2 * TODO: Give each global variable its own RefStor for locking
5 * Assignments of the same variable to itself are a big problem. The
6 underlying storge is dropped and then copied, but atm the drop code
7 cleans out the storage (e.g. any pointers or references).
9 * &array[n] is causing extra hanging locks in the interpreter.
11 * &lvalue --- do not set SRS flags!!! The address is not the local stack
14 * ptr.method() calls (other than new()) just don't understand whether
15 the method object is locked or unlocked. This isn't a problem for
16 ptr->method() calls through classes since the contents of a class is
17 always naturally unlocked (but still ref'd of course).
19 * XXX *pptr or *ptr-to-compound-type does not know whether the indirected
20 object resides in naturally locked or naturally unlocked storage.
22 --> intention - disallow &locked-storage-ptr or &locked-storage-comptype,
23 require the use of 'unlocked' on the storage to be able to take its
24 address. Relax requirements if TF_HASLVPTR is false.
26 * Do not allow the address of any return argument be taken during a
27 call. Messes up content-locked return data.
29 * Global compound types need to be fixed to not try to maintain locks.
31 * Lock global and (class) object storage in GenGetDeclarationData()
33 * on LValueStor, do not flow-through LVRSGet's. Instead require that
34 LValueStor's passed as an argument or return value have their
35 underlying object locked (and other LValueStor's do not, even if on
38 This makes .new() work as the lock will be borrowed if relaying an
39 lvs passed in, and if not relaying the caller must lock the underlying.
41 When an object is forwarded by constructing an LValueStor for it,
42 the lock must be borrowed if the object is already locked. Access
43 semantics are such that this should always be the case.
45 * insnx86.c line 149 ish.. remove rea->regused ???
47 * gen_refstor.c ~720, make DOP_ARGS_STORAGE cacheable. But to
48 work properly RAS needs to do a dummy load of each argument
49 from EA[non-cache] to EA.CACHE to ensure that the register
50 is properly initialized. And for pointers, to sync back
53 * Cacheability of compound types in RAS ?
55 * see classes strlen() code. CMP/BRA opt causes RSPUT to be improperly
56 handled (fix by optimizing out the PRSGET/RSPUT sequence for the
57 comparison? what if both args need it?)
59 * optimize const global access to immediate value XXX doesn't currently
60 happen (strlen()'s test against NIL).
62 * optimized stack storage in llvm is not cleared
64 * return/result sequencing must be validated by the resolver. break/continue
65 may not cross a result; point and result; cannot be conditionalized in one
66 path unless it is also present in the other.
68 * Add 'finalize' to class/subclass decl indicating that no further subclassing
71 * libgen initializers for global declarations w/ assignments or type defaults.
73 * resolver propagate CTF_RESOLVING flag in run-time context.
75 * resolver and runtime and generator - validate when casting pointers
76 that an integral number of objects of the target fits in the source.
77 Otherwise the cast is illegal.
79 * tests/refine.d - method call through non-replicated (i.e. not part of
80 subclass) @ref currently executes the superclass method
81 instead of looking up the actual method via the type.
83 * generally speaking must change how interfaces and classes work when making
84 field accesses. Disallow in interfaces, allow otherwise but then restrict
85 how refine operates so field offsets remain consistent across all
86 sub-classes. Disallow alias access via @ref?
88 * resolver - Disallow binding of incompatible subclasses to a superclass.
89 Do this by flagging the incompatibility.
91 * resolver - Enforce that non-interface classes cannot contain interfaces.
93 * resolver - Enforce that it is not allowed to take the sizeof() an interface
96 * resolver - Enforce that one cannot instantiate/declare an interface type.
98 * resolver - Enforce that object accessors can only be to non-generic elements.
100 * resolver - Reorder generic elements in an interface to the end for offset
103 * resolver - Do not allow pointer arithmatic on references.
105 * libgen - chained temporary data returned along one side of an expressin
106 can get corrupted by the other side. On the top-level temporary data is
109 * libgen - RunRelsRefStor in operators a mess.
111 * libgen - exp.c RunCastExpPtrToRef() is basically messing with the
112 type of the passed-in lvalue, which is really a bad idea.
114 * read, write, sizeof, arysize, and bcopy currently taking Int32 !!!'s. FIX!
117 * When realizing persistent store we have to clean out any previously
120 * map files containing structures with pointers can get corrupted. Maybe
121 we should just disallow pointers entirely for any persistent store.
123 * Persistent store mmap() file path is currently hacked up. Needs command
124 line arg to set the home directory for the program, and the default
125 should be "$HOME/.<program>/rune_%s.map". libgen/heap.c
127 * infinintely recursive aliases are not currently detected.
132 * get rid of need for 'auto' prefix in casts to distinguish between
133 variable names and types.
135 * share pens and also share font information
137 * Handle simple exposure aggregation and use clip to limit update area
138 (see tests/gfxslider.d)
140 * Wide char for SSTRING ? i.e. 'x'
142 isg for resolveType's assigned expression
144 DONE<<constructors/destructors for global elements and statement blocks>> ?
146 reference type resolver and libi/exp.c TOK_STRIND and TOK_CALL. METHOD CALLS
147 via ref types (don't screw around with the exp, dup it instead?)
149 refine: check validity of procedural refinements (different return types,
150 argument types, etc...)
152 is d_Search now obsolete?
154 Add support for 'super'. Either 'structural->super' or 'super.<blah>'
155 (for method calls), to access a superclass element that has been overloaded
158 We need to be able to copy a statement tree (procedure body) on the fly
159 in the resolver, when a subclass is passed to a procedure expecting its
160 superclass. We have to keep track of compound types (proc args) in order
161 to cache the various versions of the procedure. We already have d_Super
162 to locate the procedure in a TOK_CALL if the procedure body is NULL.
163 The modified procedure must be able to pass a different sg_Level.. i.e.
164 the one for the superclass, even though all of its types will be the actual
165 class (the subclass), in order to locate the proper declaration (a declaration
166 may be invisible to the subclass but visible to the superclass).
171 int ary[exp] ... the exp is shared across qualifications of the type,
172 which is bad. Also, what about compound types sharing the same semgroup?
175 double check 'isglobal' TF/SGF_NOINIT cache
177 cast procedures generally work but we really need to formalize
178 how ex_Type is set verses what the exp node really wants (e.g. for
179 procedures... look at ex_Type or look at the procdecl's return type?).
181 really fix cast... in libi/exp.c
183 Check global scope when indirecting through a type or class rather
184 then storage. only allow global scope for storage in that case.
186 non-global procedure called through object -- automatic "this" reference.
187 procedures embedded in class accessing class elements????? global vs
190 Implement STF_FORWARD checking (do not allow forward reference stack
191 declarations to other decls at the same semantic level.)
193 Make procedural refinement work
197 Implement heap variables (implement as initialized pointers / use '@' ?)
199 Partial procedure args resolution
201 procedure pointers (point to the declaration)