)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"0dec6d2f6f248ac0779ee9a63c64e47aacb10205","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_3ba918e0","line":10,"updated":"2015-09-15 10:23:47.000000000","message":"The bug is light on information about where these arch names are coming from in the first place ? Before adding more remappings I\u0027d like to know where these non-matching names are from.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"459e4777154073cbc80b632ec936e32ac99bb83a","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_fb6e4095","line":10,"in_reply_to":"ba15a1d1_1bf454f0","updated":"2015-09-15 10:28:25.000000000","message":"Do you mean these are reported by the kernel uname() call then ?  If so, then you\u0027ll need to be patching libvirt itself too, as it only knows about the amd64 arch name canonicalization.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"08b508dda724f9d09f42207d10b56cae76333e0f","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_7b978990","line":10,"in_reply_to":"ba15a1d1_3891179a","updated":"2015-09-15 12:59:26.000000000","message":"I don\u0027t think it is openstack\u0027s job to cope with every OS vendors\u0027 own preferred terminology. You say this is \"just\" adding Ubuntu\u0027s terminology but this sets precedent to \"just\" add other vendor\u0027s own custom terminology too. This is not a sustainable long term approach. OpenStack has defined what it considers the canonical list of architectures to me, and people should just use those when talking to OpenStack. A big part of the work to convert openstack to use enums was to make it absolutely clear what expected values are permitted for each field. The current back-compat remappings that exist are just temporary workarounds to deal with legacy data from older versions of openstack and not intended to remain long term.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":935,"name":"James Page","email":"ringo.page@gmail.com","username":"james-page"},"change_message_id":"d911120cdc8d91709ae7a49c33f66a87889880e1","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_1bf454f0","line":10,"in_reply_to":"ba15a1d1_3ba918e0","updated":"2015-09-15 10:25:53.000000000","message":"These are the architecture names in Ubuntu/Debian for these four platforms.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":935,"name":"James Page","email":"ringo.page@gmail.com","username":"james-page"},"change_message_id":"1d7c5f8f71ddf5b0cb3a9631b4407523940e7d69","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_f19f3f4e","line":10,"in_reply_to":"ba15a1d1_5b13acb8","updated":"2015-09-15 11:20:05.000000000","message":"I\u0027ve added a bit more detail to the original bug report as well - hopefully that explains the driver for this change.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"64beaa6b1a156a65b2a8a61e89b774e8fe34e349","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_f122df00","line":10,"in_reply_to":"ba15a1d1_5b13acb8","updated":"2015-09-15 11:24:48.000000000","message":"Ok, so you\u0027re saying that this is to cope with the user uploading an image to glance and setting the wrong value in the architecture field. This is just user error and not something we need to work around - they should simply use one of the documented required values, and not expect Nova to try to correct their mistakes.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":935,"name":"James Page","email":"ringo.page@gmail.com","username":"james-page"},"change_message_id":"45b63dea676513c51b46d0982323c451c07aa0ca","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_3891179a","line":10,"in_reply_to":"ba15a1d1_f122df00","updated":"2015-09-15 12:36:15.000000000","message":"I don\u0027t think its unreasonable to map well know architecture naming used by distributions into something OpenStack understands - this is not \u0027every possible way\u0027 - its if the user makes a (somewhat understandable) choice of using the Ubuntu/Debian architecture name for their image.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"},{"author":{"_account_id":935,"name":"James Page","email":"ringo.page@gmail.com","username":"james-page"},"change_message_id":"1691253b8bb3df44950e97690148aada49766020","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Add canonicalization for non-x86 architectures"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Various non-x86 architectures fail to map to underlying libvirt"},{"line_number":10,"context_line":"architecture names depending on image configuration etc."