------- |
333 | |
d4967760 |
334 | - DDuncan's Str types |
335 | |
d03bd989 |
336 | subtype 'Str' |
337 | => as 'Value' |
338 | => where { Encode::is_utf8( $_[0] ) or $_[0] !~ m/[^0x00-0x7F]/x } |
d4967760 |
339 | => optimize_as { defined($_[0]) && !ref($_[0]) }; |
340 | |
d03bd989 |
341 | subtype 'Blob' |
342 | => as 'Value' |
343 | => where { !Encode::is_utf8( $_[0] ) } |
d4967760 |
344 | => optimize_as { defined($_[0]) && !ref($_[0]) }; |
345 | |
8b59f8d6 |
346 | - type unions |
347 | |
d03bd989 |
348 | Add support for doing it with Classes which do not have |
8b59f8d6 |
349 | a type constraint yet created |
350 | |
351 | - type intersections |
352 | |
353 | Mostly just for Roles |
354 | |
355 | - inherited slot specs |
356 | |
db1ab48d |
357 | 'does' can be added to,.. but not changed |
358 | (need type unions for this) |
8b59f8d6 |
359 | |
8b59f8d6 |
360 | - proxy attributes |
361 | |
db1ab48d |
362 | a proxied attribute is an attribute |
d03bd989 |
363 | which looks like an attribute, |
364 | talks like an attribute, smells |
365 | like an attribute,.. but if you |
366 | look behind the curtain,.. its |
db1ab48d |
367 | over there.. in that other object |
368 | |
369 | (... probably be a custom metaclass) |
8b59f8d6 |
370 | |
f90e052d |
371 | - local coerce |
372 | |
373 | [13:16] mst stevan: slight problem with coerce |
374 | [13:16] mst I only get to declare it once |
375 | [13:17] mst so if I'm trying to declare it cast-style per-source-class rather than per-target-class |
376 | [13:17] mst I am extremely screwed |
377 | [13:17] stevan yes |
d03bd989 |
378 | [13:17] stevan they are not class specific |
f90e052d |
379 | [13:18] stevan they are attached to the type constraint itself |
380 | [13:18] * stevan ponders anon-coercion-metaobjects |
381 | [13:18] mst yes, that's fine |
382 | [13:19] mst but when I declare a class |
383 | [13:19] mst I want to be able to say "this class coerces to X type via <this>" |
384 | [13:19] stevan yeah something like that |
385 | [13:19] stevan oh,.. hmm |
386 | [13:20] stevan sort of like inflate/deflate? |
387 | [13:20] stevan around the accessors? |
388 | [13:25] * bluefeet has quit (Remote host closed the connection) |
389 | [13:27] mst no |
390 | [13:27] mst nothing like that |
391 | [13:27] mst like a cast |
392 | [13:31] mst stevan: $obj->foo($bar); where 'foo' expects a 'Foo' object |
393 | [13:31] mst stevan: is effectively <Foo>$bar, right? |
394 | [13:32] mst stevan: I want to be able to say in package Bar |
395 | [13:32] mst stevan: coerce_to 'Foo' via { ... }; |
396 | [13:32] mst etc. |
d03bd989 |
397 | [13:53] stevan hmm |
398 | |
f3dc2e80 |
399 | ----------------------------------------------------------- |
400 | -- Type Constraints refactor |
401 | ----------------------------------------------------------- |
402 | |
403 | - add support for locally scoped TC |
404 | |
405 | This would borrow from MooseX::TypeLibrary to prefix the TC with the name |
406 | of the package. It would then be accesible from the outside as the fully |
407 | scoped name, but the local attributes would use it first. (this would need support |
408 | in the registry for this). |
409 | |
410 | - look into sugar extensions |
411 | |
412 | Use roles as sugar layer function providers (ala MooseX::AttributeHelpers). This |
413 | would allow custom metaclasses to provide roles to extend the sugar syntax with. |
414 | |
415 | (NOTE: Talk to phaylon a bit more on this) |
416 | |
417 | - allow a switch of some kind to optionally turn TC checking off at runtime |
418 | |
419 | The type checks can get expensive and some people have suggested that allowing |
420 | the checks to be turned off would be helpful for deploying into performance |
421 | intensive systems. Perhaps this can actually be done as an option to make_immutable? |
422 | |
423 | - misc. minor bits |
424 | |
425 | * make the errors for TCs use ->message |
426 | * look into localizing the messages too |
427 | * make ANON TCs be lazy, so they can possibly be subsituted for the real thing later |
428 | * make ANON TCs more introspectable |
429 | * add this ... |
430 | |
431 | # |
432 | # Type Definition |
433 | # |
434 | subtype 'Username', |
435 | from 'Str', |
436 | where { (/[a-z][a-z0-9]+/i or fail('Invalid character(s)')) |
437 | and (length($_) >= 5 or fail('Too short (less than 5 chars)')) |
438 | } |
