Do the transformCall using the original types2-derived type of the call
(in particular, the types of the params as non-shapes). Currently, since
we were using the param types of the instantiation, we might add in
interface conversions to an interface with shapes in the one case of a
full-instantiated generic call. So, we do the transformCall() before
installing the shaped-based instantiation. transformCall() works
correctly even in the case of OCALL/FUNCINST.
Fixed two related bugs:
- Fixed case where we still were not correctly substituting the types
for a function instantiation.
- The type substituter needs to copy field flags while substituting in
tstruct.
Change-Id: I14e960737d6840a75846ede480e6650534ba3af3
Reviewed-on: https://go-review.googlesource.com/c/go/+/340259
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
- Removed gcshapeType - we're going with more granular shapes for now, and gradually
coarsening later if needed.
- Put in early return in getDictionarySym(), so the entire rest of the
function can be un-indented by one level.
- Removed some duplicated infoprint calls, and fixed one infoprint
message in getGfInfo.
Change-Id: I13acce8fdabdb21e903275b53ff78a1e6a378de2
Reviewed-on: https://go-review.googlesource.com/c/go/+/339901
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
In this case, we can't use an itab for doing a bound call, since we're
converting from an interface to an interface. We do a static or dynamic
type assert in new function assertToBound().
The dynamic type assert in assertToBound() is only needed if a bound is
parameterized. In that case, we must do a dynamic type assert, and
therefore need a dictionary entry for the type bound (see change in
getGfInfo). I'm not sure if we can somehow limit this case, since using
an interface as a type arg AND having the type bound of the type
arg be parameterized is a very unlikely case.
Had to add the TUNION case to parameterizedBy1() (which is only used for
extra checking).
Added a bunch of these test cases to 13.go, which now passes.
Change-Id: Ic22eed637fa879b5bbb46d36b40aaad6f90b9d01
Reviewed-on: https://go-review.googlesource.com/c/go/+/339898
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
- Changed some early returns to asserts (instantiateMethods and Shapify
should never take a shape arg)
- Added suggested change (by Ingo) to use copy() in getInstantiation()
- Clarified that shape types never have methods in Shapify(), removed
some TODO comments.
Change-Id: Ia2164ffe670a777f7797bbb45c7ef5e6e9e15357
Reviewed-on: https://go-review.googlesource.com/c/go/+/338971
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
The other uses of Unshapify were really only there to allow for the
dictionary checking code at the beginning of generic functions/methods.
But that will go away as soon as we start combining real shapes. If we
get rid of that code, we can get rid of the unshapify calls elsewhere.
The only tricky part is that getInstantiation now gets targs that may each
either be a shape or concrete type, and it must translate any concrete
types to shapes, while leaving the already existing shapes.
Change-Id: Ib2b9072b921f8e064958548a1078d82f1d040c9f
Reviewed-on: https://go-review.googlesource.com/c/go/+/338289
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Rewrite a method expression such as 'T.String' (where T is type param
and String is part of its type bound Stringer) as:
func(rcvr T, other params...) {
return Stringer(rcvr).String(other params...)
}
New function buildClosure2 to create the needed closure. The conversion
Stringer(rcvr) uses the dictionary in the outer function.
For a method expression like 'Test[T].finish' (where finish is a method
of Test[T]), we can already deal with this in buildClosure(). We just
need fix transformDot() to allow the method lookup to fail, since shapes
have no methods on them. That's fine, since for any instantiated
receiver type, we always use the methods on the generic base type.
Also removed the OMETHEXPR case in the main switch of node(), which
isn't needed any (and removes one more potential unshapify).
Also, fixed two small bugs with handling closures that have generic
params or generic captured variables. Need to set the instInfo for the
closure in the subst struct when descending into a closure during
genericSubst() and was missing initializing the startItabConv and gfInfo
fields in the closure info.
Change-Id: I6dadedd1378477936a27c9c544c014cd2083cfb7
Reviewed-on: https://go-review.googlesource.com/c/go/+/338129
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Introduce new dynamic dottype operations which take a dynamic
instead of static type to convert to.