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Add mappings for POWER and ARM based architectures to improve this"},{"line_number":13,"context_line":"situation:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":6,"id":"ba15a1d1_5b13acb8","line":10,"in_reply_to":"ba15a1d1_fb6e4095","updated":"2015-09-15 10:35:18.000000000","message":"Note quite; I think the archs returned from uname are ok; this is more todo with the way images get distributed with the standardized \u0027ubuntu\u0027 architecture names - this patch maps those into the correct versions for libvirt, which should match uname.","commit_id":"c9ba8633068953a07ff44e465709f417701beaa2"}],"nova/compute/arch.py":[{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"f86a40bf0684aba147a45335b3b8dbe94901e7b4","unresolved":false,"context_lines":[{"line_number":99,"context_line":"    PPC64LE,"},{"line_number":100,"context_line":"    PPCEMB,"},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"    S390,"},{"line_number":103,"context_line":"    S390X,"},{"line_number":104,"context_line":"    SH4,"},{"line_number":105,"context_line":"    SH4EB,"}],"source_content_type":"text/x-python","patch_set":5,"id":"1a4dcd0f_348001bf","line":102,"updated":"2015-08-11 15:23:09.000000000","message":"here","commit_id":"a74f6362f36ef0e76078ac155c2035e4d763467b"},{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"f86a40bf0684aba147a45335b3b8dbe94901e7b4","unresolved":false,"context_lines":[{"line_number":166,"context_line":"        newname \u003d ARMV7"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"    if newname \u003d\u003d \"arm64\":"},{"line_number":169,"context_line":"        newname \u003d AARCH64"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"    if not is_valid(newname):"},{"line_number":172,"context_line":"        raise exception.InvalidArchitectureName(arch\u003dname)"}],"source_content_type":"text/x-python","patch_set":5,"id":"1a4dcd0f_d448bdb5","line":169,"updated":"2015-08-11 15:23:09.000000000","message":"because you try to support platform other than x86, can you include others as well? \ne.g s390 already supported libvirt etc","commit_id":"a74f6362f36ef0e76078ac155c2035e4d763467b"},{"author":{"_account_id":935,"name":"James Page","email":"ringo.page@gmail.com","username":"james-page"},"change_message_id":"0a0f0993c71c42f6b0405b75b611e676b9141c4b","unresolved":false,"context_lines":[{"line_number":166,"context_line":"        newname \u003d ARMV7"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"    if newname \u003d\u003d \"arm64\":"},{"line_number":169,"context_line":"        newname \u003d AARCH64"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"    if not is_valid(newname):"},{"line_number":172,"context_line":"        raise exception.InvalidArchitectureName(arch\u003dname)"}],"source_content_type":"text/x-python","patch_set":5,"id":"ba15a1d1_1b7d346e","line":169,"in_reply_to":"1a4dcd0f_4823ae58","updated":"2015-09-15 10:24:34.000000000","message":"I\u0027d prefer not to add any further architectures under this change; its not definitive, but it does represent the archs we\u0027ve tested in Ubuntu.","commit_id":"a74f6362f36ef0e76078ac155c2035e4d763467b"},{"author":{"_account_id":11303,"name":"Markus Zoeller (markus_z)","email":"mzoeller@linux.vnet.ibm.com","username":"mzoeller"},"change_message_id":"14ae7e51191ec6f185cf463b5936835d6c642b5b","unresolved":false,"context_lines":[{"line_number":166,"context_line":"        newname \u003d ARMV7"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"    if newname \u003d\u003d \"arm64\":"},{"line_number":169,"context_line":"        newname \u003d AARCH64"},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"    if not is_valid(newname):"},{"line_number":172,"context_line":"        raise exception.InvalidArchitectureName(arch\u003dname)"}],"source_content_type":"text/x-python","patch_set":5,"id":"1a4dcd0f_4823ae58","line":169,"in_reply_to":"1a4dcd0f_d448bdb5","updated":"2015-08-11 16:18:00.000000000","message":"The libvirt architecture names for IBM z systems are \"s390x\" for 64 bit and \"s390\" for 31 bit.","commit_id":"a74f6362f36ef0e76078ac155c2035e4d763467b"}]}