[Issue 4867] [GSoC] Postblit is not usable with const objects
https://issues.dlang.org/show_bug.cgi?id=4867 RazvanN changed: What|Removed |Added Status|NEW |RESOLVED CC||razvan.nitu1...@gmail.com Resolution|--- |WONTFIX --- Comment #4 from RazvanN --- We have copy constructor now so this is not going to get fix. --
[Issue 20365] Copy constructor not invoked on static arrays of structs but the postblit works
https://issues.dlang.org/show_bug.cgi?id=20365 RazvanN changed: What|Removed |Added CC||razvan.nitu1...@gmail.com --- Comment #3 from RazvanN --- This can be nicely solved with the new hasCopyConstructor/hasPostblit traits [1]. I will look into it. [1] https://github.com/dlang/dmd/pull/10528 --
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358 --- Comment #5 from Max Samukha --- I disagree. Implicit constructors should not magically give public assess to private members. That breaks encapsulation. I bet you wouldn't like it if the compiler implicitly generated public setters for private fields. Mutating private fields with an external initializer is analogous to that. --
[Issue 18750] [Tracker] everything wrong with code generation for bt instruction
https://issues.dlang.org/show_bug.cgi?id=18750 ag0aep6g changed: What|Removed |Added Depends on||20363 Referenced Issues: https://issues.dlang.org/show_bug.cgi?id=20363 [Issue 20363] optimization with union, >> and ulong leads to wrong result --
[Issue 20363] optimization with union, >> and ulong leads to wrong result
https://issues.dlang.org/show_bug.cgi?id=20363
ag0aep6g changed:
What|Removed |Added
Keywords||wrong-code
CC||ag0ae...@gmail.com
Blocks||18750
--- Comment #1 from ag0aep6g ---
Seems to be (yet another) bug related to the `bt` instruction.
Smaller test case:
ulong test(double d)
{
ulong u = * cast(ulong*)
return (u >> 1) & 1;
}
void main()
{
ulong u = 0b10;
if (test(* cast(double*) ) == 0) assert(false); /* fails with -O */
}
The generated code for `test`:
0: 48 0f ba e0 01 bt rax,0x1
5: 19 c0 sbbeax,eax
7: f7 d8 negeax
9: c3 ret
It misses code that fills `rax`.
Referenced Issues:
https://issues.dlang.org/show_bug.cgi?id=18750
[Issue 18750] [Tracker] everything wrong with code generation for bt
instruction
--
[Issue 20365] Copy constructor not invoked on static arrays of structs but the postblit works
https://issues.dlang.org/show_bug.cgi?id=20365 Suleyman Sahmi (سليمان السهمي) changed: What|Removed |Added CC||sahmi.soulaim...@gmail.com --
[Issue 20365] Copy constructor not invoked on static arrays of structs but the postblit works
https://issues.dlang.org/show_bug.cgi?id=20365 --- Comment #2 from Suleyman Sahmi (سليمان السهمي) --- There is a whole jungle of pustblit calls in druntime that do not call the copy constructor yet. Just grep the word "postblit" in druntime source to get an idea. I contacted Razvan about it. --
[Issue 20365] Copy constructor not invoked on static arrays of structs but the postblit works
https://issues.dlang.org/show_bug.cgi?id=20365 kinke changed: What|Removed |Added CC||ki...@gmx.net --- Comment #1 from kinke --- Let's include CatAssign as well for completeness: S[] c; c ~= a; // no copy ctor calls, but 4 postblits --
[Issue 20365] New: Copy constructor not invoked on static arrays of structs but the postblit works
https://issues.dlang.org/show_bug.cgi?id=20365
Issue ID: 20365
Summary: Copy constructor not invoked on static arrays of
structs but the postblit works
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Severity: major
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: sahmi.soulaim...@gmail.com
Copy constructor case: https://run.dlang.io/is/xzVWOi
```
extern(C) void puts(const char*);
struct S
{
long[3] a;
this(ref typeof(this)) { puts("CC"); }
}
void main()
{
S[4] a;
auto b = a;
}
```
Prints nothing. Expected to print "CC" four times.