Change-Id: I5824a1fea056fe811b1226ce059e1e8da1baa335
Reviewed-on: https://go-review.googlesource.com/c/go/+/337609
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Fix the cons.go missing method error. Mark all the methods of
instantiated interface types as used. We could try to record all the
exact methods used for generic interface types, but for now, just mark
all the methods as used so that their methods are not dead-code
eliminated.
Change-Id: I35685eda82476244371379b97691a1b8506ef0f7
Reviewed-on: https://go-review.googlesource.com/c/go/+/337349
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
append is fine using shape types.
Change-Id: Iae829b9b5929d4dc7aa74bed57da13d4f6d746be
Reviewed-on: https://go-review.googlesource.com/c/go/+/337669
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
HasShape needs a TINTER case.
Add a test for x.(T) in various situations. Needs the fix above.
Also remove ONEW unshapify case. It is ok for ONEW to have a shape
type, as it will just be passed to mallocgc, or possibly used as a
stack object type, both of which are ok.
Change-Id: Ibddf8f5c8c254d32cb5ebcaca7dc94b4c00ab893
Reviewed-on: https://go-review.googlesource.com/c/go/+/337231
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
We just need to substitute from the typeparams to the shapes for the dst
type of the bound.
Removed concretify substituter, not used anymore. Also removed
shape2params, not needed anymore.
However, since the dst type is now not concrete, this gives more cases
where the linker can't find a method.
I realized that we need to call MarkUsedIfaceMethod to mark a method as
used on a particular interface, else a type's method can be still
deadcoded even though MarkTypeUsedInInterface has been called on the
concrete type. I added a new version MarkUsedIfaceMethodIndex to fit my
use case.
Change-Id: Id67b72b350889dd3688b42739c337d5d79a0d1a2
Reviewed-on: https://go-review.googlesource.com/c/go/+/337230
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
This operation computes just the data field needed to put its argument
into an interface. Used by generics because we produce the type field
of an interface using dictionaries (instead of statically).
With this operation defined, we can now assert that shape types
are never marked as used in interfaces (the only previous use
was IDATA(CONVIFACE(t))).
Change-Id: Idb1eb5f3b238285cb99413d382599c0621b7681a
Reviewed-on: https://go-review.googlesource.com/c/go/+/337109
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Add a test to make sure there's no invalid OCONVIFACEs when stenciling is done.
Use concrete types for the type of DOTTYPE and DOTTYPE2.
MarkTypeUsedInInterface - should we allow types with shape types
underneath? I think the itab CL will help with this (at least, for
a remaining cons.go issue).
Change-Id: I2c96d74e8daaca26cadc84ea94abb9a27c0bb240
Reviewed-on: https://go-review.googlesource.com/c/go/+/337069
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
This fix the case where a type param or derived type is converted to a
non-empty interface. Previously, we were converting to an empty
interface and then using DOTTYPE to convert to the correct non-empty
interface. In that case, we can get the needed itab directly from the
dictionary. This is needed for correctness from shapes, if the
destination interface is parameterized, else we will incorrectly convert
to the shape version of the interface.
Creating/writing an itab can involve generating wrappers for a bunch of
methods, which may use dictionaries. So, all the
dictionaries/instantiations are being generated on the fly and have
recursive relationships, it is simplest to finish creating/writing the
itabs at the end of the stenciling phase. So, we create a list of the
dictionaries which need to be completed by writing out their itab
entries.
The existing tests ordered.go, ifaceconv.go, and issue44688.go make use
of this optimization.
Got itab conversions for bound calls working, except for 13.go.
Also, want to get rid of the concretify, but I think we need more info
on the Bound from types2.
Change-Id: If552958a7b8a435500d6cc42c401572c367b30d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/336993
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Changes to add CONVIFACE nodes where possible in noder2, even when the
args are typeparams. The transformation to insert a CONVIFACE node can
usually happen when there an obvious assignment/conversion to an
interface type from a non-interface type. So, we now do this
tranformation for:
- direct conversions to an interface type
- function arguments that are implicitly converted to an interface
based on the parameter types.
