[Issue 24378] New: [REG 2.104] inout error with -profile=gc
https://issues.dlang.org/show_bug.cgi?id=24378
Issue ID: 24378
Summary: [REG 2.104] inout error with -profile=gc
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: regression
Priority: P1
Component: druntime
Assignee: nob...@puremagic.com
Reporter: tim.dl...@t-online.de
```
inout(char)[] foo(inout(char)[] s) pure @safe
{
s = s[0..1] ~ '%';
return s;
}
```
The above code compiled with and without -profile=gc until DMD version 2.103.
Since DMD 2.104 it only compiles without -profile=gc and produces the following
error with -profile=gc:
/path/to/dmd.linux/dmd2/linux/bin64/../../src/druntime/import/core/lifetime.d(1573):
Error: `inout` on `return` means `inout` must be on a parameter as well for
`pure nothrow @nogc @property @safe inout(char)[]()`
/path/to/dmd.linux/dmd2/linux/bin64/../../src/druntime/import/core/internal/array/concatenation.d(174):
Error: template instance
`core.internal.array.concatenation._d_arraycatnTXTrace!(inout(char)[],
inout(char)[], inout(char))._d_arraycatnTXTrace.forward!(__param_3, __param_4)`
error instantiating
onlineapp.d(3):instantiated from here:
`_d_arraycatnTXTrace!(inout(char)[], inout(char)[], inout(char))`
As a result of this issue, compiling DMD itself with -profile=gc does not work
any more.
--
[Issue 24377] New: Error: negative array dimension `3145728u * 1024u`[32bit]
https://issues.dlang.org/show_bug.cgi?id=24377
Issue ID: 24377
Summary: Error: negative array dimension `3145728u *
1024u`[32bit]
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: kde...@vogtner.de
$ cat nad.d
import std.stdio;
void main()
{
auto arr = new char [3u * 1024 * 1024 * 1024]; // <--- error
writefln ("%X", arr.length);
}
$ dmd -m32 ./nad.d
./nad.d(5): Error: negative array dimension `3145728u * 1024u` [unexpected]
$ gdc -m32 nad.d -o nad
$ ./nad
C000 [expected]
$ cat nad2.d
import std.stdio;
void main()
{
auto siz = 3u * 1024 * 1024 * 1024;
auto arr = new char [siz];
writefln ("%X", arr.length);
}
$ dmd -m32 ./nad2.d
$ ./nad2
C000 [expected]
used compilers:
- DMD64 D Compiler v2.105.3
- gcc (GCC) 12.1.0
--
[Issue 24376] New: ImportC: .di generator outputs D keywords for members
https://issues.dlang.org/show_bug.cgi?id=24376
Issue ID: 24376
Summary: ImportC: .di generator outputs D keywords for members
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Keywords: ImportC
Severity: enhancement
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: dave287...@gmail.com
This is already noted as a limitation of ImportC (the mismatch in keywords) and
in D code you can work around it using __traits. However, in the .di generator
you instead get a parse error:
C code:
struct foo {
int version;
const char* function;
int uint;
};
.di file:
// ...
struct foo
{
int version = void;
const(char)* function = void;
int uint = void;
}
// ...
Trying to import the .di file will then fail.
I think ideally you solve this with adding a mechanism for D identifiers to
have the same name as keywords. A new kind of string?
Another way to go is to have the compiler rename identifiers that are D
keywords when importing C code (say by appending an underscore). Could also
only do that in the .di generation, but then importing the C code and importing
the .di file are different.
--
[Issue 24375] New: ImportC: .di generator outputs C expression with `->` operator
https://issues.dlang.org/show_bug.cgi?id=24375
Issue ID: 24375
Summary: ImportC: .di generator outputs C expression with `->`
operator
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Keywords: ImportC
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: dave287...@gmail.com
C code:
struct foo {
int x;
};
typedef int weird[sizeof(((struct foo *)((void*)0))->x)];
outputs:
extern(C)
{
// ...
struct foo
{
int x = void;
}
alias weird = int[(cast(foo*)cast(void*)0)->x.sizeof];
// ...
}
Encountered something like this in an SDL header.
--
[Issue 24374] New: ImportC: .di generator incorrect output for anonymous structs as members
https://issues.dlang.org/show_bug.cgi?id=24374
Issue ID: 24374
Summary: ImportC: .di generator incorrect output for anonymous
structs as members
Product: D
Version: D2
Hardware: x86
OS: Mac OS X
Status: NEW
Keywords: ImportC
Severity: enhancement
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: dave287...@gmail.com
C file:
union foo {
struct { int x; } X;
};
struct bar {
struct { int x; } X;
};
Results in following in .di file:
extern(C)
{
// ...
union foo
{
struct X = void;
}
struct bar
{
struct X = void;
}
// ...
}
Which is not valid D code.
--
[Issue 24374] ImportC: .di generator incorrect output for anonymous structs as members
https://issues.dlang.org/show_bug.cgi?id=24374 dave287...@gmail.com changed: What|Removed |Added Hardware|x86 |All OS|Mac OS X|All Severity|enhancement |normal --
[Issue 24373] New: ImportC: .di generator can output invalid aliases
https://issues.dlang.org/show_bug.cgi?id=24373
Issue ID: 24373
Summary: ImportC: .di generator can output invalid aliases
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Keywords: ImportC
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: dave287...@gmail.com
Example c file:
typedef unsigned int uint;
typedef unsigned short ushort;
If you then convert to a .di file, you get something like:
extern(C)
{
// ... bunch of junk
alias uint = uint;
alias ushort = ushort;
// ...
}
Which are not valid alias declarations.
--
