You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Submit bug report to either GCC or just MinGW devs, depending on #1; probably a bug tracker, although I think mailing lists will get quicker responses (or do all bug trackers to mailing lists? idk)
The text was updated successfully, but these errors were encountered:
Done (as of a couple months ago): header inclusion order, underlying max alignment definition issue.
The underlying cause is that GCC supports some long double type (via compiler builtin __float128), which they align to 16 bytes. However, the provided malloc links directly (apparently without wrapper code?) to the malloc from Microsoft/Windows CRT. So there is a semantic inconsistency here, because max_align_t has to be the maximum primitive alignment, but malloc doesn't obide (Microsoft don't support that __float128 I think).
It doesn't look like the devs are terribly interested in either a practical solution to the inclusion order issue, nor the underlying issue.
Maybe this repository should provide a "quick fix header", but if you took the time to read through the readme, the fix is obvious. And maybe there's an easier way for you to fix the system header, but overwriting system headers is pretty icky in the first place (will probably not survive compiler updates), so I don't see a "clean solution" right now either.
Submit bug report to either GCC or just MinGW devs, depending on #1; probably a bug tracker, although I think mailing lists will get quicker responses (or do all bug trackers to mailing lists? idk)
The text was updated successfully, but these errors were encountered: