mirrors/git-annex
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions

Merge branch 'master' of ssh://git-annex.branchable.com

Browse source
This commit is contained in:
Joey Hess 2018-10-19 17:57:30 -04:00
commit 3180fa6fd6
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
3 changed files with 22 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
doc/design/external_special_remote_protocol/comment_36_f8081ce200700516efef61ec0ac86f86._comment Normal file
View file

@ -0,0 +1,8 @@
[[!comment format=mdwn
username="Ilya_Shlyakhter"
avatar="http://cdn.libravatar.org/avatar/1647044369aa7747829c38b9dcc84df0"
subject="comment 36"
date="2018-10-19T15:52:21Z"
content="""
What is the intended use case for CHECKURL-MULTI? Are there examples of external special remote implementations that use this response?
"""]]

13
doc/todo/option_to_add_user-specified_string_to_key/comment_2_915c1c31fcb23455aa11331cd34aa92d._comment Normal file
View file

@ -0,0 +1,13 @@
[[!comment format=mdwn
username="Ilya_Shlyakhter"
avatar="http://cdn.libravatar.org/avatar/1647044369aa7747829c38b9dcc84df0"
subject="comment 2"
date="2018-10-19T17:54:49Z"
content="""
Another use case for a user-defined string field in a key is, for URL or WORM keys, to include additional information unique to the object. E.g. for an s3:// URI (handled by an external special remote) could embed the ETag in the key. Or for a dx:// URI to include the DNAnexus file ID .
From [[internals/key_format]], it seems like adding -uXXXXXX- (with XXXXXX a user-specified string) would not break compatibility?
"""]]

1
doc/todo/preserve_file_extensions_in_WORM_and_URL_keys.mdwn Normal file
View file

@ -0,0 +1 @@
Right now, when computing a WORM key from a relative path or a URL key from a URL, if the original string is longer than a SHA256 checksum, its tail is replaced with its md5. Unfortunately, this eats up the file extension(s) at the end, causing the issues that \*E backends solve. It would be better to keep the tail of the path and replace the start or the middle with the md5, preserving extensions (as configured in annex.maxextensionlength) the same way \*E backends do. Maybe also, add a config option for the length beyond which the replacement-with-checksum happens?