













































































































































(edited)


























































































































2




















❯ ./bin/llvm-lit tools/mlir/test/CAPI/
-- Testing: 10 tests, 10 workers --
PASS: MLIR :: CAPI/quant.c (1 of 10)
PASS: MLIR :: CAPI/pdl.c (2 of 10)
PASS: MLIR :: CAPI/transform_interpreter.c (3 of 10)
PASS: MLIR :: CAPI/transform.c (4 of 10)
PASS: MLIR :: CAPI/sparse_tensor.c (5 of 10)
PASS: MLIR :: CAPI/llvm.c (6 of 10)
PASS: MLIR :: CAPI/translation.c (7 of 10)
PASS: MLIR :: CAPI/pass.c (8 of 10)
PASS: MLIR :: CAPI/ir.c (9 of 10)
PASS: MLIR :: CAPI/execution_engine.c (10 of 10)
Testing Time: 0.26s
Total Discovered Tests: 10
Passed: 10 (100.00%)




MlirAttribute compile_unit =
mlirLLVMDICompileUnitAttrGet(ctx, id, 0x1C, file, foo, false, 0x1);
// CHECK: #llvm.di_compile_unit<id = distinct[0]<"foo">, sourceLanguage =
// DW_LANG_Rust, file = <"foo" in "bar">, producer = "foo", isOptimized =
// false, emissionKind = Full>
mlirAttributeDump(compile_unit);
obviously the test i add to llvm would mention rust










































(edited)
1
1



(edited)
new-approach branch





(edited)






-i'' help with macos?-i














main.ger2.8303

s/main/central/ whatevermain.ger2.8303 that generates a message to main.banmaster. However our link to main is down. Now that ban could wait until it's back up, but that's not desirable. It should be in effect locally atleast














tail-log main.ger10.8306 and get logs. I know we can bolt all of these things on with bash scripts and call it a day but I don't really like it



main.ger2.8303 that generates a message to main.banmaster. However our link to main is down. Now that ban could wait until it's back up, but that's not desirable. It should be in effect locally atleast 




tail-log main.ger10.8306 and get logs. I know we can bolt all of these things on with bash scripts and call it a day but I don't really like it 










Hacker News • 2024-02-16 09:43:39Z 













uint64_t args[] = { 1 };
MlirAttribute expression_elem =
mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, args);





uint64_t args[] = { 1 };
MlirAttribute expression_elem =
mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, args); MlirAttribute expression_elem = mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, (uint64_t){1});


Incompatible integer to pointer conversion passing 'uint64_t' (aka 'unsigned long') to parameter of type 'const uint64_t *' (aka 'const unsigned long *'); take the address with & (fix available)clang(-Wint-conversion) (edited)

MlirAttribute expression_elem =
mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, &(uint64_t){1});






MlirAttribute expression_elem =
mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, &(uint64_t){1}); 1UL not work?


1UL not work? 








uint64_t is UINT64_C(1)




uint64_t is UINT64_C(1) 




































uint_fast16_t i; (edited)

short, int, long, long longunsigned variants


short, int, long, long long 











uint_fast16_t uint16_t or uint64_t?






uint_fast16_t uint16_t or uint64_t? 


mul



mul (register-constant-multiplication) is faster for 64 bitimul for both 32 and 64 bit has a latency of 3 cycles and it takes 1 cycle


idiv is faster for 32-bit than for 64-bit

mul (register-constant-multiplication) is faster for 64 bit 

mul instruction is faster for 64 bit than for 32 bit

x * 1234 does not compile down to shifts, but to a multiplication


x / 113 compile to btw?



















































ERROR state when a request cannot be added. Now only the invalid request is aborted immediately. The ERROR state is now only entered when a curl function fails in a way where recovery is not possible.
Fix occasional deadlock when HTTP client is shutting down after entering the ERROR state, by also immediately aborting new requests when the HTTP client is already in the ERROR state.
Remove unused CHttp::EState::STOPPING....




go implement it for me then pls ty

go implement it for me then pls ty 

CMapLayers the less i believe it's possible to write
(edited)

1


graphics_threaded it's float










STATE_QUEUED to STATE_RUNNING, but not from STATE_ABORTED. Previously, the code usedd exchange to change the state and compared non-atomically afterwards. This led to STATE_ABORTED being replaced by STATE_RUNNING in a race. With compare_exchange_strong, this can no longer happen.
Fixes #8044.







































const unsigned char *sqlite3_column_text(sqlite3_stmt*, int iCol);
do i need to consider anything if im going to use unsigned char ?const char *username = reinterpret_cast<const char*>(sqlite3_column_text(pstmt, 0)); (edited)

const unsigned char *sqlite3_column_text(sqlite3_stmt*, int iCol);
do i need to consider anything if im going to use unsigned char ? char *, I think

const char *username = reinterpret_cast<const char*>(sqlite3_column_text(pstmt, 0)); (edited)

unsigned char and char are pretty much interchangeable