Menu
▾
▴
#1615 Confusing warnings on signed char comparison
When comparing a signed char and unsigned char, the following warning is printed:
warning 185: comparison of 'signed char' with 'unsigned char' requires promotion to int
1. When performing arithmetics (*, /, %) on the signed operand, only with % the warning is printed. What is the logic behind this?
2. When performing / or % on an unsigned bitfield, the warning is confusingly printed (while it is not printed when performing a direct comparison). Why?
SDCC : mcs51/gbz80/z80/ds390/pic16/pic14/TININative/ds400/hc08 2.9.7 #5677 (Feb 7 2010) (MINGW32)
Jan Waclawek
With the attached bug2.c, the following warnings are generated,
bug2.c:11: warning 185: comparison of 'signed char' with 'unsigned char' requires promotion to int
bug2.c:12: warning 185: comparison of 'signed char' with 'unsigned char' requires promotion to int
bug2.c:15: warning 185: comparison of 'signed char' with 'unsigned char' requires promotion to int
bug2.c:17: warning 185: comparison of 'signed char' with 'unsigned char' requires promotion to int
I can understand there are warnings for "s.a / 1" and "s.a % 1", since s.a is 7-bit-field, and their result types are promoted to "signed char", and there is no waning for "s.a * 1" since the result type is promoted to "signed int".
But I am confused with that there is no warning for "q / 1", but there is warning for "q % 1". Why "q / 1" is promoted to int, while "q % 1" remains signed char? Should the result type be the same for char / char and char % char ?
Ben
By adding "case '%': " to line 2316 SDCCsymt.c, schar%schar was promoted to int, but some regression test cases failed.
Running ds390 regression tests
Summary for 'ds390': 2 failures, 9279 tests, 1746 test cases, 6804853 bytes, 901116852 ticks
Failure: gen/ds390/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running hc08 regression tests
Summary for 'hc08': 2 failures, 9240 tests, 1747 test cases, 2530961 bytes, 31789256 ticks
Failure: gen/hc08/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running host regression tests
Summary for 'host': 0 failures, 9200 tests, 1747 test cases, 0 bytes, 0 ticks
making huge library
Running mcs51-huge regression tests
Summary for 'mcs51-huge': 1 failures, 9293 tests, 1746 test cases, 3516296 bytes, 457061376 ticks
Failure: gen/mcs51-huge/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running mcs51-large regression tests
Summary for 'mcs51-large': 1 failures, 9291 tests, 1746 test cases, 3055423 bytes, 383109888 ticks
Failure: gen/mcs51-large/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running mcs51-medium regression tests
Summary for 'mcs51-medium': 7 failures, 7302 tests, 1746 test cases, 2668197 bytes, 268918092 ticks
Failure: gen/mcs51-medium/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Failure: gen/mcs51-medium/onebyte/onebyte_attrL_volatile_attrR_volatile
Running mcs51-small regression tests
Summary for 'mcs51-small': 1 failures, 7307 tests, 1746 test cases, 2250290 bytes, 221564460 ticks
Failure: gen/mcs51-small/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running mcs51-stack-auto regression tests
Summary for 'mcs51-stack-auto': 1 failures, 7307 tests, 1746 test cases, 2506991 bytes, 231922188 ticks
Failure: gen/mcs51-stack-auto/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running mcs51-xstack-auto regression tests
Summary for 'mcs51-xstack-auto': 1 failures, 7308 tests, 1746 test cases, 3743400 bytes, 400707288 ticks
Failure: gen/mcs51-xstack-auto/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running s08 regression tests
Summary for 's08': 2 failures, 9242 tests, 1747 test cases, 2430235 bytes, 30863063 ticks
Failure: gen/s08/gcc-torture-execute-divmod-1/gcc-torture-execute-divmod-1
Running stm8 regression tests
Summary for 'stm8': 0 failures, 9295 tests, 1747 test cases, 1926373 bytes, 7539671 ticks
Running ucgbz80 regression tests
Summary for 'ucgbz80': 0 failures, 9278 tests, 1747 test cases, 2621411 bytes, 20115802 ticks
Running ucr2k regression tests
Summary for 'ucr2k': 0 failures, 9286 tests, 1747 test cases, 2289093 bytes, 11489227 ticks
Running ucr3ka regression tests
Summary for 'ucr3ka': 0 failures, 9286 tests, 1747 test cases, 2299016 bytes, 10620168 ticks
Running ucz180 regression tests
Summary for 'ucz180': 0 failures, 9286 tests, 1747 test cases, 2384136 bytes, 12421589 ticks
Running ucz80 regression tests
Summary for 'ucz80': 0 failures, 9297 tests, 1747 test cases, 2318778 bytes, 12574920 ticks
Bug looks like to be fixed. Compiler warn on line with comment 'warning (expected)' only.