blob: 4218c8751ad7380be328a4d3273c98bcb4e1c432 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
|
# Reported 93/05/09 by Jim Wilson: IEEE single-precision FLT_MIN value gets
# assembled incorrectly. (Off by one ulp.)
proc dotest {} {
set testname "IEEE FLT_MIN, single-precision"
set x 0
gas_start "x930509a.s" "-al"
while 1 {
expect {
-re " 00 ?00 ?80 ?00\[ \]+.single" { pass $testname; set x 1 }
-re " 00 ?80 ?00 ?00\[ \]+.single" { pass $testname; set x 1 }
-re ".single" { fail $testname; set x 1 }
-re "\[^\n\]*\n" { }
timeout { perror "timeout\n"; break }
eof { break }
}
}
gas_finish
if !$x then { fail "$testname (listing didn't match)" }
}
# C54x alignment/addressing is different, so the listing looks different
# float encoding is tested in c54x-specific tests.
# No floating point support in assembly code for CRIS.
if { ![istarget vax*-*-*] && ![istarget *c54x*-*-*]
&& ![istarget cris-*-*] && ![istarget arc*-*-*] } then {
dotest
}
|
