)]}'
{"diskimage_builder/lib/disk-image-create":[{"author":{"_account_id":7118,"name":"Ian Wienand","email":"iwienand@redhat.com","username":"iwienand"},"change_message_id":"427ee2f1fb1aa36cc455763ac951f112172a72e4","unresolved":false,"context_lines":[{"line_number":434,"context_line":"export TMP_IMAGE_DIR"},{"line_number":435,"context_line":"# Try the \u0027old fashioned\u0027 way calling the block device"},{"line_number":436,"context_line":"# phase. If this gives no result, use the configuration based approach:"},{"line_number":437,"context_line":"eval_run_d block-device \"IMAGE_BLOCK_DEVICE[S]*\u003d\""},{"line_number":438,"context_line":""},{"line_number":439,"context_line":"if [ -z ${IMAGE_BLOCK_DEVICE} ] ; then"},{"line_number":440,"context_line":"    # For compatibily reasons in addition to the YAML configuration"}],"source_content_type":"application/x-shellscript","patch_set":1,"id":"3f79a3b5_d059765c","line":437,"updated":"2018-10-18 23:15:22.000000000","message":"I agree that if using this path, and IMAGE_BLOCK_DEVICE is set by these scripts, we don\u0027t then set any value for IMAGE_BLOCK_DEVICES per below, and the bootloader element will try and eval something that doesn\u0027t exist and bail.\n\nSo you are using block-device.d scripts to do setup?  Is there something that can\u0027t be done with the block-device YAML files?\n\nSince this is supposed to only be a path for backwards compatibility, and the only thing that uses the multiple-device list it is the bootloader on PPC which has to find the boot partition, I\u0027m thinking it probably makes more sense to have this defined by default as an empty array\n\n IMAGE_BLOCK_DEVICES\u003d(  )\n\nthat would seem to keep the intent of any existing scripts just continuing to work.  Because any existing block-device.d script is not going to know to output IMAGE_BLOCK_DEVICES as an eval-able array.\n\n*If* you require this -- and now it means we have legacy block device scripts that want to configure the PPC bootloader -- with I don\u0027t think is really what you want to do -- then we should make sure we also document that legacy scripts should output this in the developer documentation.  But I doubt that\u0027s actually what\u0027s going on here.","commit_id":"194f35473f78e4bb0a80f2750a2a187d3cb0c1e7"},{"author":{"_account_id":4146,"name":"Clark Boylan","email":"cboylan@sapwetik.org","username":"cboylan"},"change_message_id":"95b18410ec10febc21420e50909ed76e492f2462","unresolved":false,"context_lines":[{"line_number":434,"context_line":"export TMP_IMAGE_DIR"},{"line_number":435,"context_line":"# Try the \u0027old fashioned\u0027 way calling the block device"},{"line_number":436,"context_line":"# phase. If this gives no result, use the configuration based approach:"},{"line_number":437,"context_line":"eval_run_d block-device \"IMAGE_BLOCK_DEVICE[S]*\u003d\""},{"line_number":438,"context_line":""},{"line_number":439,"context_line":"if [ -z ${IMAGE_BLOCK_DEVICE} ] ; then"},{"line_number":440,"context_line":"    # For compatibily reasons in addition to the YAML configuration"}],"source_content_type":"application/x-shellscript","patch_set":1,"id":"3f79a3b5_932f8888","line":437,"range":{"start_line":437,"start_character":43,"end_line":437,"end_character":47},"updated":"2018-10-18 23:03:52.000000000","message":"This is evaluated as a grep regex. Considering that we don\u0027t want an arbitrary number of S\u0027s can we rewrite this to be:\n\n  \"IMAGE_BLOCK_DEVICES\\?\u003d\"","commit_id":"194f35473f78e4bb0a80f2750a2a187d3cb0c1e7"},{"author":{"_account_id":16068,"name":"Tobias Henkel","email":"tobias.henkel@bmw.de","username":"tobias.henkel"},"change_message_id":"acdbd74a113bd4344def2cb60827476dbfff7c18","unresolved":false,"context_lines":[{"line_number":434,"context_line":"export TMP_IMAGE_DIR"},{"line_number":435,"context_line":"# Try the \u0027old fashioned\u0027 way calling the block device"},{"line_number":436,"context_line":"# phase. If this gives no result, use the configuration based approach:"},{"line_number":437,"context_line":"eval_run_d block-device \"IMAGE_BLOCK_DEVICE[S]*\u003d\""},{"line_number":438,"context_line":""},{"line_number":439,"context_line":"if [ -z ${IMAGE_BLOCK_DEVICE} ] ; then"},{"line_number":440,"context_line":"    # For compatibily reasons in addition to the YAML configuration"}],"source_content_type":"application/x-shellscript","patch_set":1,"id":"3f79a3b5_b1cab944","line":437,"in_reply_to":"3f79a3b5_b135d9de","updated":"2018-10-19 06:08:14.000000000","message":"An alternative solution to my problem: If1d834c17b6aab6ac8ff57e84ecf57f6fa1ae34f","commit_id":"194f35473f78e4bb0a80f2750a2a187d3cb0c1e7"},{"author":{"_account_id":16068,"name":"Tobias Henkel","email":"tobias.henkel@bmw.de","username":"tobias.henkel"},"change_message_id":"dc7ff0ba18b702f9412160eba22328e0b2f18667","unresolved":false,"context_lines":[{"line_number":434,"context_line":"export TMP_IMAGE_DIR"},{"line_number":435,"context_line":"# Try the \u0027old fashioned\u0027 way calling the block device"},{"line_number":436,"context_line":"# phase. If this gives no result, use the configuration based approach:"},{"line_number":437,"context_line":"eval_run_d block-device \"IMAGE_BLOCK_DEVICE[S]*\u003d\""},{"line_number":438,"context_line":""},{"line_number":439,"context_line":"if [ -z ${IMAGE_BLOCK_DEVICE} ] ; then"},{"line_number":440,"context_line":"    # For compatibily reasons in addition to the YAML configuration"}],"source_content_type":"application/x-shellscript","patch_set":1,"id":"3f79a3b5_b135d9de","line":437,"in_reply_to":"3f79a3b5_d059765c","updated":"2018-10-19 05:11:28.000000000","message":"So what I\u0027m actually trying to do is to use docker for downloading large docker images of unknown size. Further I couldn\u0027t get the docker daemon to work within any of the chrooted phases so I do it in the block-device phase using bwrap and the docker daemon installed into the target. The problem here is that the image size is calculated before that and the only workaround I found so far was to re-calculate the needed image size and creating the block device by myself (actually using exactly the same code as this is done originally.\n\nNow thinking of that, what do you think about evaluating DIB_IMAGE_SIZE instead of IMAGE_BLOCK_DEVICES as this is the actual thing I need to overwrite. That would eliminate my need to create the block device by myself.","commit_id":"194f35473f78e4bb0a80f2750a2a187d3cb0c1e7"}]}