439 | on_fail { MyException->throw(value => $_[0], message => $_[1]) }; |
440 | |
441 | # fail() will just return false unless the call is made via |
442 | $tc->check_or_fail($value); |
443 | |
444 | * and then something like this: |
445 | |
446 | subtype Foo => as Bar => where { ... } => scoped => -global; |
447 | subtype Foo => as Bar => where { ... } => scoped => -local; |
448 | |
449 | # or |
450 | |
451 | subtype Foo => as Bar => where { ... } => in __PACKAGE__ ; |
452 | |
453 | # or (not sure if it would be possible) |
454 | |
455 | my $Foo = subtype Bar => where { ... }; |
456 | |
457 | # ---------- |
458 | |
459 | [17:10] <autarch> stevan: it should do it if I pass coerce => 1 as part of the attribute definition |
460 | [17:12] <stevan> autarch: what I am not 100% sure of is how to tell it to deep coerce and when to not |
461 | [17:13] <stevan> cause a basic coerce is from A to B |
462 | [17:13] <autarch> hmm |
463 | [17:13] <stevan> which is valid for collection types too |
464 | [17:13] <stevan> deep coercion is what you are asking for |
465 | [17:13] <autarch> yeah |
466 | [17:13] <stevan> so perhaps we add deep_coerce => 1 |
467 | [17:13] <stevan> which will do it |
468 | [17:13] <autarch> that's fine for me |
469 | [17:13] <stevan> k |
470 | |
471 | coerce_deeply => 1 # reads better |
472 | |
7af2c1d2 |
473 | ------------------------------------------------------------------------------- |
474 | INTERNALS |
475 | ------------------------------------------------------------------------------- |
476 | |
477 | - rationalize all the get_X methods for classes (and roles) |
478 | |
479 | We have get_attribute, get_attributes_list, get_all_attributes, |
480 | etc. First, we need to make the method names consistent. If something |
481 | returns an attribute vs a name, that needs to be clear from the method |
482 | name. We also need to make sure that local vs. "entire inheritance |
483 | chain" is clear from the name. |
484 | |
7af2c1d2 |
485 | This is mostly a CMOP change. |
486 | |
487 | - Metaclass constructors |
488 | |
489 | There's a _lot_ of different conventions in here. Some things to consider: |
490 | |
491 | * new vs _new |
492 | * allowing new( 'name', %args ) vs ( name => 'name', %args ) |
493 | * Method->wrap vs Method->new |
494 | |
7af2c1d2 |
495 | - Moose::Meta::TypeConstraint::Parameter{izable,ized} |
496 | |
497 | The relationship between these two classes is very odd. In particular, |
498 | this line in Parameterized is insane: |
499 | |
500 | foreach my $type (Moose::Util::TypeConstraints::get_all_parameterizable_types()) { |
501 | |
502 | Why does it need to loop through all parameterizable types? Shouldn't |
503 | it know which parameterizable type it "came from"? |
504 | |
505 | - Moose::Util::TypeConstraints vs Moose::Meta::Type{Coercion,Constraint} |
506 | |
507 | The Util module has _way_ too much functionality. It needs to be |
508 | refactored so it's a thin sugar layer on top of the meta API. As it |
509 | stands now, it does things like parse type names (and determine if |
510 | they're valid), manage the registry, and much more. |
511 | |
7af2c1d2 |
512 | - Anything with a _(meta)?class method |
513 | |
514 | Every method that returns a class name needs to become a rw attribute |
515 | that can be set via the constructor. |
d03bd989 |
516 | |
cac484fa |
517 | - The Moose::Error stuff |
518 | |
519 | This is sort of half-implemented. We still use Carp directly, and the |
520 | internals can't decide how to throw an error (is it |
521 | Moose->throw_error, __PACKAGE__->throw_error, what?). |
522 | |
523 | The internals need to be made consistent before we expose this to the |
524 | rest of the world. |
525 | |
8b59f8d6 |
526 | ------------------------------------------------------------------------------- |
527 | TO PONDER |
528 | ------------------------------------------------------------------------------- |
529 | |
530 | - Moose "strict" mode |
531 | |
532 | use Moose 'strict'; This would allow us to have all sort of expensive tests |
d03bd989 |
533 | which can be turned off in prod. |
534 | |
8b59f8d6 |
535 | - Moose::Philosophy.pod |
536 | |
537 | To explain Moose from a very high level |
538 | |
687e52bb |
539 | - moosedoc |
8b59f8d6 |
540 | |
687e52bb |
541 | We certainly have enough meta-information to make pretty complete POD docs. |
d03bd989 |
542 | |
543 | |
544 | |