[Issue 21110] New: OOB memory access, safety violation
https://issues.dlang.org/show_bug.cgi?id=21110
Issue ID: 21110
Summary: OOB memory access, safety violation
Product: D
Version: D2
Hardware: x86_64
OS: Other
Status: NEW
Severity: regression
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: bcarnea...@gmail.com
void main() {
import std.stdio : writeln;
int[] dst;
int[] a;
int[] b;
a.length = 3;
b.length = 3;
dst.length = 4;
dst[] = a[] + b[];
writeln(dst[3]);
}
Prior to 2.077, per Steven Schveighoffer, the above safely failed with an array
violation. Currently on his and my test machines, the above runs to completion
printing out some number (402696894 on my machine, 402653184 on his).
Annotating main with @safe didn't change anything.
--
[Issue 21091] [ICE] Segmentation fault: Module::importAll(Scope*) (this=0x0, prevsc=0x0) at dmd/dmodule.d:1124
https://issues.dlang.org/show_bug.cgi?id=21091 Dlang Bot changed: What|Removed |Added Keywords||pull --- Comment #3 from Dlang Bot --- @RazvanN7 created dlang/dmd pull request #11503 "Fix Issue 21091 - [ICE] Segmentation fault: Module::importAll(Scope*)…" fixing this issue: - Fix Issue 21091 - [ICE] Segmentation fault: Module::importAll(Scope*) (this=0x0, prevsc=0x0) at dmd/dmodule.d:1124 https://github.com/dlang/dmd/pull/11503 --
[Issue 21109] Wrong result when using sort() on enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109 Andrej Mitrovic changed: What|Removed |Added Summary|Possibly wrong codegen when |Wrong result when using |using enum arrays |sort() on enum arrays --
[Issue 21109] Possibly wrong codegen when using enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109 --- Comment #4 from Andrej Mitrovic --- (In reply to Andrej Mitrovic from comment #3) > I think this is related to `sort`. It's possible that sort compares by > pointers if two structs are otherwise equivalent. > > And then using `enum` has a different effect on arrays, possibly optimizing > and storing the same pointer in two structs (making them equal). Sorry the actual reasoning should be: `enum` always allocates a new array, therefore the pointer addresses change each time. It's possible opCmp does a value comparison first, and then a pointer comparison. But I don't think pointers should be compared at all.. at least not when dealing with arrays and not naked pointers in a struct. --
[Issue 21109] Possibly wrong codegen when using enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109 --- Comment #3 from Andrej Mitrovic --- I think this is related to `sort`. It's possible that sort compares by pointers if two structs are otherwise equivalent. And then using `enum` has a different effect on arrays, possibly optimizing and storing the same pointer in two structs (making them equal). --
[Issue 15824] 'scope' attribute ignored when making an alias to a scope delegate
https://issues.dlang.org/show_bug.cgi?id=15824 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |FIXED --- Comment #3 from Mathias LANG --- https://github.com/dlang/dmd/pull/10018 deprecated this. --
[Issue 11754] Disallow changing the default parameters of overridden inherited functions
https://issues.dlang.org/show_bug.cgi?id=11754 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |WONTFIX --- Comment #3 from Mathias LANG --- That seems like a pointless restriction. Closing as WONTFIX. --
[Issue 21109] Possibly wrong codegen when using enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109 --- Comment #2 from Andrej Mitrovic --- I've confirmed the issue with ldc2 as well. So it looks like it's a front-end bug. --
[Issue 21109] Possibly wrong codegen when using enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109 --- Comment #1 from Andrej Mitrovic --- Using DMD v2.093.0. --
[Issue 13112] Ignore constness when copying dynamic array argument to static array parameter
https://issues.dlang.org/show_bug.cgi?id=13112
Mathias LANG changed:
What|Removed |Added
Status|NEW |RESOLVED
CC||pro.mathias.l...@gmail.com
Resolution|--- |INVALID
--- Comment #1 from Mathias LANG ---
The problem is not constness, but that dynamic arrays do not convert implicitly
to static array (which IMO is not an issue).
