The only remaining explicit use of `typedef` in the codegen (from my
search) is in `addForwardStructFormat` which from what I understand
won't do anything in NIFC.
Running `ctags` on files with quoted symbols (e.g. `$`) would list \`
instead of the full ident. Issue was the result getting reassigned at
the end to a \` instead of appending
closes https://github.com/nim-lang/RFCs/issues/554
Adds a symmetric difference operation to the language bitset type. This
maps to a simple `xor` operation on the backend and thus is likely
faster than the current alternatives, namely `(a - b) + (b - a)` or `a +
b - a * b`. The compiler VM implementation of bitsets already
implemented this via `symdiffSets` but it was never used.
The standalone binary operation is added to `setutils`, named
`symmetricDifference` in line with [hash
sets](https://nim-lang.org/docs/sets.html#symmetricDifference%2CHashSet%5BA%5D%2CHashSet%5BA%5D).
An operator version `-+-` and an in-place version like `toggle` as
described in the RFC are also added, implemented as trivial sugar.
refs #8064, refs #24010
Error messages for standalone explicit generic instantiations are
revamped. Failing standalone explicit generic instantiations now only
error after overloading has finished and resolved to the default `[]`
magic (this means `[]` can now be overloaded for procs but this isn't
necessarily intentional, in #24010 it was documented that it isn't
possible). The error messages for failed instantiations are also no
longer a simple `cannot instantiate: foo` message, instead they now give
the same type mismatch error message as overloads with mismatching
explicit generic parameters.
This is now possible due to the changes in #24010 that delegate all
explicit generic proc instantiations to overload resolution. Old code
that worked around this is now removed. `maybeInstantiateGeneric` could
maybe also be removed in favor of just `explicitGenericSym`, the `result
== n` case is due to `explicitGenericInstError` which is only for niche
cases.
Also, to cause "ambiguous identifier" error messages when the explicit
instantiation is a symchoice and the expression context doesn't allow
symchoices, we semcheck the sym/symchoice created by
`explicitGenericSym` with the given expression flags.
#8064 isn't entirely fixed because the error message is still misleading
for the original case which does `values[1]`, as a consequence of
#12664.
The remaining followup from #24259. A body for building the type doesn't
seem necessary here since the types with array fields are generally
atomic/already built from `getTypeDescAux`.
The first commit reverts the revert of #23787.
The second fixes lambdalifting in convolutedly nested
closures/closureiters. This is considered to be the reason of #24094,
though I can't tell for sure, as I was not able to reproduce #24094 for
complicated but irrelevant reasons. Therefore I ask @jmgomez, @metagn or
anyone who could reproduce it to try it again with this PR.
I would suggest this PR to not be squashed if possible, as the history
is already messy enough.
Some theory behind the lambdalifting fix:
- A closureiter that captures anything outside its body will always have
`:up` in its env. This property is now used as a trigger to lift any
proc that captures such a closureiter.
- Instantiating a closureiter involves filling out its `:up`, which was
previously done incorrectly. The fixed algorithm is to use "current" env
if it is the owner of the iter declaration, or traverse through `:up`s
of env param until the common ancestor is found.
---------
Co-authored-by: Andreas Rumpf <rumpf_a@web.de>
fixes#24305, refs #23807
Since #23014 `nkHiddenAddr` is produced to fast assign array elements in
iterators. However the array access inside this `nkHiddenAddr` can get
folded at compile time, generating invalid code. In #23807, compile time
folding of regular `addr` expressions was changed to be prevented in
`transf` but `nkHiddenAddr` was not updated alongside it.
The method for preventing folding in `addr` in #23807 was also faulty,
it should only trigger on the immediate child node of the address rather
than all nodes nested inside it. This caused a regression as outlined in
[this
comment](https://github.com/nim-lang/Nim/pull/24322#issuecomment-2419560182).
To fix both issues, `addr` and `nkHiddenAddr` now both shallowly prevent
constant folding for their immediate children.
fixes#24319
`byRefLoc` (`mapType`) requires the Loc `a` to have the right type.