- EQ/NE comparison of an interface and a non-interface
With this change, we can remove some special case checks for CONVIFACE
nodes after transformation in node(), and instead just have the one
check in the CONVIFACE check.
Change-Id: I7cf2ef920aebf9e5553210aeaf19f344e128ca62
Reviewed-on: https://go-review.googlesource.com/c/go/+/336992
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Make sure the instantiating types are the type parameters of the containing
function (or types derived from those).
The one exception is the argument of a OFUNCINST, whose type
parameters are those of the declaration site of the function or method
being instantiated. We skip those types.
Change-Id: I4b3ff22eb8a81a76476930cf8ed2a7dd6489d8b8
Reviewed-on: https://go-review.googlesource.com/c/go/+/336352
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
METHVALUE in a generic function (that is not called) was not causing
buildClosure() to be called and therefore not using dictionaries. Also,
had to add an extra check to make sure that if we have a FUNCINST
node above a METHVALUE, we only call buildClosure once.
Change-Id: I49756152fc343e5ac1c449e697960fc2a0f482ae
Reviewed-on: https://go-review.googlesource.com/c/go/+/336429
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
The builtins.go test is derived from
cmd/compile/internal/types2/testdata/check/builtins.go2, after removing
the error cases. Added a few extra tests for len/cap/append.
Fixed one bug, which is that DELETE operations can't be transformed if
their argument is a typeparam. Also, the tranform of LEN/CAP calls does
not need to be delayed. Removed out-date references to the old
typechecker in the comments.
Change-Id: If7a21506a7ff63ff7c8e87ccd614ef4ff3a0d3c8
Reviewed-on: https://go-review.googlesource.com/c/go/+/336010
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Remaining stuff from Dan's CL 335412 to get tests passing.
- Removed PTRLIT case in node() that was added with the gcshape change.
I don't think it is needed anymore.
- Modified (*itab).init() to add an irrelevant pointer (the itab
itself) for the '==' entry of interfaces with 'comparable'. That
entry is not used, so we just want to avoid the error that the given
type doesn't actually have an '==' method.
Change-Id: I9b2d4c8342c0b2048d76e003d95023f4ccd559f5
Reviewed-on: https://go-review.googlesource.com/c/go/+/336149
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Trust: Dan Scales <danscales@google.com>
Trust: Keith Randall <khr@golang.org>
FUNCINST nodes aren't instantiated correctly. Skip those types when
adding to the set of types considered for the dictionary. Those types
include those which are uninstantiated(have tparams), and those with
type parameters that aren't a parameter of the containing function
(they are the type parameter of the function being called).
Allow func types to be put in the dictionary.
Change-Id: I26bab85d3eebc2f54d02b4bba5e31407faf7c5b2
Reviewed-on: https://go-review.googlesource.com/c/go/+/336129
Reviewed-by: Dan Scales <danscales@google.com>
Trust: Dan Scales <danscales@google.com>
Trust: Keith Randall <khr@golang.org>
Calling MarkTypeUsedInInterface before inlining is problematic (when
using the containing function as the reference source), because we end
up attaching a relocation to the original function, which is then
deadcode removed because all instances of it were inlined. (All other
current uses of MarkTypeUsedInInterface happen after inlining, so they
don't have this problem.)
Instead, attach the relocation to the dictionary in which the type
appears. This should always work, because if the dictionary is ever
deadcode eliminated, then the type can never be put in an interface,
which means we indeed don't need its methods.
This CL is a bit overkill, as not *all* types in the dictionary need
this mark, but it is sufficient for now.
Change-Id: I2d39456691f1c3df8efe3bddc3b9017e0ef37a7f
Reviewed-on: https://go-review.googlesource.com/c/go/+/336012
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Dan Scales <danscales@google.com>
Still 1-1 with real types, but now with their own names!
Shape types are implicitly convertible to (and convertible from)
the types they represent.