Postblit case: https://run.dlang.io/is/OX4O8I
```
extern(C) void puts(const char*);
struct S
{
long[3] a;
this(this) { puts("PB"); }
}
void main()
{
S[4] a;
auto b = a;
}
```
prints "PB" four times as expected.
--
[Issue 20363] optimization with union, >> and ulong leads to wrong result
https://issues.dlang.org/show_bug.cgi?id=20363 Ketmar Dark changed: What|Removed |Added CC||ket...@ketmar.no-ip.org --
[Issue 20362] dmd fails to infer scope parameter for delegate
https://issues.dlang.org/show_bug.cgi?id=20362 ZombineDev changed: What|Removed |Added CC||petar.p.ki...@gmail.com --
[Issue 20114] -checkaction=context evaluates operand second time on assertion failure
https://issues.dlang.org/show_bug.cgi?id=20114 --- Comment #2 from Dlang Bot --- dlang/dmd pull request #10512 "Issue 20114 - -checkaction=context evaluates operand second time on a…" was merged into master: - 0ae5f75fce4b8a24341e0b4317f01aedd740aa6e by MoonlightSentinel: Issue 20114 - -checkaction=context evaluates operand second time on assertion failure https://github.com/dlang/dmd/pull/10512 --
[Issue 20114] -checkaction=context evaluates operand second time on assertion failure
https://issues.dlang.org/show_bug.cgi?id=20114 ZombineDev changed: What|Removed |Added Status|NEW |RESOLVED CC||petar.p.ki...@gmail.com Resolution|--- |FIXED --
[Issue 20364] New: [REG2.069] changing length for array of typeof(null) elements kills program
https://issues.dlang.org/show_bug.cgi?id=20364
Issue ID: 20364
Summary: [REG2.069] changing length for array of typeof(null)
elements kills program
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Severity: regression
Priority: P1
Component: druntime
Assignee: nob...@puremagic.com
Reporter: sahmi.soulaim...@gmail.com
Test case:
```
void main()
{
typeof(null)[] result;
result.length = 1;
}
```
--
[Issue 20363] New: optimization with union, >> and ulong leads to wrong result
https://issues.dlang.org/show_bug.cgi?id=20363
Issue ID: 20363
Summary: optimization with union, >> and ulong leads to wrong
result
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: bugzi...@d-ecke.de
auto test(double val)
{
union A
{
double a;
ulong b;
}
A fb;
fb.a = val;
auto sgn = ((fb.b >> 63) & 1) ? "-" : "+";
import std.stdio;
writeln(sgn);
}
void main()
{
test(-double.nan);
}
When running this with "rdmd -O" I get "+" not the expected "-". Without the -O
flag it works.
--
[Issue 20359] core.atomic.cas throws with null as `ifThis`
https://issues.dlang.org/show_bug.cgi?id=20359 John Hall changed: What|Removed |Added CC||john.michael.h...@gmail.com --- Comment #1 from John Hall --- Writing cas(cast(shared), null, cast(shared)b); also causes an error. --
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358 Basile-z changed: What|Removed |Added Hardware|x86_64 |All --
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358 Basile-z changed: What|Removed |Added CC||b2.t...@gmx.com --- Comment #4 from Basile-z --- I vote for closing as invalid. Default constructors should be seen as real constructors so the private fields are not written directly but rather copied from the constructor parameters to the fields (conceptually). Finally explicit constructors can be used in a way that prevents initialization of private fields, if required. --
[Issue 20362] New: dmd fails to infer scope parameter for delegate
https://issues.dlang.org/show_bug.cgi?id=20362
Issue ID: 20362
Summary: dmd fails to infer scope parameter for delegate
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: atila.ne...@gmail.com
With dmd 2.089.0, *without* using -preview=dip1000, this fails to compile:
void stringify(scope void delegate(scope const char[]) sink) {
sink("oops");
}
void main() {
string str;
stringify((chars) {str ~= chars; });
}
bug.d(7): Error: function bug.stringify(scope void delegate(scope
const(char[])) sink) is not callable using argument types (void)
bug.d(7):cannot pass argument __lambda1 of type void to parameter scope
void delegate(scope const(char[])) sink
It compiles successfully with -preview=dip1000. A workaround is to explicitly
add `scope` to the `chars` parameter of the delegate, and then it compiles with
and without -preview=dip1000.