The difference you're seeing with variables is because initializer are treated
differently. I would say that this behavior is more likely to be the issue than
not having implicit conversion from static to dynamic array.
The following compiles (but obviously triggers a range error):
```
void f(char[1] arr) {}
void main()
{
string s;
f(s[0 .. 1]);
}
```
So closing as invalid.
--
[Issue 21109] New: Possibly wrong codegen when using enum arrays
https://issues.dlang.org/show_bug.cgi?id=21109
Issue ID: 21109
Summary: Possibly wrong codegen when using enum arrays
Product: D
Version: D2
Hardware: All
OS: Mac OS X
Status: NEW
Severity: critical
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: andrej.mitrov...@gmail.com
-
import std.algorithm;
import std.stdio;
struct S
{
int i;
int[] arr;
}
void main ()
{
enum q1 = [
S(2, [0, 0]),
];
enum q2 = [
S(2, [0, 0]),
];
enum q3 = [
S(2, [1, 1]),
];
foreach (idx; 0 .. 10) // 10 iterations
{
const unique_count = [q1, q2, q3].sort.uniq.count;
writefln("Uniques: %s", unique_count);
}
}
-
This produces:
```
$ dmd -run test.d
> Uniques: 2
Uniques: 2
Uniques: 2
Uniques: 2
Uniques: 2
Uniques: 3
Uniques: 2
Uniques: 2
Uniques: 2
Uniques: 2
```
Notice how in one run the number of uniques was 3!
If I change `enum` to `auto` the bug goes away.
--
[Issue 16479] Missing substitution while mangling C++ template parameter for functions
https://issues.dlang.org/show_bug.cgi?id=16479 Mathias LANG changed: What|Removed |Added Status|REOPENED|RESOLVED Resolution|--- |FIXED --- Comment #13 from Mathias LANG --- Moved to https://issues.dlang.org/show_bug.cgi?id=21108 as the original issue has been mostly fixed and only a few test-cases remains, which can be worked around. --
[Issue 21108] New: Wrong mangling for extern(C++, std) (identifier namespace)
https://issues.dlang.org/show_bug.cgi?id=21108
Issue ID: 21108
Summary: Wrong mangling for extern(C++, std) (identifier
namespace)
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Keywords: C++
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: pro.mathias.l...@gmail.com
Taken from: https://issues.dlang.org/show_bug.cgi?id=16479#c11
```
extern (C++, std)
struct pair(T1, T2) {}
extern (C++)
void func_20413(pair!(int, float), pair!(float, int));
```
actual: _Z10func_20413St4pairIifEStS_IfiE
expected: _Z10func_20413St4pairIifES_IfiE
A C++ string namespace `extern (C++, "std") struct pair(T1, T2) {}` works as
expected (testcase from https://issues.dlang.org/show_bug.cgi?id=20413).
--
[Issue 21096] [ICE] Segmentation fault in dmd.hdrgen.sizeToBuffer at dmd/hdrgen.d:3153
https://issues.dlang.org/show_bug.cgi?id=21096 Dlang Bot changed: What|Removed |Added Keywords||pull --- Comment #1 from Dlang Bot --- @RazvanN7 created dlang/dmd pull request #11502 "Fix Issue 21096 - [ICE] Segmentation fault in dmd.hdrgen.sizeToBuffer at dmd/hdrgen.d:3153" fixing this issue: - Fix Issue 21096 - [ICE] Segmentation fault in dmd.hdrgen.sizeToBuffer at dmd/hdrgen.d:3153 https://github.com/dlang/dmd/pull/11502 --
[Issue 17928] [scope] `in` is not treated as `const scope`
https://issues.dlang.org/show_bug.cgi?id=17928 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |FIXED --- Comment #8 from Mathias LANG --- Implemented in https://github.com/dlang/dmd/pull/10769 --
[Issue 13752] add traits to query storage class of variables (isTLS, isGlobal)