Without `lfEnforceDeref`, it produces the wrong type for `deref (var
array)`, which may come from `mitems`.
fixes#18896fixes#20886
```nim
type
PFile {.importc: "FILE*", header: "<stdio.h>".} = distinct pointer
# import C's FILE* type; Nim will treat it as a new pointer type
```
This is an excerpt from the Nim manual. In the old Nim versions, it
produces a void pointer type instead of the `FILE*` type that should
have been generated. Because these C types tend to be opaque and adapt
to changes on different platforms. It might affect the portability of
Nim on various OS, i.e. `csource_v2` cannot build on the apline platform
because of `Time` relies on Nim types instead of the `time_t` type.
ref https://github.com/nim-lang/Nim/pull/18851
follows up #24259
This is the remaining missing use of `StructInitializer` in
`getDefaultValue` after #24259 and #24302. The only remaining direct C
code in getDefaultValue is [this
line](922f7dfd71/compiler/ccgexprs.nim (L3525))
which creates a global array variable, which isn't implemented yet. Next
steps would be all remaining variable and `typedef` declarations, then
hopefully we can move on to general statements and expressions.
follows up #24259
This was the only use of the `STRING_LITERAL` macro in `nimbase.h`, so
this macro is now removed. We don't have to remove it though, maybe
people use it.
`StructInitializer` is now used for most braced initializers in the C
generation, mostly in `genBracedInit`, `getNullValueAux`,
`getDefaultValue`. The exceptions are:
* the default case branch initializer for objects uses C99 designated
initializers with field names, which are not implemented for
`StructInitializer` yet (`siNamedStruct`)
* the uses in `ccgliterals` are untouched so all of ccgliterals can be
done separately and in 1 go
There is one case where `genBracedInit` does not use cbuilder, which is
the global literal variable for openarrays. The reason for this is
simply that variables with C array type are not implemented, which I
thought would be best to leave out of this PR.
For the simplicity of the implementation, code in `getNullValueAuxT`
that reset the initializer back to its initial state if the `Sup` field
did not have any fields itself, is now disabled. This was so the
compiler does not generate `{}` for the Sup field, i.e. `{{}}`, but
every call to `getNullValueAuxT` still generates `{}` if the struct
doesn't have any fields, so I don't know if it really breaks anything.
The case where the Sup field doesn't have any fields but the struct does
also would have generated `{{}, field}`.
Worst case, we can implement either the "resetting" or just disable the
generation of the `Sup` field if there are no fields total. But a better
fix might be to always generate `{0}` if the struct has no fields, in
line with the `char dummy` field that gets added for all objects with no
fields. This doesn't seem necessary for now but might be for the NIFC
output, in which case we can probably keep the logic contained inside
cbuilder (if no fields generated for `siOrderedStruct`/`siNamedStruct`,
we add a `0` for the `dummy` field). This would stipulate that all uses
of struct initializers are exhaustive for every field in structs.
fixes#24296fixes#24295
Templates use `expectedType` for type inference. It's justified that
when templates don't have an actual return type, i.e., `untyped` etc.
When the return type of templates is specified, we should not infer the
type
```nim
template g(): string = ""
let c: cstring = g()
```
In this example, it is not reasonable to annotate the templates
expression with the `cstring` type before the `fitNode` check with its
specified return type.
fixes#24288, refs #23625
Since #23625 "cannot evaluate" errors during VM code generation are
"soft" errors with `nim check`, i.e. the code generation isn't halted
(except for the current proc which `return`s which can cause wrong
codegen) and the expression is still attempted to be evaluated. Now,
these errors signal to the VM that the current generated VM code cannot
be evaluated, and so instead of evaluating, an error node is returned.
This keeps the benefit of the "soft" errors without potentially crashing
the compiler on improperly generated VM code. Although maybe the
compiler might not be able to handle the generated error node in some
cases.
This fixes the chame example in #24288 but this is not tested in CI.
Presumably it or the compiler was doing something like `compiles()` on
code that can't run in the VM.
