| 00:00.12 | CosmicPenguin | Thats good, because you're scaring me over here |
| 00:01.05 | kergoth | pb_: just realized that the manifest generated do_stage will be wayyyyy off for -native builds |
| 00:01.08 | kergoth | hmmm |
| 00:01.48 | kergoth | and we cant just arbitrarily replace bindir with STAGING_BINDIR, since STAGING_BINDIR is always the native bin path, not target, ever |
| 00:02.58 | pb_ | oh yeah, right |
| 00:03.07 | pb_ | hrm |
| 00:03.16 | kergoth | if bindir was arch dependent, then we could, and itd "just work" |
| 00:03.30 | kergoth | but then non -native .oe's wouldnt be able to use STAGING_BINDIR anymore for their native scripts |
| 00:03.34 | kergoth | heh |
| 00:03.36 | kergoth | damnit |
| 00:05.18 | kergoth | i guess we could make it staging_dir/${HOST_SYS}/bin in manifest output, so the current binary arch will be the one used, then the target arch .oe's could reference staging_dir/${BUILD_SYS}/bin/ |
| 00:07.06 | pb_ | yah, that could work |
| 00:10.52 | treke_ | I need to learn how to do some fancy routing shiznit |
| 00:12.17 | pb_ | best check the advanced routing HOWTO |
| 00:13.28 | treke_ | probably |
| 00:14.08 | kergoth | on the plus side, the native oeclass now has to do less. the staging dirs will automatically flip, since HOST_SYS changes |
| 00:16.26 | kergoth | heh, CFLAGS and LDFLAGS just got simpler too |
| 00:16.34 | pb_ | yay |
| 00:16.44 | kergoth | simpliciation rocks |
| 00:16.50 | kergoth | er, simplification too |
| 00:16.58 | pb_ | heh |