https://issues.dlang.org/show_bug.cgi?id=13752 Mathias LANG changed: What|Removed |Added CC||andrej.mitrov...@gmail.com --- Comment #1 from Mathias LANG --- *** Issue 12474 has been marked as a duplicate of this issue. *** --
[Issue 12474] Implement the getStorageClass trait
https://issues.dlang.org/show_bug.cgi?id=12474 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |DUPLICATE --- Comment #4 from Mathias LANG --- Regarding storage class, we have getParameterStorageClass (implemented in https://github.com/dlang/dmd/pull/6829) for parameter. For variables, we have another issue (13752), so marking as duplicate. *** This issue has been marked as a duplicate of issue 13752 *** --
[Issue 7846] Forward referenced storage class is not applied in certain cases
https://issues.dlang.org/show_bug.cgi?id=7846 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |FIXED --- Comment #3 from Mathias LANG --- Fixed in v2.084: ``` 2.063 to 2.083.1: Success with output: - int true const(int) false - Since 2.084.1: Success with output: - const(int) false const(int) false - ``` --
[Issue 6931] scope parameter storage class not checked at all
https://issues.dlang.org/show_bug.cgi?id=6931 Mathias LANG changed: What|Removed |Added Status|NEW |RESOLVED CC||pro.mathias.l...@gmail.com Resolution|--- |FIXED --- Comment #6 from Mathias LANG --- Solved by DIP1000. --
[Issue 21092] [ICE] Segmentation fault in ExpressionPrettyPrintVisitor::visit(CommaExp*) at dmd/hdrgen.d:2293
https://issues.dlang.org/show_bug.cgi?id=21092 Dlang Bot changed: What|Removed |Added Keywords||pull --- Comment #2 from Dlang Bot --- @RazvanN7 created dlang/dmd pull request #11500 "Fix Issue 21092 -[ICE] Segmentation fault in ExpressionPrettyPrintVis…" fixing this issue: - Fix Issue 21092 -[ICE] Segmentation fault in ExpressionPrettyPrintVisitor::visit(CommaExp*) at dmd/hdrgen.d:2293 https://github.com/dlang/dmd/pull/11500 --
[Issue 21107] New: Cannot define an r/w property inside a function
https://issues.dlang.org/show_bug.cgi?id=21107
Issue ID: 21107
Summary: Cannot define an r/w property inside a function
Product: D
Version: D2
Hardware: x86_64
OS: Linux
Status: NEW
Severity: normal
Priority: P1
Component: dmd
Assignee: nob...@puremagic.com
Reporter: por...@narod.ru
It wonders me that the following program does not compile:
void main() {
@property int x() { return 0; }
@property void x(int) { }
}
x.d(3): Error: declaration x is already defined
Is it a DMD v2.090.1 bug?
If it should not be compilable indeed, it is nevertheless a bug: The error
message should be more clear.
--
[Issue 21106] New: Add type inference for constructor calls
https://issues.dlang.org/show_bug.cgi?id=21106
Issue ID: 21106
Summary: Add type inference for constructor calls
Product: D
Version: D2
Hardware: All
OS: All
Status: NEW
Severity: enhancement
Priority: P1
Component: dlang.org
Assignee: nob...@puremagic.com
Reporter: por...@narod.ru
I propose to add type inference (similar to one for functions) for constructor
calls, that is to make the following program legal:
class C(T) {
this(T) { }
}
void main() {
immutable x = new C(0);
}
This would simplify programming eliminating the need to create a factory
function or factory functions for each template class.
Pro: shorter programs.
Con: some consider it a good practice to make a factory function ALWAYS.
Pro: even if you create a factory function, the body of this function itself
would be shortened.