Change-Id: I0133a8d8fbeb369380574b075a32b3c987e314d5
Reviewed-on: https://go-review.googlesource.com/c/go/+/335170
Run-TryBot: Keith Randall <khr@golang.org>
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Dan Scales <danscales@google.com>
We have almost 200 uses of the "assert" helper functions in noder and
typecheck. Clearly the tiny bit of extra convenience of writing a
one-line assertion rather than an if+panic is helpful, so we might as
well add functions for this to base itself so that it's easier to
write more informative error messages.
Change-Id: I06e2db2f0455af063937b25a53ca42f9413cf496
Reviewed-on: https://go-review.googlesource.com/c/go/+/336050
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Distinguish the gcshape of all top-level instantiated type from normal
concrete types, even if they have the exact same underlying "shape",
because in a function instantiation, any method call on this type arg
will be a generic method call (requiring a dictionary), rather than a
direct method call on the underlying type (no dictionary).
So, we add the instshape prefix to the gcshape name for instantiated
types, and we make it a defined type with that name.
Change-Id: I33056269d24f3451a2632a5ce6a481108f533c9c
Reviewed-on: https://go-review.googlesource.com/c/go/+/335169
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
When comparing a value whose type is a type parameter to an interface,
we need to convert that type parameter to an interface using the dictionary
entries.
Change-Id: I409c9e36e376fe4ef8163407d0fd4e84496d5b65
Reviewed-on: https://go-review.googlesource.com/c/go/+/334150
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
In the case that a generic function/method f does a method call on a
type param allowed by its bound, an instantiation of f may do a direct
method call of a concrete type or a method call defined on a generic
type, depending on whether the passed type in a concrete type or an
instantiated type with the appropriate method defined. See the test case
boundmethod.go added to this change.
In order to keep the dictionary format the same for all instantiations
of a generic function/method, I decided to have an optional
sub-dictionary entry for "bounds" calls. At the point that we are
creating the actual dictionary, we can then fill in the needed
sub-dictionary, if the type arg is an instantiated type, or a zeroed
dictionary entry, if type arg is not instantiated and the method will be
on a concrete type.
In order to implement this, I now fill in n.Selection for "bounds"
method calls in generic functions as well. Also, I need to calculate
n.Selection correctly during import for the case where it is now set -
method calls on generic types, and bounds calls on typeparams.
With this change, the dictionaries/sub-dictionaries are correct for
absdiff.go. The new test boundmethod.go illustrates the case where the
bound sub-dict entry is not used for a dictionary for stringify[myint],
but is used for a dictionary for stringify[StringInt[myint]].
Change-Id: Ie2bcb971b7019a9f1da68c97eb03da2333327457
Reviewed-on: https://go-review.googlesource.com/c/go/+/333456
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
1.go, 12.go: similar to calculating type sizes, we delay computing
instantiations during import until we get up to a top-level type, in
order to make sure recursive types are complete. But we should always
delay calculating sizes when we delay instantiating types, since
otherwise we may try to calculate the size of an incomplete type. So,
needed to add Defer/ResumeCheckSize in (*importReader).typ where we also
defer instantiations. (iimport.go)
2.go: when doing type substition, we have to handle named, parameterized
basic types i.e. the type has a type parameter even though the
underlying type is a basic type that doesn't depend on the parameter.
(subr.go)
3.go: for go 1.18, we allow arbitrary types in interfaces. We had
already allowed union types and tilde types, but didn't allow regular
non-interface types in Go 1.17 for compatibility. Just skip an error
in the case of 1.18. (size.go)
5.go: types2 and types1 differ in how they print out interfaces. types1
puts a space between "interface" and "{", types2 does not. So, since
some typenames come from types2 and some from types1, we need to remove
the space when printing out type arguments. (iimport.go/subr.go)
9.go: in subst.node(), we were missing the KeyExpr case where a node has
no type. The assertion is just there, to make sure we understand all the
cases where there is no type to translate. We could just remove the
whole error check. (stencil.go)
13.go: in subst.node(), missed handling the case where a method
expression is immediate called (which of course, is quite unusual, since
then there's no real reason to have used the method expression syntax in
that case). Just needed to add ir.OMETHEXPR in the OCALL switch
statement. (stencil.go)
Change-Id: I202cbe9541dfafe740e3b84b44982d6181738ea0
Reviewed-on: https://go-review.googlesource.com/c/go/+/333165
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Up to this point, we were only handling typeparams that were converted
to empty or non-empty interfaces. But we have a dictionary entry for
each derived type (i.e. type derived from typeparams) as well. So, when
doing a conversion, look for the source type in both the type params and
derived types of the generic info, and then use the appropriate
dictionary entry.
