uppercase values in CHECKSUM attribute in descriptor cause checksum validation to fail

[lydiam]

UF, and possibly other Affiliates, sometimes send descriptors that contain the MD5 CHECKSUM values in uppercase. This appears to cause checksum validation problems in the submission service and the SIP is rejected for failing checksum validation.

An example can be found in retsina:/var/daitss2/snafu_wips/E000019HP_DNG8DP, although WIP-creation seems to lowercase the checksum originally found in the descriptor.

From /var/daitss2/snafu_wips/E000019HP_DNG8DP/metadata: CHECKSUM="1e8415f2a5157cda51ef9601c04af8f8"

[daitss@retsina 0]$ md5sum data 1e8415f2a5157cda51ef9601c04af8f8 data [daitss@retsina 0]$ pwd /var/daitss2/snafu_wips/E000019HP_DNG8DP/files/original/0

The original SIP descriptor: /var/daitss2/test-packages/UFE0015542B

[daitss@retsina UFE0015542B]$ grep CHECKSUM *xml

[cchou]

Perhaps submission can check checksum case insensitively.

[lydiam]

I may not have the correct example in this ticket: I had two versions of the same package, one with uppercase checksum and one with lowercase, and I believe that my example looks at the lowercased WIP. Bad WIP example, but SIP UFE0015542B is a good example of a SIP with an uppercase CHECKSUM.

[lydiam]

The reject message from the above package: MD5 for westwood_g.pdf - expected: 1E8415F2A5157CDA51EF9601C04AF8F8; computed 1e8415f2a5157cda51ef9601c04af8f8

[flazz]

please continue this here: http://github.com/daitss/core/issues/issue/44