Grinding a small figdraw-based program under `nim ic` and diffing its
output against the classic backend surfaced eight bugs, four of which
silently produced a wrong binary rather than an error.
Frontend / build graph (`deps.nim`):
* Dead `when`-guarded imports were compiled anyway. `when someStrdefine
== "x": import y` is `cvUnknown` to the scanner, which conservatively
keeps the edge — right for an edge, but it also gave `y` its own `nim m`
rule, so a build died on a package the user never installed because they
never selected that backend. Track which edges are speculative and drop
a speculative subtree that cannot compile; if the guard was in fact
live, the discovery fixpoint puts the node back with the honest `cannot
open file`.
* Deleting a still-imported module went unnoticed: no mtime moves, so
nothing re-fires and `nim ic` relinked a stale binary while `nim c`
reported `cannot open file`. Report an unresolvable import from a
non-speculatively reached module during the graph scan.
* Macro-generated imports were discovered once and then forgotten.
Discovery only ran after a failure and the graph is re-derived
statically every run, so on a warm build the discovered module had no
rules at all and editing it changed nothing. Seed the graph from the
`.s.deps` sidecars up front.
* Config changes invalidated nothing. nifmake decides staleness from
file mtimes and never looks at a rule's command line, so `-d:foo=bar` /
`--mm:` / `--threads:` regenerated the build file with the new switches
and re-fired zero rules. Reify the configuration as a file and make it
an input of every rule.
* Command-line switches never reached the children: they replay the
project's config files, never the driver's argv, so `nim ic --opt:speed`
produced a byte-identical debug binary (likewise `--panics`,
`--experimental`, `--passC`). Forward the driver's switches, minus the
ones that must differ per child.
Artifacts and codegen:
* A failed `nim m` still wrote its `.s.bif` and cookies, so nifmake saw
the rule as satisfied on the next run: `nim ic` then reported success
for a program that does not compile, and generated code from
error-bearing AST (or hit an internal error in `ccgexprs`). Never
persist an artifact when `errorCounter > 0`.
* Top-level destructors were never injected. `sfInjectDestructors` lives
on the module symbol, which `moduleFromNifFile` rebuilds from scratch,
so `genTopLevelStmt` skipped `injectDestructorCalls` entirely: a
module-level `block: let h = openHandle()` never ran `=destroy`. Persist
the flag as a `(modflags)` record. `injectdestructors` also has to
tolerate the `nkReplayAction` entries the loader prepends to `topLevel`.
* `nfFirstWrite` / `nfLastRead` were dropped by the serializer. A sym
node is written as a bare NIF `SymUse` token, which has nowhere to put
node flags, so the frontend's move analysis never reached the backend:
EVERY first assignment to a destructor-bearing local compiled as
`=sink`, i.e. `=destroy` on still-zeroed memory followed by a copy, and
no read was ever a move. Wrap a sym use in `(nflags ...)` when it
carries persistent node flags.
fixes#25803
This pull request introduces a new approach for generating C++
constructor expressions in the Nim compiler's C++ backend, ensuring that
type-prefixed construction is used when needed (such as in assignments),
rather than just braced initializer lists. It also adds a new test case
(with corresponding C++ header) to verify correct behavior when
assigning to types with overloaded assignment operators and
constructors.
**C++ code generation improvements:**
* Added `genCppConstructorExpr` in `ccgtypes.nim`, which generates a
full type-prefixed constructor expression (e.g., `Foo(a, b)`) for C++
code generation, as opposed to just a braced initializer list. This is
important for contexts like assignments where the type must be explicit.
* Updated `resetLoc` in `cgen.nim` to use `genCppConstructorExpr`
instead of `genCppInitializer` when initializing imported C++ types,
ensuring correct code generation for assignments.
**Testing:**
* Added a new C++ header file, `tcpp_default_ctor_assignment.h`,
defining a struct `AmbiguousAssign` with overloaded assignment operators
and constructors to test ambiguous assignment scenarios.
* Added a corresponding Nim test, `tcpp_default_ctor_assignment.nim`,
which exercises construction and assignment for the imported C++ type,
ensuring the new code generation logic works as intended.
First performance numbers:
time tests/arc/torcbench -- YRC
true peak memory: true
real 0m0,163s
user 0m0,161s
sys 0m0,002s
time tests/arc/torcbench -- ORC
true peak memory: true
real 0m0,107s
user 0m0,104s
sys 0m0,003s
So it's 1.6x slower. But it's threadsafe and provably correct. (Lean and
model checking via TLA+ used.)
