Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 12:06
    jatinderaujla synchronize #3679
  • 12:04
    jatinderaujla ready_for_review #3679
  • 07:37
    jatinderaujla opened #3679
  • 07:00
    jatinderaujla closed #3678
  • 06:58
    jatinderaujla edited #3678
  • 06:54
    jatinderaujla ready_for_review #3678
  • 06:31
    jatinderaujla edited #3678
  • 06:24
    jatinderaujla opened #3678
  • Jul 03 03:19
    vivekwassan synchronize #3657
  • Jul 03 03:17
    vivekwassan synchronize #3657
  • Jul 03 03:16
    vivekwassan synchronize #3657
  • Jul 01 20:32
    roldanbob edited #3666
  • Jul 01 20:31
    roldanbob edited #3666
  • Jul 01 16:49
    Naros closed #3677
  • Jul 01 15:49
    MartinMedek synchronize #3545
  • Jul 01 15:20
    MartinMedek synchronize #3545
  • Jul 01 15:02
    MartinMedek synchronize #3545
  • Jul 01 14:06
    MartinMedek synchronize #3545
  • Jul 01 13:57
    MartinMedek synchronize #3545
  • Jul 01 13:53
    MartinMedek synchronize #3545
Jiri Pechanec
@jpechane
decimals were also done in stages
Gunnar Morling
@gunnarmorling
right
and about keeping support for float as a means of compat?
Jiri Pechanec
@jpechane
I don't think so
It is incubating mode
Gunnar Morling
@gunnarmorling
well, MicroDuration is used by PG
Jiri Pechanec
@jpechane
So this is a fair change
Gunnar Morling
@gunnarmorling
This message was deleted
For cassandra and NanoDuration, no doubt
Jiri Pechanec
@jpechane
Damn you are right
Gunnar Morling
@gunnarmorling
as said, i'd still "just" change it and see how it goes
we'd document it in the upgrade notes
Jiri Pechanec
@jpechane
I've missed that you are talking about two types
Gunnar Morling
@gunnarmorling
ah, sorry
if it's really a blocker, people always can add an SMT as a stop-gap measure
so MicroDuration existed before
Jiri Pechanec
@jpechane
Yes
Gunnar Morling
@gunnarmorling
but Cassandra has nano resolution
so we decided to add NanoDuration
and this started of this discussion
ok, i'll send a PR then
about the mapping in general, i'm not very convinced of it
as it's always an estimation
durations are given like "1 month, 2 days ..."
Jiri Pechanec
@jpechane
@gunnarmorling I propose to log the Jira for new precise duration
Gunnar Morling
@gunnarmorling
so any conversion to microseconds is prone to slight differences
Jiri Pechanec
@jpechane
That will use seconds + nanos as complex type
Gunnar Morling
@gunnarmorling
in the end, one would need to have a string mapping that describes the months, days and so on
Jiri Pechanec
@jpechane
So if somebody raises the issue we can just redirect him to it
Gunnar Morling
@gunnarmorling
i don't think that's enough
you cannot really convert a month to seconds
Jiri Pechanec
@jpechane
@gunnarmorling Well, the problem is that interval and duration are different types
And correctly we should support both
Gunnar Morling
@gunnarmorling
yes, right
Jiri Pechanec
@jpechane
As Oracle's duration is real duration
Gunnar Morling
@gunnarmorling
anyways, i'd expand that all driven by user demand
those types seem not to be that commonly used
Jiri Pechanec
@jpechane
Well, I think the logged Jira will not do any harm
There can always be a volunteer who'd like to take care of it
Gunnar Morling
@gunnarmorling
i don't like JIRA issues we don't have real plans to tackle
we already have many of those
makes it hard to focus on the more actionable ones
Jiri Pechanec
@jpechane
And are you sure we will remeber this discussion 6 month from now?
As we've come up with a preliminary design
Gunnar Morling
@gunnarmorling
grrr
Jiri Pechanec
@jpechane
@gunnarmorling Or add it as a comment to the DBZ-1497
There will not be Jira
Gunnar Morling
@gunnarmorling
too late :)
Jiri Pechanec
@jpechane
Sorry, I tried to come up with a compromise :-)