This is blocking this PR:
https://github.com/dlang/dmd/pull/10506
--
[Issue 20361] New: denormalized floatingpoint number literals do not work
https://issues.dlang.org/show_bug.cgi?id=20361
Issue ID: 20361
Summary: denormalized floatingpoint number literals do not work
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: bugzi...@d-ecke.de
void main()
{
float f = 1e-40f;
}
Compiling with dmd produces:
test.d(3): Error: number 1e-40f is not representable
But it is representable. The bit pattern is:
0 001000101101110
--
[Issue 3858] mixin attribute is ignored
https://issues.dlang.org/show_bug.cgi?id=3858 RazvanN changed: What|Removed |Added CC||razvan.nitu1...@gmail.com --- Comment #3 from RazvanN --- Maybe a solution would be to parse string mixins at parse time and simply glue the resulting AST to the main AST. --
[Issue 3852] Default struct constructors needed
https://issues.dlang.org/show_bug.cgi?id=3852 RazvanN changed: What|Removed |Added CC||razvan.nitu1...@gmail.com --- Comment #2 from RazvanN --- How come disbling the default constructor via `disable this();` does not solve your issues? I am tempted to close this as WONTFIX as the board is probably going to be against adding the possibility of defining default constructors. --
[Issue 3825] (D1 only) AAs entries are default initialized before the new value is evaluated
https://issues.dlang.org/show_bug.cgi?id=3825 RazvanN changed: What|Removed |Added Status|REOPENED|RESOLVED CC||razvan.nitu1...@gmail.com Resolution|--- |WONTFIX --- Comment #10 from RazvanN --- D1 is no longer supported, so I am closing this. --
[Issue 3818] Generic error message for wrong foreach
https://issues.dlang.org/show_bug.cgi?id=3818 RazvanN changed: What|Removed |Added Keywords||bootcamp CC||razvan.nitu1...@gmail.com --
[Issue 3790] [OOP] Forwarding constructors to super class
https://issues.dlang.org/show_bug.cgi?id=3790 RazvanN changed: What|Removed |Added Status|NEW |RESOLVED CC||razvan.nitu1...@gmail.com Resolution|--- |INVALID --- Comment #5 from RazvanN --- Closing on the basis of Dmitry's proposed solution. --
[Issue 3773] Incorrectly returning an enum error, points to enum line iso error line
https://issues.dlang.org/show_bug.cgi?id=3773 RazvanN changed: What|Removed |Added Status|NEW |RESOLVED CC||razvan.nitu1...@gmail.com Resolution|--- |FIXED --- Comment #1 from RazvanN --- I cannot reproduce this in git master. I get error for trying to return something from void functions, which is the correct behavior. --
[Issue 3771] warn or disallow method as property usage without @property attribute
https://issues.dlang.org/show_bug.cgi?id=3771 RazvanN changed: What|Removed |Added Status|NEW |RESOLVED CC||razvan.nitu1...@gmail.com Resolution|--- |INVALID --
[Issue 20360] std.variant doesn't do postblit/dtor correctly for large structs
https://issues.dlang.org/show_bug.cgi?id=20360 --- Comment #1 from John Colvin --- related to https://issues.dlang.org/show_bug.cgi?id=12944 --
[Issue 20360] std.variant doesn't do postblit/dtor correctly for large structs
https://issues.dlang.org/show_bug.cgi?id=20360 John Colvin changed: What|Removed |Added Keywords||industry --
[Issue 20360] New: std.variant doesn't do postblit/dtor correctly for large structs
https://issues.dlang.org/show_bug.cgi?id=20360
Issue ID: 20360
Summary: std.variant doesn't do postblit/dtor correctly for
large structs
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Severity: major
Priority: P1
Component: phobos
Assignee: nob...@puremagic.com
Reporter: john.loughran.col...@gmail.com
unittest
{
import std.variant;
static struct S
{
int* p;
ubyte[100] u;
this(int a)
{
p = new int(a);
}
this(this)
{
p = new int(*p);
}
~this()
{
p = null;
}
}
Variant v = S(4);
assert(v.peek!S.p !is null); // fails
}
--
[Issue 11038] static has no effect as a block attribute for imports
