As noted as part of #54 (see comment), the preservation service was losing file sizes during a publishing update for files that were not being updated. Processing the updated POD file during an update was blowing away the size values (and other metadata extracted from the files, like the checksum). To fix this, the file-extracted metadata were written out as annotations to the metadata (which are protected from updates to the POD). As part of this, the code needed to be cleaned up where assumptions were made that annotations would not have such metadata.
As noted as part of #54 (see comment), the preservation service was losing file sizes during a publishing update for files that were not being updated. Processing the updated POD file during an update was blowing away the size values (and other metadata extracted from the files, like the checksum). To fix this, the file-extracted metadata were written out as annotations to the metadata (which are protected from updates to the POD). As part of this, the code needed to be cleaned up where assumptions were made that annotations would not have such metadata.