Hi Jeremy, On Sun, 2008-07-06 at 11:09 +1000, Jeremy Apthorp wrote: > I tried compiling sqlite3 on my own machine, and the issue does not > appear. It seems to be in fact unrelated to that -D and more related > to the build environment. Do you still have your build log? If yes, please do send it to me. By the way, I couldn't reproduce this bug. :-|
> There are several comments in the sqlite3 > source (see for example sqlite3/src/util.c) indicating that > -ffast-math is broken. Was this package built with -ffast-math? Was -3 > built differently to -2? Build logs are available to public[1][2]. None of them shows that gcc option -ffast-math was in use. A quick diff shows no interesting difference between the two build logs. > I believe there is now a patch in the sqlite3 CVS tree that #errors > when -ffast-math is specified. Yes, there is[3]. Regards, Laszlo/GCS [1] http://buildd.debian.org/fetch.cgi?&pkg=sqlite3&ver=3.5.9-2&arch=i386&stamp=1212432963 [2] http://buildd.debian.org/fetch.cgi?pkg=sqlite3;ver=3.5.9-3;arch=i386;stamp=1214437066 [3] http://www.sqlite.org/cvstrac/chngview?cn=5345 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
