Right now we "ask" users to compute the checksums themselves, in order to be able to check if we have some content.
We should also allow users to upload their own files (within reasonable limits…) and verify if we have them or not.
Description
Description
Status | Assigned | Task | ||
---|---|---|---|---|
Migrated | gitlab-migration | T32 web UI: checksum search | ||
Migrated | gitlab-migration | T48 upload-based content search |
Event Timeline
Comment Actions
For now, upload the file, compute its sha1, dismiss the content, check if sha1 in swh storage exists.
Answers the query.
Sample:
$ curl -X POST -F filename=@/path/to/file http://localhost:6543/api/1/uploadnsearch/ { "found": false, "sha1": "e95097ad2d607b4c89c1ce7ca1fef2a1e4450558" }%
(code 200)
Regarding the limit, olasd and i spoke of apache's settings for this.
In the mean time, it's flask which limits it.
This is a parameter in swh-web-ui's of 16Mb for now.
(Throws a 413 entity too large if too much -> tested with a tarball of 21Mb)