https://issues.dlang.org/show_bug.cgi?id=11038 Dlang Bot changed: What|Removed |Added Keywords||pull --- Comment #1 from Dlang Bot --- @RazvanN7 created dlang/dmd pull request #10541 "Fix Issue 11038 - static has no effect as a block attribute for imports" fixing this issue: - Fix Issue 11038 - static has no effect as a block attribute for imports https://github.com/dlang/dmd/pull/10541 --
[Issue 20356] exceeding template expansion limits
https://issues.dlang.org/show_bug.cgi?id=20356 Dlang Bot changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #3 from Dlang Bot --- dlang/dmd pull request #10538 "Fix #20356 raises recursive expansion limit" was merged into stable: - b4a814c02a2bfdf048440ac70a6e93fff596110f by Robert Schadek: Fix Issue 20356 raises recursive expansion limit * Raises the limit from 300 to 500 and gives clearer error msg. * make the test give the right error msg * moved the constant to globals.Global * changed some more constants * rename https://github.com/dlang/dmd/pull/10538 --
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358
--- Comment #3 from Max Samukha ---
(In reply to Simen Kjaeraas from comment #2)
> The solution here would be to define the default constructor as private when
> a struct has private members.
That would be better, but simply making the default constructor private would
outlaw other valid use cases:
struct S {
int x;
private int y;
int z;
}
// partial initialization should still be allowed
S s = {1};
S s2 = {1, z: 2}; // etc
In other words, private should be the respective parameters of the default
constructor.
--
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358 Simen Kjaeraas changed: What|Removed |Added CC||simen.kja...@gmail.com --- Comment #2 from Simen Kjaeraas --- Supporting Max's argument: test.d's author should have no knowledge of S's private fields and their allowed values. A change in S's internals would lead to breakage in unrelated modules. The solution here would be to define the default constructor as private when a struct has private members. Inside s.d this would allow the exact same usage as now, while not giving other modules access to internals that arguably shouldn't be available to them. Constructors would need to be defined explicitly to initialize private fields when called from other modules. I started out agreeing with RazvanN on this, but I think the above is the best solution. --
[Issue 20357] format should obey space flag when printing nan or inf
https://issues.dlang.org/show_bug.cgi?id=20357 Dlang Bot changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #2 from Dlang Bot --- dlang/phobos pull request #7263 "Fix Issue 20357 - format should obey space flag when printing nan or inf" was merged into master: - ea3d45e4c97954109c4aac6f52aaa7a9de95afdf by Bernhard Seckinger: Fix Issue 20357 - format should obey space flag when printing nan or inf https://github.com/dlang/phobos/pull/7263 --
[Issue 20358] External initialization of private struct fields should be disallowed
https://issues.dlang.org/show_bug.cgi?id=20358 RazvanN changed: What|Removed |Added CC||razvan.nitu1...@gmail.com --- Comment #1 from RazvanN --- This issue is invalid. Default constructor is allowed to initialize private fields, otherwise you would end up with partially initialized objects. If this behavior does not suit you, you can disable the default constructor. I suggest closing this as invalid. --
[Issue 20356] exceeding template expansion limits
https://issues.dlang.org/show_bug.cgi?id=20356 Dlang Bot changed: What|Removed |Added Keywords||pull --- Comment #2 from Dlang Bot --- @burner updated dlang/dmd pull request #10538 "Fix #20356 raises recursive expansion limit" fixing this issue: - Fix Issue 20356 raises recursive expansion limit * Raises the limit from 300 to 500 and gives clearer error msg. * make the test give the right error msg * moved the constant to globals.Global https://github.com/dlang/dmd/pull/10538 --
