Add randomly generated slug to container names to prevent collisions #6140
Conversation
shin-
force-pushed
the
4688_no_sequential_ids
branch
2 times, most recently
from
69be0a3
to
554eca2
Compare
shin-
force-pushed
the
4688_no_sequential_ids
branch
from
554eca2
to
06fb161
Compare
shin-
force-pushed
the
4688_no_sequential_ids
branch
3 times, most recently
from
3a19eeb
to
1165301
Compare
Signed-off-by: Joffrey F <joffrey@docker.com>
shin-
force-pushed
the
4688_no_sequential_ids
branch
from
ae3f9bf
to
70654cf
Compare
shin-
changed the title
|
Relating to #1516 , this fixes concerns 2 and 3. The costly queries are unfortunately still there, but this PR is a compromise intended to cause minimal disruption to existing workflows. While this can be seen as a breaking change in the sense that container names can no longer be inferred the way they could before, it should be noted that
|
| @@ -151,3 +152,21 @@ def unquote_path(s): | |||
| if s[0] == '"' and s[-1] == '"': | |||
| return s[1:-1] | |||
| return s | |||
|
|
|||
|
|
|||
| def generate_random_id(): | |||
There was a problem hiding this comment.
Why do we use such a complicated algorithm to generate a random identifier? And is there any specific reason we don't want a fully numerical ID?
Otherwise uuid.uuid4().hex sounds good to generate a hexadecimal identifier, no?
There was a problem hiding this comment.
The code is a transposition of https://github.com/moby/moby/blob/master/pkg/stringid/stringid.go ; we could go with a UUID instead.
tests/acceptance/cli_test.py
Outdated
| assert 'Removing v2-full_web_' in result.stderr | ||
| assert 'Removing v2-full_other_' in result.stderr | ||
| assert 'Removing v2-full_web_run_' in result.stderr | ||
| assert 'Removing v2-full_web_run_' in result.stderr |
Signed-off-by: Joffrey F <joffrey@docker.com>
shin-
force-pushed
the
4688_no_sequential_ids
branch
from
70654cf
to
5916639
Compare
Fixes #4688