--
[Issue 20380] std.net.curl: misbehaviour on arch 'arm32v7 gnueabihf' (Raspberry/Raspbian/Buster)
https://issues.dlang.org/show_bug.cgi?id=20380 --- Comment #9 from Iain Buclaw --- (In reply to Iain Buclaw from comment #8) > GCC 10.2 is out https://gcc.gnu.org/gcc-10/ Changelog entry for std.net.curl change: https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=libphobos/ChangeLog;h=bf6eefd3bd1a8b300d9993e253a90679575fc4be;hb=ee5c3db6c5b2c3332912fb4c9cfa2864569ebd9a#l1 Built and ran dub on aarch64 server, issue no longer occurs. --
[Issue 21105] Casting from a function pointer to a delegate
https://issues.dlang.org/show_bug.cgi?id=21105 Dennis changed: What|Removed |Added CC||dkor...@live.nl --- Comment #1 from Dennis --- Are you aware of std.functional.toDelegate? https://dlang.org/phobos/std_functional.html#toDelegate Implicit casting to delegate probably needs a DIP or at least some good rationale. --
[Issue 21105] New: Casting from a function pointer to a delegate
https://issues.dlang.org/show_bug.cgi?id=21105 Issue ID: 21105 Summary: Casting from a function pointer to a delegate Product: D Version: D2 Hardware: All OS: All Status: NEW Severity: enhancement Priority: P1 Component: dlang.org Assignee: nob...@puremagic.com Reporter: por...@narod.ru Please add the following features: - explicit casting from a function pointer to the corresponding delegate (with the obvious semantics) - implicit casting from a function pointer to the corresponding delegate - `cast(delegate) ...` that would also do this cast. The state would be set to null by such casts. --
[Issue 20380] std.net.curl: misbehaviour on arch 'arm32v7 gnueabihf' (Raspberry/Raspbian/Buster)
https://issues.dlang.org/show_bug.cgi?id=20380 Iain Buclaw changed: What|Removed |Added Depends on||19367 Referenced Issues: https://issues.dlang.org/show_bug.cgi?id=19367 [Issue 19367] std.net.curl does not understand HTTP/2 status lines --
[Issue 19367] std.net.curl does not understand HTTP/2 status lines
https://issues.dlang.org/show_bug.cgi?id=19367 Iain Buclaw changed: What|Removed |Added Blocks||20380 Referenced Issues: https://issues.dlang.org/show_bug.cgi?id=20380 [Issue 20380] std.net.curl: misbehaviour on arch 'arm32v7 gnueabihf' (Raspberry/Raspbian/Buster) --
[Issue 20380] std.net.curl: misbehaviour on arch 'arm32v7 gnueabihf' (Raspberry/Raspbian/Buster)
https://issues.dlang.org/show_bug.cgi?id=20380 Iain Buclaw changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #8 from Iain Buclaw --- GCC 10.2 is out https://gcc.gnu.org/gcc-10/ --
[Issue 21089] With vm.overcommit_memory=0, DMD can't link if it uses more than half the total (ram+swap) memory in the system.
https://issues.dlang.org/show_bug.cgi?id=21089 Dlang Bot changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #3 from Dlang Bot --- dlang/dmd pull request #11497 "Fix issue 21089: spawn linker with vfork to avoid need to overcommit memory" was merged into master: - c7aef70550c291928b20d3daa9bb4d643c460693 by Mathis Beer: Fix issue 21089: spawn linker with vfork to avoid need to overcommit memory - dmd is large, ld is small. https://github.com/dlang/dmd/pull/11497 --
[Issue 21089] With vm.overcommit_memory=0, DMD can't link if it uses more than half the total (ram+swap) memory in the system.
https://issues.dlang.org/show_bug.cgi?id=21089 --- Comment #2 from Dlang Bot --- @FeepingCreature updated dlang/dmd pull request #11497 "Spawn linker with vfork to avoid need to overcommit memory" fixing this issue: - Fix issue 21089: spawn linker with vfork to avoid need to overcommit memory - dmd is large, ld is small. https://github.com/dlang/dmd/pull/11497 --
[Issue 2789] Functions overloads are not checked for conflicts
https://issues.dlang.org/show_bug.cgi?id=2789
Iain Buclaw changed:
What|Removed |Added
CC||ibuc...@gdcproject.org
--- Comment #19 from Iain Buclaw ---
(In reply to Witold Baryluk from comment #18)
> Interesting, gdc 9.2 does report duplicate:
>
> :16:6: error: example.B.m at :16:6 conflicts with
> example.B.m at :12:6
>
>16 | int m() {
>
>
>
> The issue is still present in dmd nightly, dmd 2.089, ldc 1.20, ldc trunk.
>
> Link with an example: https://godbolt.org/z/xE6fK7
This is checked in the codegen pass. Would be better if the front-end did it
for us.
--
