Thread (28 messages) flat view 28 messages, 2 authors, 2018-06-09

Re: [PATCH 20/20] abbrev: add a core.validateAbbrev setting

From: Martin Ågren <hidden>
Date: 2018-06-09 15:47:26

On 9 June 2018 at 00:41, Ævar Arnfjörð Bjarmason [off-list ref] wrote:
Instead of trying really hard to find an unambiguous SHA-1 we can with
core.validateAbbrev=false, and preferably combined with the new
support for relative core.abbrev values (such as +4) unconditionally
print a short SHA-1 without doing any disambiguation check. I.e. it
This first paragraph read weirdly the first time. On the second attempt
I knew how to structure it and got it right. It might be easier to read
if the part about core.appreb=+4 were in a separate second sentence.

That last "it" is "the combination of these two configs", right?
allows for picking a trade-off between performance, and the odds that
future or remote (or current and local) short SHA-1 will be ambiguous.
quoted hunk ↗ jump to hunk
diff --git a/Documentation/config.txt b/Documentation/config.txt
index abf07be7b6..df31d1351f 100644
--- a/Documentation/config.txt
+++ b/Documentation/config.txt
@@ -925,6 +925,49 @@ means to add or subtract N characters from the SHA-1 that Git would
 otherwise print, this allows for producing more future-proof SHA-1s
 for use within a given project, while adjusting the value for the
 current approximate number of objects.
++
+This is especially useful in combination with the
+`core.validateAbbrev` setting, or to get more future-proof hashes to
+reference in the future in a repository whose number of objects is
+expected to grow.
Maybe s/validateAbbrev/validateAbbrev = false/?
+
+core.validateAbbrev::
+       If set to false (true by default) don't do any validation when
+       printing abbreviated object names to see if they're really
+       unique. This makes printing objects more performant at the
+       cost of potentially printing object names that aren't unique
+       within the current repository.
Good. I understand why I'd want to use it, and why not.
++
+When printing abbreviated object names Git needs to look through the
+local object store. This is an `O(log N)` operation assuming all the
+objects are in a single pack file, but `X * O(log N)` given `X` pack
+files, which can get expensive on some larger repositories.
This might be very close to too much information.
++
+This setting changes that to `O(1)`, but with the trade-off that
+depending on the value of `core.abbrev` we may be printing abbreviated
+hashes that collide. Too see how likely this is, try running:
++
+-----------------------------------------------------------------------------------------------------------
+git log --all --pretty=format:%h --abbrev=4 | perl -nE 'chomp; say length' | sort | uniq -c | sort -nr
+-----------------------------------------------------------------------------------------------------------
++
+This shows how many commits were found at each abbreviation length. On
+linux.git in June 2018 this shows a bit more than 750,000 commits,
+with just 4 needing 11 characters to be fully abbreviated, and the
+default heuristic picks a length of 12.
These last few paragraphs seem like too much to me.
++
+Even without `core.validateAbbrev=false` the results abbreviation
+already a bit of a probability game. They're guaranteed at the moment
+of generation, but as more objects are added, ambiguities may be
+introduced. Likewise, what's unambiguous for you may not be for
+somebody else you're communicating with, if they have their own clone.
This seems more useful.
++
+Therefore the default of `core.validateAbbrev=true` may not save you
+in practice if you're sharing the SHA-1 or noting it now to use after
+a `git fetch`. You may be better off setting `core.abbrev` to
+e.g. `+2` to add 2 extra characters to the SHA-1, and possibly combine
+that with `core.validateAbbrev=false` to get a reasonable trade-off
+between safety and performance.
Makes sense. As before, I'd suggest s/SHA-1/object ID/.

I do wonder if this documentation could be a bit less verbose without
sacrificing too much correctness and clarity. No brilliant suggestions
on how to do that, sorry.

Martin
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help