Of course there is always the chance that the implementation is wrong
and doesn't match the model.
`hashType` proc returned the same hash value from different instanced
generics types like `D[int64]` and `D[F]`.
That caused the struct type with wrong field types.
object/tuple type size check code is generated when it is compiled with
`-d:checkAbi` option.
TODO:
- [ ] test writing of .nif files
- [x] implement loading of fields in PType/PSym that might not have been
loaded
- [ ] implement interface logic
- [ ] implement pragma "replays"
- [ ] implement special logic for `converter`
- [ ] implement special logic for `method`
- [ ] test the logic holds up for `export`
- [ ] implement logic to free the memory of PSym/PType if memory
pressure is high
- [ ] implement logic to close memory mapped files if too many are open.
---------
Co-authored-by: demotomohiro <gpuppur@gmail.com>
Co-authored-by: ringabout <43030857+ringabout@users.noreply.github.com>
Co-authored-by: Jacek Sieka <arnetheduck@gmail.com>
fixes#24847
Object constructors call `fillObjectFields` when a field inside the
constructor does not have a location, however when the field is from a
base type this does not process it. Now `fillObjectFields` also calls
itself for the base type to fix this but not sure if this is a good
solution as `fillObjectFields` is used in other places too.
fixes#24705
```nim
proc xxx(v: static int) =
echo v
xxx(10)
xxx(20)
```
They are mangled as `_ZN14titaniummangle7xxx_s10E` and
`_ZN14titaniummangle7xxx_s20E` with `--debugger:native`. Static
parameters are prefixed with `_s` to distinguish simple cases like
`xxx(10, 15)` and `xxx(101, 5)` if `xxx` supports two `static[int]`
parameters
fixes#24725
`lacksMTypeField` doesn't take the base types into consideration. And
for ` {.inheritable, pure.}`, it shouldn't generate a `m_type` field.
The lower half of cgen contains the main proc and HCR init code which
cause a large diff, so they are excluded from this PR. In general things
like generated defines, line directives, the stacktrace macros (`nimfr_`
etc) are also not done, since there are not exact equivalents for these
in NIFC (NIFC does generate line directives but based on the NIF line
info mechanism).
This adapts most of ccgstmts to cbuilder, missing C++ exceptions and
`genCaseGeneric`/`genIfForCaseUntil`. Rebased version of #24399 which
was split into some other PRs and since has had conflicts with #24416.
Finishes `genEnumInfo` as followup to #24351. As #24381 mentions this
covers every use of `for` loops in the codegen.
---------
Co-authored-by: Andreas Rumpf <rumpf_a@web.de>
Doing this early is useful so we can move the indentation logic into
`Builder` itself rather than mix it with the block logic in `ccgstmts`
(the `if` statements in #24381 have not been indented properly either).
However it also means `Builder` is now used for code that still
generates raw C code, so the diff won't be as clean when these get
updated.
C++ member procs are not implemented, and `codegenDecl` in general is
also not adapted, although it had to be included in the generation of
proc params for simplicity. Guessing `codegenDecl` and C++ stuff are
supposed to map to pragmas on NIFC, or are just not supported.
Most of what ccgexprs uses is now ported to cbuilder, so this PR makes
around ~25% of ccgexprs use it, along with adding `if` stmts (no
`while`/`switch` and `for` which is only used as `for (tmp = a; tmp < b;
tmp++)`). The `if` builder does not add indents for blocks since we
can't make `Builder` an object yet rather than an alias to `string`,
this will likely be one of the last refactors.
Somewhat unrelated but `ccgtypes` is not ready yet because proc
signatures are not implemented.
The types of the typed operations are not necessarily correct, the mixed
types of the operands in some cases are left the same to keep the C
integer promotion working as expected. To fix any wrong types the
integer promotion can be done explicitly instead, but this is a behavior
change to codegen.
Different from `intLiteral`, we add procs that just try to generate
integer values, assuming they're not an edge case like `<= low(int32)`
or `> high(int32)`.
This PR is somewhat large, worst case it can be split into one with just
assignments and one with just fields/derefs etc.
Assignments with calls as values have not been touched so they can be
done when calls are implemented. Similarly codegen with complex logic
i.e. `genEnumInfo`, `genTypeInfoV2`, `unaryExpr` is not completely
ported yet so they can be done in standalone PRs.
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.
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
`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.
