From: Gilles E. <g....@fr...> - 2011-05-11 17:56:31
|
Looking in the test suite result, the difference was one test - busybox-1.17.4 FAIL 'date-R-works' That fail on gcc-4.4.6 the first time but work the second time I tried. This test should be racy, it mostly work but sometime fail, so that should be only noise. - gdb-7.2 there is less failures, the diff show Running ./gdb.base/miscexprs.exp ... -FAIL: gdb.base/miscexprs.exp: print value of ibig.i[100] | 1 -FAIL: gdb.base/miscexprs.exp: print value of sbig.s[90] & 127 -FAIL: gdb.base/miscexprs.exp: print value of !ibig.i[100] -FAIL: gdb.base/miscexprs.exp: print value of !sbig.s[90] -FAIL: gdb.base/miscexprs.exp: print value of !ibig.i[100] -FAIL: gdb.base/miscexprs.exp: print value of !ibig.i[100] -FAIL: gdb.base/miscexprs.exp: print value of !sbig.s[90] * 10 -FAIL: gdb.base/miscexprs.exp: print value of ibig.i[100] * sbig.s[90] -FAIL: gdb.base/miscexprs.exp: print value of fbig.f[100] * dbig.d[202] -FAIL: gdb.base/miscexprs.exp: print value of !(sbig.s[90] * 2) -FAIL: gdb.base/miscexprs.exp: print value of ibig.i[100] << 2 -FAIL: gdb.base/miscexprs.exp: print value of sbig.s[90] >> 4 -FAIL: gdb.base/miscexprs.exp: print value of lbig.l[333] >> 6 Running ./gdb.base/float.exp ... Running ./gdb.base/longjmp.exp ... FAIL: gdb.base/longjmp.exp: next over longjmp(1) @@ -221,8 +207,8 @@ === gdb Summary === -# of expected passes 5178 -# of unexpected failures 20 +# of expected passes 5191 +# of unexpected failures 7 So that's not bad. The full gdb test had some other FAIL, the total FAIL number is reduced from 38 to 25. Looking at the log diff, using some sed -i -e 's|4\.4\.5|4.4.6|g' to reduce the noise, I find nothing alarming, just the usual noise log. So gcc-4.4.6 should just be something good. The drawback is that there is 497 files that are different and not yet include in update. I haven't checked which include gcc version and which are differently compiled. Is it the time to update gcc? Gilles |