Added some cases to ifaceconv.go (e.g. converting []T to an empty
interface).
Change-Id: I7bbad0128bec20ccccd93ae1d65c1ffd44ca79a0
Reviewed-on: https://go-review.googlesource.com/c/go/+/333011
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
The (*T).M method expression is where M is a value method, but the type
(*T) is a pointer to the main type. In this case, after following any
embedded fields, we need to add an extra star operator when using the
receiver arg in the closure call.
Thanks to Cuong for finding/pointing out an example for this case
(typeparam/mdempsky/14.go) This example also shows that we now need the
ability to export/import OEFACE and OIDATA, which I added.
Change-Id: Ida0f81ce757fff78fec6276c60052ed71d207454
Reviewed-on: https://go-review.googlesource.com/c/go/+/333014
Run-TryBot: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Dan Scales <danscales@google.com>
The Func.ClosureCalled flag is an optimization used by escape analysis
to detect closures that were directly called, so we know we have
visibility of the result flows. It's not needed by any other phases of
the compiler, so we might as well calculate it within escape analysis
too.
This saves some trouble during IR construction and trying to maintain
the ClosureCalled flag through inlining and copying.
Passes toolstash -cmp.
Change-Id: Ic53cecb7ac439745c0dfba2cd202b9cc40f1e47c
Reviewed-on: https://go-review.googlesource.com/c/go/+/332691
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Trust: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
We often need to create a function/method instantiation, but not a
dictionary, because the call to the instantiation will be using a
sub-dictionary. Also, main dictionaries are only need for concrete,
non-gcshape types, whereas instantiations will be for gcshape types (or
concrete types, for strict stenciling).
Created a helper function getDictOrSubdict() to reduce duplicated code.
Also, moved gfGetGfInfo() call in getDictionarySym() inside conditional
where it is needed, to avoid extra work when dictionary has already been
created.
Change-Id: I06587cb2ddc77de2f991e9f9eaf462d2c5a5d45e
Reviewed-on: https://go-review.googlesource.com/c/go/+/332550
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Added new struct instInfo for information about an instantiation (of a
generic function/method with gcshapes or concrete types). We use this to
remember the dictionary param node, the nodes where sub-dictionaries
need to be used, etc. The instInfo map replaces the Stencil map in
Package.
Added code to access sub-dictionary entries at the appropriate call
sites. We are currently still calculating the corresponding main
dictionary, even when we really only need a sub-dictionary. I'll clean
that up in a follow-up CL.
Added code to deal with "generic" closures (closures that reference some
generic variables/types). We decided that closures will share the same
dictionary as the containing function (accessing the dictionary via a
closure variable). So, the getGfInfo function now traverses all the
nodes of each closure in a function that it is analyzing, so that a
function's dictionary has all the entries needed for all its closures as
well. Also, the instInfo of a closure is largely shared with its
containing function. A good test for generic closures already exists
with orderedmap.go.
Other improvements:
- Only create sub-dictionary entries when the function/method
call/value or closure actually has type params in it. Added new test
file subdict.go with an example where a generic method has an
instantiated method call that does not depend not have type params.
