Gabriel Ebner

> If one message, then how do we handle a non-terminating case. Just as we do now, via timeouts. Right now, we only report trace messages at the end of...

@leodemoura The `EOF` value in the scanner is completely arbitrary (it just shouldn't be 0 or a character that we actually want to read). I'll change it to `0xFF`. I...

The problem already happens when reading `init/core.olean`. Bisecting shows that it happens at the `psigma.has_sizeof` instance. Weirdly enough, switching the `sigma` and `psigma` instances makes the error go away. I'm...

Thanks for figuring out that it works with the olean file generated on amd64. I have tried to add a few checks in the code to narrow down the problem....

This is terribly, terribly weird. We seem to lose 3 bytes when writing the blob section of the olean file. All of these bytes have the value `0xFF` (there are...

> Are these 3 bytes the only ff bytes? That was my first thought as well, but no, other 0xFF bytes survive just fine. > Even worse: they get lost...

The 4096 byte boundaries align perfectly with the 8192 byte buffer that libstdc++ uses: ``` write(4, "oleanfile\0\3\2\1\377xFo\376\0\0\377\0\2R\260TK\0Prop"..., 8192) = 8192 write(4, "\2\377\0\0\4\1PInfo\0\0\347\200\nATTR\0\0i\377\0\0\3\350\0\347\0\0"..., 8192) = 8192 write(4, "\205\10\2x\342\202\201\0\4\0\1\10\2y\342\202\201\0\4\0\1\10\2x\342\202\202\0\4\0P\10"..., 8192) = 8192...

And I have a reproducible minimal test: ```c++ #include #include int main() { constexpr auto fn = "buffer_0xff.out"; constexpr unsigned size = 1000000; { // write a million times 0xFF...

Upstream knows about this breakage: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=47275 > By changing the signedness of char you change the ABI and everything must use the same ABI. > Indeed, all the bets are...

@abliss The latest git version can now compile the standard library (well, half of it, my rpi is slow). Can you confirm that it compiles on your device, and could...