)]}'
{"src/erasurecode_helpers.c":[{"author":{"_account_id":597,"name":"Pete Zaitcev","email":"zaitcev@kotori.zaitcev.us","username":"zaitcev"},"change_message_id":"d06859b43497c8fc185b3f4c97741f396c00b529","unresolved":false,"context_lines":[{"line_number":223,"context_line":"    return buf;"},{"line_number":224,"context_line":"}"},{"line_number":225,"context_line":""},{"line_number":226,"context_line":"int _get_data_ptr_array_from_fragments(char **data_array, char **fragments,"},{"line_number":227,"context_line":"                                       int num_fragments)"},{"line_number":228,"context_line":"{"},{"line_number":229,"context_line":"    int i \u003d 0;"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"1fa4df85_13e4210c","line":226,"updated":"2020-03-10 22:29:02.000000000","message":"I disagree with this for a couple of reasons.\n\n1. The function _get_data_ptr_array_from_fragments is not used in liberasurecode, so it has to be purely an external API for the sake of Quadiron. Which is fine, but if so, it should not be named starting with an underscore.\n\n2. The return value is always the same as the parameter num_fragments. Is that intentional? Why even have a return value then?\n\nThe protocol of get_data_ptr_array_from_fragments is not particularly C-like. It fills returned array with NULL where fragments are missing, but returns the non-NULL count. But in this case, it differs. Please confirm that this is intentional.\n\nIn fact, please confirm that this is even used at all.","commit_id":"ea119b371fdf86b54b29de83e1fa605d2ead1ab1"}],"src/erasurecode_preprocessing.c":[{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"73f325c3ba7c8dc068deb475f0d0ae4187d675be","unresolved":false,"context_lines":[{"line_number":361,"context_line":""},{"line_number":362,"context_line":"    /* Copy fragment data into cstring (fragments should be in index order) */"},{"line_number":363,"context_line":"    for (i \u003d 0; i \u003c num_data \u0026\u0026 orig_data_size \u003e 0; i++) {"},{"line_number":364,"context_line":"        char* fragment_data \u003d _get_data_ptr_from_fragment(data[i]);"},{"line_number":365,"context_line":"        int fragment_size \u003d get_fragment_payload_size(data[i]);"},{"line_number":366,"context_line":"        int payload_size \u003d orig_data_size \u003e fragment_size ? fragment_size : orig_data_size;"},{"line_number":367,"context_line":"        memcpy(internal_payload + string_off, fragment_data, payload_size);"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_c68416d6","line":364,"updated":"2019-04-04 22:47:16.000000000","message":"I can\u0027t help but wonder whether this is going to mess with libphazr\u0027s expectations... the discussion at https://review.openstack.org/#/c/424353/5/src/erasurecode_preprocessing.c@72 makes me think this would break them.\n\nI\u0027ll add Jim Cheung as a reviewer, see if we can get him to test it.\n\nMaybe we need to differentiate between an encode data offset (for use in prepare_fragments_for_encode) and a decode one (to use here)...","commit_id":"ea119b371fdf86b54b29de83e1fa605d2ead1ab1"},{"author":{"_account_id":29916,"name":"vianney rancurel","email":"vianney.rancurel@gmail.com","username":"vrancurel"},"change_message_id":"60398d89d58f020f374eb871e1e4eebd809036f2","unresolved":false,"context_lines":[{"line_number":361,"context_line":""},{"line_number":362,"context_line":"    /* Copy fragment data into cstring (fragments should be in index order) */"},{"line_number":363,"context_line":"    for (i \u003d 0; i \u003c num_data \u0026\u0026 orig_data_size \u003e 0; i++) {"},{"line_number":364,"context_line":"        char* fragment_data \u003d _get_data_ptr_from_fragment(data[i]);"},{"line_number":365,"context_line":"        int fragment_size \u003d get_fragment_payload_size(data[i]);"},{"line_number":366,"context_line":"        int payload_size \u003d orig_data_size \u003e fragment_size ? fragment_size : orig_data_size;"},{"line_number":367,"context_line":"        memcpy(internal_payload + string_off, fragment_data, payload_size);"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_dc6db403","line":364,"in_reply_to":"5fc1f717_c68416d6","updated":"2019-04-06 01:02:00.000000000","message":"Libphazr shall benefit from these changes, but maybe they have to adapt their code.\n\nActually most of the code does not need the offset to the data. If we have a metadata header it is logical to provide a function to access the actual payload.","commit_id":"ea119b371fdf86b54b29de83e1fa605d2ead1ab1"}]}