Change-Id: I691b9dc024a89d2305fcf1d8ba8540e53c9d103f
Reviewed-on: https://go-review.googlesource.com/c/go/+/331516
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
For method references (only), selectorExpr() now computes n.Selection,
which is the generic method that is selected. This allows us to compute
as needed the proper sub-dictionary for method reference. Also cleans up
some code for distinguishing method references from references to a
field that has a function value (especially in the presence of embedded
fields).
Change-Id: I9c5b789c15537ff48c70ca7a6444aa0420178a3a
Reviewed-on: https://go-review.googlesource.com/c/go/+/332095
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Created functions to create GC shape type and names, based on a proposal
from Keith. Kept unsigned and signed integer types as different, since
they have different shift operations.
Included adding in alignment fields where padding is
required between fields, even though that seems like it will be fairly
uncommon to use.
Added some extra unusual struct typeparams (for testing the gcshape
names/types) in index.go test.
Change-Id: I8132bbd28098bd933435b8972ac5cc0b39f4c0df
Reviewed-on: https://go-review.googlesource.com/c/go/+/329921
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
This CL handles I(x) where I is an interface type and x has
typeparam type.
Change-Id: Ib99de2b741d588947f5e0164255f6365e98acd8a
Reviewed-on: https://go-review.googlesource.com/c/go/+/326189
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Separate generation of instantiation and dictionary name generation.
Add code to add subdictionaries to a dictionary. Not quite working
yet, as we need to trigger generation of the subdictionaries for methods.
Change-Id: I0d46053eba695b217630b06ef2f990f6a0b52d83
Reviewed-on: https://go-review.googlesource.com/c/go/+/331209
Trust: Keith Randall <khr@golang.org>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
Go spec call them "method values", not "partial calls". Note that
we use "OMETHVALUE" (as opposed to "OMETHODVALUE") to be consistent
with "OMETHEXPR".
Change-Id: I1efd985d4b567a1b4b20aeb603eb82db579edbd5
Reviewed-on: https://go-review.googlesource.com/c/go/+/330837
Trust: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Run-TryBot: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
This is code in progress to generate the two main other types of entries
in dictionaries:
- all types in the instantiated function derived from the type
arguments (which are currently concrete, but will eventually be
gcshapes)
- pointers (i.e. mainly the unique name) to all needed sub-dictionaries
In order to generate these entries, we now generate cached information
gfInfo about generic functions/methods that can be used for creating the
instantiated dictionaries. We use the type substituter to compute the
right type args for instantiated sub-dictionaries.
If infoPrintMode is changed to true, the code prints out all the
information gathered about generic functions, and also the entries in
all the dictionaries that are instantiated. The debug mode also prints
out the locations where we need main dictionaries in non-instantiated
functions.
Other changes:
- Moved the dictionary generation back to stencil.go from reflect.go,
since we need to do extra analysis for the new dictionary entries. In
the process, made getInstantiation generate both the function
instantiation and the associated dictionary.
- Put in small change for now in reflect.go, so that we don't try
generate separate dictionaries for Value[T].get and the
auto-generated (*Value[T]).get. The auto-generated wrapper shouldn't really
need a dictionary.
- Detected, but not handling yet, a new case which needs
dictionaries - closures that have function params or captured
variables whose types are derived from type arguments.
- Added new tests in dictionaryCapture for use of method
value/expressions in generic functions and for mutually recursive
generic functions.
Change-Id: If0cbde8805a9f673a23f5ec798769c85c9c5359b
Reviewed-on: https://go-review.googlesource.com/c/go/+/327311
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
- Fix handling of method expressions with embedded fields. Fix an
incorrect lookup for method expressions, which have only the
top-level type (and don't have DOT operations for the embedded
fields). Add the embedded field dot operations into the closure.
- Don't need a dictionary and so don't build a closure if the last
embedded field reached in a method expression is an interface value.
- Fix methodWrapper() to use the computed 'dot' node in the
generic-only part of the code.
- For a method expression, don't create a generic wrapper if the last
embedded field reached before the method lookup is an interface.
Copied cmd/compile/internal/types2/testdata/fixedbugs/issue44688.go2 to
test/typeparam/issue44688.go, made it fully runnable (rather than just
for compilation), and added a bunch more tests.