I would accept nicer ways of tracking non-evaluability than
`c.cannotEval = true` but I tried to keep it as harmless as possible.
fixes#24021
The field checking for case object branches at some point generates a
negated set `contains` check for the object discriminator. For enum
types, this tries to generate a complement set and convert to a
`contains` check in that instead. It obtains this type from the type of
the element node in the `contains` check.
`buildProperFieldCheck` creates the element node by changing a field
access expression like `foo.z` into `foo.kind`. In order to do this, it
copies the node `foo.z` and sets the field name in the node to the
symbol `kind`. But when copying the node, the type of the original
`foo.z` is retained. This means that the complement is performed on the
type of the accessed field rather than the type of the discriminator,
which causes problems when the accessed field is also an enum.
To fix this, we properly set the type of the copied node to the type of
the kind field. An alternative is just to make a new node instead.
A lot of text for a single line change, I know, but this part of the
codebase could use more explanation.
fixes#24087, refs https://forum.nim-lang.org/t/341, refs #14222, refs
#14221
The Nim compiler calls `cl` for linking as well as compilation. This
means that options to the linker have to be passed after a `/link`
argument. But the Nim compiler doesn't include this option normally,
because users may still want to pass non-linker options to `cl` at link
time.
To deal with this, a workaround is used: every single library link
option adds `/link` before it. The linker simply ignores extraneous
`/link` arguments and gives a warning instead, since it's an
unrecognized option to the linker. This is really hacky but otherwise we
need to separate linker arguments into arguments passed either to the
compiler or to the linker at link time, and this behavior wouldn't be
meaningful outside of MSVC.
I can't really test this manually but I did test that the linker ignores
`/link`. I also can't really do more than this, I don't really use MSVC
so I wouldn't know how to navigate it, or how people use it. Ideally
someone who knows more about/uses MSVC can give their input or take
over.
fixes#24269, refs #20095
Instead of checking the package of the *used sym* to determine whether a
stylecheck should trigger, we check the package of the lineinfo instead.
Before #20095 this checked for the current compilation context module
instead which caused issues with generic procs, but the lineinfo should
more closely match the AST.
I figured this might cause issues with includes etc but the foreign
package test specifically tests for an include and passes, so maybe the
package determining logic accounts for this already. This still might
not be the correct logic, I'm not too familiar with the package handling
in the compiler.
Package PRs, both merged:
- json_rpc: https://github.com/status-im/nim-json-rpc/pull/226
- json_serialization:
https://github.com/status-im/nim-json-serialization/pull/99
fixes#24274
The code in the `if` branch replaces the current destination `d` with a
new one. But the location `d` can be an assignment location, in which
case the provided expression isn't generated. To fix this, don't trigger
this code for when the location already exists. An alternative would be
to call `putIntoDest` in this case as is done below.
fixes#24091, refs #24092
Any instantiations resolving to a generic body type now gives an error.
Due to #24092, this does not error in cases like matching against `type
M` in generics because generic body type symbols are just not
instantiated. But this prevents parameters with type `type M` from being
used, although there doesn't seem to be any code which does this. Just
in case such code exists, we still allow `typedesc` types resolving to
generic body types.
fixes#23587
As explained in the issue, `getOrDefault` has a parameter named
`default` that can be a proc after generic instantiation. But the
parameter having a proc type [overrides all other
overloads](f73e03b132/compiler/semexprs.nim (L1203))
including the magic `system.default` overload and causes a compile error
if the proc doesn't match the normal use of `default`. To fix this, the
`result = default(B)` initializer call is removed because it's not
needed, `result` is always set in `getOrDefaultImpl` when a default
value is provided.
This is still a suspicious behavior of the compiler but `tables` working
has a higher priority.
fixes#18649, refs #24183
Same as in #24183 for templates, we now process pragma nodes in generics
so that macro symbols are captured and the pragma arguments are checked,
but ignoring language pragma keywords.
A difference is that we cannot process call nodes as is, we have to
process their children individually so that the early untyped
macro/template instantiation in generics does not kick in.
The only remaining use of `struct` after this is in
`genConstSeq`/`genConstSeqV2` which use `genBracedInit`, I figured these
should be done in the PR that adapts `genBracedInit` in general to
cbuilder.
fixes#24233
Integer literals with type `int` can match `int64` with a generic match.
