shafik added a subscriber: rsmith.
shafik added a comment.
Thank you for this fix.
================
Comment at: clang/lib/Sema/SemaInit.cpp:808
unsigned NumElems = numStructUnionElements(ILE->getType());
- if (RDecl->hasFlexibleArrayMember())
+ if (!RDecl->isUnion() && RDecl->hasFlexibleArrayMember())
++NumElems;
----------------
Fznamznon wrote:
> Just for some context, numStructUnionElements checks that there is a flexible
> array member and returns number_of_initializable_fields-1 for structs. For
> unions it just returns 1 or 0, so flexible array member caused adding one
> more element to initlistexpr that was never properly handled.
>
> Instead of doing this change, we could probably never enter initialization
> since the record (union) declaration is not valid, but that is not the case
> even for other types of errors in code, for example, I've tried declaring
> field of struct with a typo:
>
> ```
> struct { cha x[]; } r = {1};
> ```
> Initialization is still performed by clang.
> Also, it seems MSVC considers flexible array member inside union as valid, so
> the test code is probably not always invalid.
I am not sure what to think here, looking at gcc documentation for this
extension: https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
and using the following code:
```
struct f1 {
int x; int y[];
} f1 = { 1, { 2, 3, 4 } }; // #1
struct f2 {
struct f1 f1; int data[3];
} f2 = { { 1 }, { 2, 3, 4 } }; // #2
struct { char x[]; } r = {1}; // #3
```
gcc rejects 2 and 3 even though 2 comes from their documentation. Clang warns
on 2 and MSVC rejects 2
CC @aaron.ballman @rsmith
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D147626/new/
https://reviews.llvm.org/D147626
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
