mreyescdl
commented
3 years ago Tracing the request through Apache, UI and DB query shows that the "+" sign is present in the UI, but not used in the DB query. This results in a 404 response. The UI seems to convert the "+" sign to whitespace prior to querying the DB.
Apache
------
/api/presign-file/ark%3A%2F13030%2Fm5bs499p/6/producer%2FBaltic%2B6Regions_genera_sizes.xlsx?no_redirect=true
UI
--
/api/presign-file/ark:%2F13030%2Fm5bs499p/6/producer%2FBaltic+6Regions_genera_sizes.xlsx?no_redirect=true
DB Query
--------
SELECT `inv_files`.* FROM `inv_files` INNER JOIN `inv_versions` ON `inv_versions`.`id` = `inv_files`.`inv_version_id` INNER JOIN `inv_objects` ON `inv_objects`.`id` = `inv_files`.`inv_object_id` WHERE (inv_objects.ark = 'ark:/13030/m5bs499p') AND (inv_versions.number = '6') AND (inv_files.pathname = 'producer/Baltic 6Regions_genera_sizes.xlsx') ORDER BY `inv_files`.`id` ASC LIMIT 1;
terrywbrady
This is a sister ticket to Dryad #1116.
Regarding presigned file URL: http://merritt.cdlib.org/api/presign-file/ark%3A%2F13030%2Fm5kq31dt/1/producer%2FTDR%20Acc%2BH2O.zip?no_redirect=true Per @sfisher: There are a number of presigned URL problems from Merritt. It seems like most of these contain the character
+which is correctly encoded as%2Bin our request URLs like above. I believe Merritt is prematurely decoding %2B as a plus sign (which is a space in some query strings). However, %2B should be passed through as a character to the script that operates on the filenames, not decoded in an earlier layer and interpreted as a space.See for example https://www.w3schools.com/tags/ref_urlencode.ASP or https://tools.ietf.org/html/rfc3986 or many encoding/decoding libraries in languages such as Java or Ruby.
The above issue is believed to be a new one that surfaced after the recent Rails5 upgrade.
PS. It seems to be present in the Merritt UI, also. https://merritt.cdlib.org/m/ark%253A%252F13030%252Fm5kq31dt/1 and try to download
TDR Acc+H2O.zip