> Would't it be simpler to deal with the serialization issue by bumping the
> serialVersionID?
simpler yes, but it's a different thing
The PR makes the serialized forms for commons-csv versions 1.9.0 and 1.10.0
compatible.
Given that serialization has been broken for several versions in commons-
On Fri, 21 Oct 2022 at 11:57, wrote:
>
> > Would't it be simpler to deal with the serialization issue by bumping the
> > serialVersionID?
> simpler yes, but it's a different thing
> The PR makes the serialized forms for commons-csv versions 1.9.0 and 1.10.0
> compatible.
>
> Given that serializa
I updated git master with Javadoc to signal that Serialization in CSVFormat is
not supported from one version to the next. I also bumped the serial version ID
from 1 to 2. All of this is noted in changes.xml.
Gary
On 2022/10/21 11:27:43 sebb wrote:
> On Fri, 21 Oct 2022 at 11:57, wrote:
> >
>
On 2022/10/20 22:56:05 Alex Herbert wrote:
> On Thu, 20 Oct 2022 at 23:43, Alex Herbert wrote:
> >
> > I did not have time to track through whether this behaviour changed
> > after the initial implementation of the flag. I would think not as the
> > original behaviour is from 1.0. This would map t
I've removed serialization stuff from the PR and rebased it to master
https://github.com/apache/commons-csv/pull/276
kind regards,
Markus
From: Gary D. Gregory
Sent: Friday, October 21, 2022 14:18
To: dev@commons.apache.org
Subject: Re: [VOTE] Release Apache Commons CSV 1.10.0 based on RC1
O
Hi Markus,
The PR has conflicts and does not test its changes.
I'd like feedback from the community on whether or not git master as it is
now is OK for the duplicate header behavior or if changing the default is
needed and if doing so is just ping-pongonging from one incompatibility to
another, b
Hello Alex,
Ichecked the architecture of MultivariateOptimizer family to compareto what I
have done so far. I think that what I have done can berefactored to fit in the
general framework as extension ofGradientMultivariateOptimizer even though,
main differences are :
-
There is no ne
Hi Gary,
the PR did not have conflicts 30 minutes ago. So we must have encountered a
race condition between you updating master and me rebasing and force-pushing my
pr.
I've done that again just now. There are no conflicts.
At the same time I see parts of my PR have dribbled into master already,
Hello.
I haven't looked at the code, but thanks for your interest in contributing
to Commons Math.
The more straightforward path is indeed to adapt your code to what is
currently in the "...legacy" module. However, you can really consider
the other venue, i.e. a design specific to this family of
Ok, I don't see any conflict anymore but the PR title does not match the
code: "CSVFormat.duplicateHeaderMode requires default DISALLOW"
I am ok with the changes as they stand now but we have to resolve if the
default should be changed.
Gary
On Fri, Oct 21, 2022, 09:44 wrote:
> Hi Gary,
>
> th
>From my investigation the behaviour from 1.0 to 1.6 was to throw on
duplicates, but ignore empty duplicates (so allowing pad columns in a
CSV).
The allowDuplicates flag was added in 1.7 to allow duplicates. DEFAULT
behaviour was true! So the behavioural compatibility was broken in
release 1.7. No
11 matches
Mail list logo
