From d3532ebe2b6ebcdac37bc2eec66add55f39e4018 Mon Sep 17 00:00:00 2001 From: yarikoptic Date: Thu, 15 Dec 2022 01:24:55 +0000 Subject: [PATCH] initial TODO for making it possible to use find/findref without penalty --- ...dd_--all___40__or_alike__41___to_find_and_findref.mdwn | 8 ++++++++ 1 file changed, 8 insertions(+) create mode 100644 doc/todo/add_--all___40__or_alike__41___to_find_and_findref.mdwn diff --git a/doc/todo/add_--all___40__or_alike__41___to_find_and_findref.mdwn b/doc/todo/add_--all___40__or_alike__41___to_find_and_findref.mdwn new file mode 100644 index 0000000000..aa02cd747a --- /dev/null +++ b/doc/todo/add_--all___40__or_alike__41___to_find_and_findref.mdwn @@ -0,0 +1,8 @@ +Original discussion and need is within [datalad/issues/7038](https://github.com/datalad/datalad/issues/7038) -- we had been using `annex find` and `annex findref` with `--copies 0` to simply force those two commands to return information for all keys used in the current tree, regardless of their local presence. + +Unfortunately as that issue shows, adding `--copies 0` results in significant run time penalty over a run without it, or some different invocation with similar (albeit not as generic, since would miss keys without size) `--largerthan -1`. + +I found no other way to make `find` (or `findref`) to do what is needed. So I felt that may be it does need some additional option, or may be some special ad-hoc value for some of the possible "fast" expressions (like `--largerthan idonotcare`). + +[[!meta author=yoh]] +[[!tag projects/datalad]]