5724 2014-01-14 14:46:27 +0000 bitbake -e sometimes is not useful 2014-03-27 10:18:13 +0000 1 1 7 Build System, Metadata & Runtime BitBake bitbake unspecified x86 Multiple RESOLVED FIXED Medium enhancement 1.6 1 gary richard.purdie bluelightning poky.bs.watcher poky.watcher --- --- Don't know oldest_to_newest 39801 0 gary 2014-01-14 14:46:27 +0000 39802 1 gary 2014-01-14 14:49:22 +0000 Sometimes 'bitbake -e recipe' only prints an error and not the useful variable dump. To recreate this using OE-core, simply remove x11 from DISTRO_FESTURES, e.g. in local.conf DISTRO_FEATURES_remove = "x11" Then you will see this error: $ bitbake -e virtual/libx11 WARNING: Host distribution "Fedora-13" has not been validated with this version of the build system; you may possibly experience unexpected failures. It is recommended that you use a tested distribution. Loading cache: 100% |##################################################################################| ETA: 00:00:00 Loaded 1188 entries from dependency cache. ERROR: Nothing PROVIDES 'virtual/libx11' ERROR: libx11-diet PROVIDES virtual/libx11 but was skipped: missing required distro feature ['x11'] (not in DISTRO_FEATURES) ERROR: libx11 PROVIDES virtual/libx11 but was skipped: missing required distro feature ['x11'] (not in DISTRO_FEATURES) ERROR: nativesdk-libx11 PROVIDES virtual/libx11 but was skipped: missing required distro feature ['x11'] (not in DISTRO_FEATURES) It seems to only happen on virtual targets. A similar test $ bitbake -e gst-plugins-base will yield the correct dump. 39867 2 bluelightning 2014-01-16 15:51:18 +0000 I wonder if we should put errors out to stderr so that they don't get redirected by default, and thus would have been visible in this case. This would be easy to do but would be a change in established behaviour. (That's assuming we can't just bypass errors like this when -e is used.) 41779 3 richard.purdie 2014-03-26 16:35:35 +0000 I'd like to understand which bug we're trying to fix here. Removing x11 from DISTRO_FEATURES and then trying to build libx11 is going to generate an error since the x11 recipes don't parse in that scenerio and nothing does provide libx11. The error message provided does even say what the problem is. So is the real issue this bug is addressing that error messages go to stdout? If we fix that are we done here? I suspect there is some context I'm missing, maybe a discussion between Gary and Paul about a specific bug this triggered? 41800 4 gary 2014-03-26 21:18:38 +0000 This came as a result of this thread: https://lists.yoctoproject.org/pipermail/yocto/2014-January/017825.html Basically, I was trying to create a QTe environment, so X11 was removed from DISTRO_FEATURES which yielded an inconsistency error. The recommendation was to use 'bitbake -e' to diagnose it, but this proved unworkable because the process bailed before printing any of the useful variables which might help diagnose the situation. 41820 5 richard.purdie 2014-03-27 10:18:13 +0000 Code to make stdout/stderr work better is now in: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=6ef47ec5fdec8f4b1315dc269b50a02c5b58fbaf