)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"c5f231d965d92efede5bdd3c9f88443209130af4","unresolved":true,"context_lines":[{"line_number":40,"context_line":"Case (2) and (3) are less reasonable. An empty etag value in the"},{"line_number":41,"context_line":"container listing is unexpected, and the encrypter has the plaintext"},{"line_number":42,"context_line":"etag available, so it seems more reasonable that the encrypter should"},{"line_number":43,"context_line":"replace an empty etag value with the encrypted plaintext etag."},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"All three cases make it harder to reason about the encrypter\u0027s"},{"line_number":46,"context_line":"behaviour with respect to the override-etag footer."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"b2c572c0_2391276b","line":43,"updated":"2025-02-26 18:32:56.000000000","message":"I just don\u0027t remember why we seem to have gone out of our way years ago to allow an imagined other middleware specifying ``overide-etag\u003d\u0027\u0027``.","commit_id":"f3083d32616b6ceb7027691f1283bf38eae8c883"},{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"c5f231d965d92efede5bdd3c9f88443209130af4","unresolved":true,"context_lines":[{"line_number":46,"context_line":"behaviour with respect to the override-etag footer."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This patch therefore modifies the encrypter middleware so that it"},{"line_number":49,"context_line":"*always* adds a non-empty override-etag footer. If the value of the"},{"line_number":50,"context_line":"override-etag footer or header provided by other middlewares is the"},{"line_number":51,"context_line":"empty string (cases 2 and 3), then the encrypter now sets the"},{"line_number":52,"context_line":"override-etag footer based on the plaintext etag. If no body has been"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"b1574692_3d108eac","line":49,"range":{"start_line":49,"start_character":0,"end_line":49,"end_character":46},"updated":"2025-02-26 18:32:56.000000000","message":"I suppose the piece I most care about is that the encrypter should always add an override-etag footerm (even if it is the empty string), if only to make it easier to reason about what is already a complex set of interactions between headers and footers.\n\nIt\u0027s just odd to think that we might send an empty etag to the listing. And odder to think that it might have params appended to it, like:\n\n``x-blah-blah-override-etag \u003d ;foo\u003dbar``","commit_id":"f3083d32616b6ceb7027691f1283bf38eae8c883"}]}