Change-Id: I90c1aa569e1c7272e986c9d2ae683e553c3a38a1
Reviewed-on: https://go-review.googlesource.com/c/go/+/329550
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
I initially made NewClosureFunc take an "outerfn *Func" parameter
because I was planning on having it handle closure naming, until
remembering that naming needs to wait until typecheck for noder.
We don't actually need the *Func yet, just to know whether it's
non-nil. So change the parameter to a bool, which simplifies callers a
little.
Change-Id: Ie83ee4a1ed0571ac6d3879ffd8474c6c3c1a9ff9
Reviewed-on: https://go-review.googlesource.com/c/go/+/327450
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
typecheck.tcClosure is complicated with many code flows because all of
its callers setup the closure funcs in slightly different ways. E.g.,
it's non-obvious who's responsible for setting the underlying func's
Sym or adding it to target.Decls, or how to write new code that
constructs a closure without interfering with existing code.
This CL refactors everything to use three common functions in package
ir: NewClosureFunc (which handle creating the Func, Name, and
ClosureExpr and wiring them together), NameClosure (which generates
and assigns its unique Sym), and UseClosure (which handles adding the
Func to target.Decls).
Most IR builders can actually name the closure right away, but the
legacy noder+typecheck path may not yet know the name of the enclosing
function. In particular, for methods declared with aliased receiver
parameters, we need to wait until after typechecking top-level
declarations to know the method's true name. So they're left anonymous
until typecheck.
UseClosure does relatively little work today, but it serves as a
useful spot to check that the code setting up closures got it right.
It may also eventually serve as an optimization point for early
lifting of trivial closures, which may or may not ultimately be
beneficial.
Change-Id: I7da1e93c70d268f575b12d6aaeb2336eb910a6f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/327051
Trust: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
When converting a variable of generic type to an interface, use the
entry in the dictionary for the type field instead of using the
compile-time type (which we only have when fully stenciling).
Note: this isn't all the conversions. Conversions often get processed
in the ir.OCALL case. Those aren't handled yet.
Change-Id: I9a6a4c572e3c54a8e8efad98365184dbb94c4487
Reviewed-on: https://go-review.googlesource.com/c/go/+/325330
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Dan Scales <danscales@google.com>
The instantiated functions are created in the source package of the
generic function, so all lookups of symbols should be relative to that
package, so all symbols are consistently in the source package.
Fixes#46575
Change-Id: Iba67b2ba8014a630c5d4e032c0f2f2fbaaedce65
Reviewed-on: https://go-review.googlesource.com/c/go/+/325529
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
There's no outer function in these cases, so we won't be reading
the dictionary as a subdictionary from the outer scope's dictionary.
It will always be a compile-time constant.
Change-Id: I754b126652a6ffb62255734d53fcec29d77cfa9e
Reviewed-on: https://go-review.googlesource.com/c/go/+/324949
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Reviewed-by: Dan Scales <danscales@google.com>
Change markType to scan generic types and methods, so that inlineable
functions inside generic functions/methods will be properly marked for
export, which means inlining inside instantiated functions will work
correctly.
Also, fix handling of closures for instantiated functions. Some code
needs to be adjusted, since instantiated functions/methods are compiled
as if in the package of the source generic function/type, rather than in
the local package. When we create the closure struct, we want to make
sure that the .F field has the same package as the other fields for the
closure variables. Also, we need to disable a check in tcCompLit() when
being done for an instantiated function, since fields of the closure
struct will be from the source package, not the local package.
Re-enabled part of the orderedmapsimp test that was disabled because of
these issues.
Change-Id: Ic4dba8917da0a36b17c0bdb69d6d6edfdf14104a
Reviewed-on: https://go-review.googlesource.com/c/go/+/324331
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Deal with export/import of recursive generic types. This includes
typeparams which have bounds that reference the typeparam.
There are three main changes:
- Change export/import of typeparams to have an implicit "declaration"
(doDecl). We need to do a declaration of typeparams (via the
typeparam's package and unique name), because it may be referenced
within its bound during its own definition.
