On 17/10/06 at 13:25 +0100, Stephen Gran wrote: > This one time, at band camp, Lucas Nussbaum said: > > During a rebuilt of all packages in etch, I discovered that your package > > failed to build on AMD64. > > > clamav-milter.c:1065:84: error: operator '>=' has no left operand > > clamav-milter.c:1586:84: error: operator '>=' has no left operand > > > The full build log is available from > > http://ox.blop.info/bazaar/buildlogs/20061016/amd64/clamav_0.88.4-2.buildlog > > I don't know how you managed to provoke that failure. This only happens > when the preprocessor defines SENDMAIL_VERSION_A and SENDMAIL_VERSION_B > are undef'ed. But the build log shows: > > applying patch 02_milter_sendmail_version_patch to ./ ... > patching file clamav-milter/clamav-milter.c > Hunk #1 succeeded at 29 with fuzz 2 > > Now, clamav-config.h will likely have #undef SENDMAIL_VERSION_A, since > sendmail wasn't installed during the build. However, it is #included at > line 27, and the patch output shows my patch which #defines these applying > at line 29, so the expected define should be late enough to take effect. > Very confusing. > > I guess I can revisit the patch to clean up the fuzz, and make sure it is > applied after clamav-config.h is #include'd, but it looks to me like it > already is, and it still managed to fail. Given that this works on every > other buildd, I am a bit curious to know how this went wrong. Can you > help debug at all? Do any of my esteemed readers have any other ideas?
Hi Stephen, I'm using a not fully configured nullmailer package, just because I need a mail-transport-agent in the chroot. It might not help. I'll see if I can improve the situation and keep you updated. Lucas
signature.asc
Description: Digital signature