Normally this would generate an conversion via `isFromIntLit`, but when
it matches with a generic match (`isGeneric`) the node is left alone and
continues to have type `int` (related to #4858, but separate; since
`isFromIntLit > isGeneric` it doesn't propagate). This did not cause
problems on the C backend up to this point because either the compiler
generated a cast when generating the C code or it was implicitly casted
in the C code itself. On the JS backend however, we need to generate
`int64` and `int` values differently, so we copy the integer literal and
give it the matched type now instead.
This is somewhat risky even if CI passes but it's required to make the
times module work without [this
workaround](7dfadb8b4e/lib/pure/times.nim (L219-L238))
on `--jsbigint64:on` (the default).
CI exposed an issue: When matching an int literal to a generic parameter
in a generic instantiation, the literal is only treated like a value if
it has `int literal` type, but if it has the type `int`, it gets
transformed into literally the type `int` (#12664, #13906), which breaks
the tests t14193 and t12938. To deal with this, we don't give it the
type `int` if we are in a generic instantiation and preserve the `int
literal` type.
fixes#24186
When encountering pragma nodes in templates, if it's a language pragma,
we don't process the name, and only any values if they exist. If it's
not a language pragma, we process the full node. Previously only the
values of colon expressions were processed.
To make this simpler, `whichPragma` is patched to consider bracketed
hint/warning etc pragmas like `{.hint[HintName]: off.}` as being a
pragma of kind `wHint` rather than an invalid pragma which would have to
be checked separately. From looking at the uses of `whichPragma` this
doesn't seem like it would cause problems.
Generics have [the same
problem](a27542195c/compiler/semgnrc.nim (L619))
(causing #18649), but to make it work we need to make sure the
templates/macros don't get evaluated or get evaluated correctly (i.e.
passing the proc node as the final argument), either with #23094 or by
completely disabling template/macro evaluation when processing the
pragma node, which would also cover `{.pragma.}` templates.
fixes#24228, refs #22022
As described in
https://github.com/nim-lang/Nim/issues/24228#issuecomment-2392462221,
instantiating generic routines inside `typeof` causes all code inside to
be treated as being in a typeof context, and thus preventing compile
time proc folding, causing issues when code is generated for the
instantiated routine. Now, instantiated generic procs are treated as
never being inside a `typeof` context.
This is probably an arbitrary special case and more issues with the
`typeof` behavior from #22022 are likely. Ideally this behavior would be
removed but it's necessary to accomodate the current [proc `declval` in
the package `stew`](https://github.com/status-im/nim-stew/pull/190), at
least without changes to `compileTime` that would either break other
code (making it not eagerly fold by default) or still require a change
in stew (adding an option to disable the eager folding).
Alternatively we could also make the eager folding opt-in only for
generic compileTime procs so that #22022 breaks nothing whatsoever, but
a universal solution would be better. Edit: Done in #24230 via
experimental switch
`cbuilder` is now split into `cbuilderbase`, `cbuilderexprs`,
`cbuilderdecls`, with all the struct builder code up to this point going
in `cbuilderdecls`.
Variable builders are added, with local, global and constant variables
implemented, but not threadvars.
A builder for struct (braced) initializers is added. The field names
have to be passed to build each field (so they can be used in `oconstr`
in nifc), but they're not used in the output code if a flag
`orderCompliant` is enabled, which means the initializer list is
generated in order of the built fields. The version which uses the names
on C is not implemented (C99 designated initializers), so this flag has
to be enabled for now.
The struct builders now generate the struct as an inline expression if a
name isn't provided rather than a statement. This means we can now use
`addSimpleStruct` etc for the type of fields, but we can't replace
`addFieldWithStructType` because of `#pragma pack(pop)`.
Doc comments are added to every usable proc but may still not be
sufficient.
split from #24198
This is a required refactor for the only good solution I've been able to
think of for #4858 etc. Explanation:
---
`sigmatch` currently [disables
bindings](d6a71a1067/compiler/sigmatch.nim (L1956))
(except for binding to other generic parameters) when matching against
constraints of generic parameters. This is so when the constraint is a
general metatype like `seq`, the type matching will not treat all
following uses of `seq` as the type matched against that generic
parameter.
