Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Another truncated sindex key bug #3444

Closed
mlucy opened this issue Dec 16, 2014 · 9 comments
Closed

Another truncated sindex key bug #3444

mlucy opened this issue Dec 16, 2014 · 9 comments
Assignees
@mlucy
Milestone
1.15.3

Comments

@mlucy
Copy link
Member

mlucy commented Dec 16, 2014

In next:

> def pop i; {id: i, c: 'a'*300 + ("%020d") % i}; end
> r.table('test').insert((0...1024).map{|i| pop(i)}).run
> r.table('test').max(index: 'c')['c'].run
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa\
00000000000000000009",
> r.table('test').max('c')['c'].run
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa\
00000000000000001023"

Oddly enough, if you use autogenerated keys the bug goes away, at least in this particular test case.

This is blocking the changefeed tests, so I'm going to look into this.
@mlucy mlucy self-assigned this Dec 16, 2014
@mlucy mlucy added this to the 1.16 milestone Dec 16, 2014
@danielmewes
Copy link
Member

1.15.x is also broken when using orderBy({index: r.desc("c")}).limit(1)

@danielmewes
Copy link
Member

Interestingly, both of these work:
1 forward ordering

r.table('test').orderBy({index: r.asc("c")}).limit(1)

2 Not using limit (interesting, could we do something wrong when batching?)

r.table('test').orderBy({index: r.desc("c")})

@danielmewes
Copy link
Member

Replacing the two batchspec.with_at_most(right - index) by just batchspec in slice_datum_stream_t::next_raw_batch() avoids the problem. It's obviously not a solution, but might help nailing down the root cause.

@danielmewes danielmewes modified the milestones: 1.15.3, 1.16 Dec 17, 2014
@danielmewes
Copy link
Member

Putting into 1.15.3. @mlucy and I have been working on a fix.
(@AtnNn)

@danielmewes danielmewes mentioned this issue Dec 17, 2014
Closed
40 tasks
danielmewes pushed a commit that referenced this issue Dec 17, 2014
Fix for #3444
a3ad5f6
@danielmewes
Copy link
Member

Sorry, used the wrong issue number in a commit. The commit that GitHub linked here actually doesn't fix this one.

@danielmewes
Copy link
Member

EDIT: Posted in the wrong thread. No fix yet.

@danielmewes
Copy link
Member

Gosh, not sure what I did yesterday with the different issues. There's no fix for this one yet (though we know how to fix it). CR 2400 is not about this one, but I meant to post that message into #3449 for which the fix in that CR actually is.

@danielmewes danielmewes mentioned this issue Dec 17, 2014
Closed
danielmewes pushed a commit that referenced this issue Dec 18, 2014
Fix and regression test for #3444
fc06e44
@danielmewes
Copy link
Member

Actual fix now in branch daniel_3444 and CR 2412 by @mlucy

@danielmewes
Copy link
Member

The fix is in v1.15.x and next fc06e44

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@mlucy mlucy

Labels
None yet
Projects
None yet
Milestone
1.15.3
Development

No branches or pull requests
2 participants
@danielmewes @mlucy