- We delay most of the processing of the Instantiate call until we
finish the creation of the top-most type (similar to the way we
delay CheckSize). This is because we can't do the full instantiation
properly until the base type is fully defined (with methods). The
functions delayDoInst() and resumeDoInst() delay and resume the
processing of the instantiations.
- To do the full needed type substitutions for type instantiations
during import, I had to separate out the type subster in stencil.go
and move it to subr.go in the typecheck package. The subster in
stencil.go now does node substitution and makes use of the type
subster to do type substitutions.
Notable other changes:
- In types/builtins.go, put the newly defined typeparam for a union type
(related to use of real/imag, etc.) in the current package, rather
than the builtin package, so exports/imports work properly.
- In types2, allowed NewTypeParam() to be called with a nil bound, and
allow setting the bound later. (Needed to import a typeparam whose
bound refers to the typeparam itself.)
- During import of typeparams in types2 (importer/import.go), we need
to keep an index of the typeparams by their package and unique name
(with id). Use a new map typParamIndex[] for that. Again, this is
needed to deal with typeparams whose bounds refer to the typeparam
itself.
- Added several new tests absdiffimp.go and orderedmapsimp.go. Some of
the orderemapsimp tests are commented out for now, because there are
some issues with closures inside instantiations (relating to unexported
names of closure structs).
- Renamed some typeparams in test value.go to make them all T (to make
typeparam uniqueness is working fine).
Change-Id: Ib47ed9471c19ee8e9fbb34e8506907dad3021e5a
Reviewed-on: https://go-review.googlesource.com/c/go/+/323029
Trust: Dan Scales <danscales@google.com>
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
When converting from a generic function to a concrete implementation,
add a dictionary argument to the generic function (both an actual
argument at each callsite, and a formal argument of each
implementation).
The dictionary argument comes before all other arguments (including
any receiver).
The dictionary argument is checked for validity, but is otherwise unused.
Subsequent CLs will start using the dictionary for, e.g., converting a
value of generic type to interface{}.
Import/export required adding support for LINKSYMOFFSET, which is used
by the dictionary checking code.
Change-Id: I16a7a8d23c7bd6a897e0da87c69f273be9103bd7
Reviewed-on: https://go-review.googlesource.com/c/go/+/323272
Trust: Keith Randall <khr@golang.org>
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Dan Scales <danscales@google.com>
We have a value typecheck(3) that indicates that a node in a generic
function still needs transformation (via the functions in transform.go).
But it is not very desirable to export/import the value of typecheck(3).
So, I changed the stenciling code to just try to transform all relevant
node types during node copy. Almost all tranform functions were already
idempotent. I only had to add an extra if check before calling
transformAssign() in the OAS case. We still use the typecheck(3) in
noder to determine when higher-nodes have to delay transformation
because one or more of their args are delaying transformation.
Added new test mapsimp.go that required these tranformations after import.
As an additional change, export/import of OINDEX requires exporting the
type using w.exoticType() rather than w.typ(), in order to handle
generic functions. Since generic functions can have pre-transform
operations, the index operation can have a tuple type (multiple return
from a map lookup).
Added printing of imported function bodies in -W=3 debug mode.
Change-Id: I220e2428dc5f2741e91db146f075eb5b6045f451
Reviewed-on: https://go-review.googlesource.com/c/go/+/322191
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
The name substitution for stenciling was incorrectly handling non-local
names. Made changes to explicitly built the vars[] name substitution map
based on the local variables (similar to what inlining substitution
does). Then, we we are stenciling a name node, we do NOT make a copy of
the name node if it is not in vars[], since it is then a reference to an
external name. Added new function localvar() to create the new nodes for
the local variables and put them in the vars[] map.
New test listimp2.go, added missing test calls in list2.go
Change-Id: I8946478250c7bf2bd31c3247089bd50cfeeda0fd
Reviewed-on: https://go-review.googlesource.com/c/go/+/322190
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