However to solve #4858 etc we need to bind `or` types with a conversion
match to the type they are supposed to be converted to (i.e. matching
`int literal(123)` against `int8 | int16` should bind `int8`[^1], not
`int`). The generic parameter constraint binding needs some way to keep
track of this so that matching `int literal(123)` against `T: int8 |
int16` also binds `T` to `int8`[^1].
The only good way to do this IMO is to generate a new "binding context"
when matching against constraints, then binding the generic param to
what the constraint was bound to in that context (in #24198 this is
restricted to just `or` types & concrete types with convertible matches,
it doesn't work in general).
---
`semtypinst` already does something similar for bindings of generic
invocations using `LayeredIdTable`, so `LayeredIdTable` is now split
into its own module and used in `sigmatch` for type bindings as well,
rather than a single-layer `TypeMapping`. Other modules which act on
`sigmatch`'s binding map are also updated to use this type instead.
The type is also made into an `object` type rather than a `ref object`
to reduce the pointer indirection when embedding it inside
`TCandidate`/`TReplTypeVars`, but only on arc/orc since there are some
weird aliasing bugs on refc/markAndSweep that cause a segfault when
setting a layer to its previous layer. If we want we can also just
remove the conditional compilation altogether and always use `ref
object` at the cost of some performance.
[^1]: `int8` binding here and not `int16` might seem weird, since they
match equally well. But we need to resolve the ambiguity here, in #24012
I tested disallowing ambiguities like this and it broke many packages
that tries to match int literals to things like `int16 | uint16` or
`int8 | int16`. Instead of making these packages stop working I think
it's better we resolve the ambiguity with a rule like "the earliest `or`
branch with the best match, matches". This is the rule used in #24198.
fixes#18396, fixes#20142
Set types with base types matching less than a generic match (so
subrange matches, conversion matches, int conversion matches) are now
considered mismatching, as their representation is different on the
backends (except VM and JS), causing codegen issues. An exception is
granted for set literal types, which now implicitly convert each element
to the matched base type, so things like `s == {'a', 'b'}` are still
possible where `s` is `set[range['a'..'z']]`. Also every conversion
match in this case is unified under the normal "conversion" match, so a
literal doesn't match one set type better than the other, unless it's
equal.
However `{'a', 'b'} == s` or `{'a', 'b'} - s` etc is now not possible.
when it used to work in the VM. So this is somewhat breaking, and needs
a changelog entry.
Previously, the compiler never differentiated between `untyped`/`typed`
argument default values and other types, it considered any parameter
with a type as typed and called `semExprWithType`, which both
typechecked it and disallowed `void` expressions. Now, we perform no
typechecking at all on `untyped` template param default values, and call
`semExpr` instead for `typed` params, which allows expressions with
`void` type.
fixes#23010, split from #24195
When resemming bracket nodes, the compiler currently unconditionally
makes a new node with an array type based on the node. However the VM
can generate bracket nodes with `seq` types, which this erases. To fix
this, if a bracket node already has a type, we still resem the bracket
node, but don't construct a new type for it, instead using the type of
the original node.
A version of this was rejected that didn't resem the node at all if it
was typed, but I can't find it. The difference with this one is that the
individual elements are still resemmed.
This should fix the break caused by #24184 so we could redo it after
this PR but it might still have issues, not to mention the related
pre-existing issues like #22793, #12559 etc.
`addSimpleStruct` is just so the compiler doesn't use so much extra
computation on analyzing the `typ` parameter for `addStruct`, which
doesn't change anything for `seq` types. We could probably still get
away with using `addStruct` instead, or making `addStruct` accept `nil`
as the `typ` argument but this would be even more computation.
There were a lot of hidden issues with `addStruct` being a template &
template argument substitution, so most of the behavior is moved into
`startStruct`/`finishStruct` procs.
This is turning out to be a lot of code for just a couple of changed
lines, we might have to split `cbuilder` into multiple modules.
