On Fri, 19 Dec 2025, Pali Rohár wrote:
On Friday 19 December 2025 11:41:30 Martin Storsjö wrote:
On Wed, 17 Dec 2025, Pali Rohár wrote:
Changes were applied on top of the c4ae0f47b9adf0b6b91dcc906f5b1fbfe342bb42
Pali Rohár (5):
crt: Use native _getc_nolock when possible
crt: Add missing _imp symbol for mingw-w64 emulated _nolock functions
headers: Unify _getc_nolock and _putc_nolock declarations
crt: Provide _fgetc_nolock() function for all msvcrt builds
crt: Provide _fputc_nolock() function for all msvcrt builds
These changes look good to me, and pass fine through CI. But I'm wondering
about how widely you've tested that calling these functions actually work (I
primarily care about msvcrt and ucrt), as I'm not sure if anything in our
testsuite calls them and there's no new test added here.
If you've tested building and linking code calling all the functions
touched, both for msvcrt and ucrt (I know the patches only intend to change
things for msvcrt, but you never know), this should be ok to merge.
// Martin
Yes, I tested it just manually on stdin and stdout. I have not written
any automated test or some heavy testing.
Ok, I pushed this patchset.
// Martin
_______________________________________________
Mingw-w64-public mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
