Thanks Jun for your clarification. I think one of the advantages for MXNet is
it's language binding. Although the number of customers using Scala is great
less than Python, it is still one important language we need to support.
About the first question, I would say we all in. At least all
We should reach an agreement on the responsibility/ownership before moving
forward.
1. Who will take the ownership of fixing Scala build failure if an operator
is changed/added in C++?
2. What is the expected turn around time of fixing the scala build failure.
3. What if we are short of resources
Hi All,
Thank you all for your feedback. This changes do influence a lot on the PRs
related to operator changes. I will take what Marco proposed to place that in
the nightly build rather than every CI we runs.
Thanks,
Qing
On 6/20/18, 8:40 PM, "Hao Jin" wrote:
I don't think we should
I don't think we should keep this hash check thing. As Qing stated before
on this thread, if there's any change in documentation the hash value is
also changed and then flagged as a problem, how will that break any user's
code? I would go for something like Marco's proposal: moving this to an
Agreed, for new APIs we can turn into a nightly job and report on dev@. The
goal here is not to burden anyone but to catch changes on the backend
before it propagates and break user's code and co-ordinate changes across
language bindings which IMO is essential, so I would like to keep for
changes
we should have a betting pool on what the current coverage is.
On Wed, Jun 20, 2018 at 7:54 PM Naveen Swamy wrote:
> +1 to collect data on coverage.
>
> On Wed, Jun 20, 2018 at 7:38 PM, Marco de Abreu <
> marco.g.ab...@googlemail.com.invalid> wrote:
>
> > Hello,
> >
> > for now, this is only a
I think we should go with an asychronous approach using a nightly job that
detects the changes and reports them accordingly. We could then send out a
report if there is a mismatch.
I agree with the already stated points that we should not put the burden of
adding frontend API bindings to somebody
+1 to collect data on coverage.
On Wed, Jun 20, 2018 at 7:38 PM, Marco de Abreu <
marco.g.ab...@googlemail.com.invalid> wrote:
> Hello,
>
> for now, this is only a test for myself and a way to gather data in order
> to give us the possibility to make a data-driven decision. So far, we have
> not
I understand the concerns here. However the problem here is that since the
Scala APIs are generated using Macros and we do not(may not ever) have full
test coverage on each of the APIs, we will not know for example if an
operator/API changes on the backend and that propagates to Scala users,
their
Hi Qing,
What is the exact risk of changing / adding operators? could you
provide an example? I also feel the way you proposed is little bit too
heavy to developers, and not quite friendly to new contributors.
On Wed, Jun 20, 2018 at 10:22 AM Haibin Lin wrote:
>
> I appreciate the effort and
While I think test coverage is a nice information to have. I would object
to using this as a metric to decide whether a PR should be merged.
Code-cov act as a mere coverage of APIs, which is a useful aspect, it can
be misleading in many cases, especially when such change involves
cross-language
Hello,
I'd like to introduce test coverage metrics of PRs using https://codecov.io/.
This tool will aggregate coverage reports across multiple runs, platforms
and technologies and gives contributors as well as reviewers a new tool
that allows to improve the quality of a pull request.
Since we
+1
On Wed, Jun 20, 2018 at 8:43 AM Steffen Rochel
wrote:
> I had a discussion yesterday with Jun Wu (wujun@gmail.com) to get a
> better understanding about the concerns raised, that users might get
> confused and maintenance efforts.
> I agree with Jim that building and fostering the
+1
On Mon, Jun 18, 2018, 6:52 PM Anirudh wrote:
> Hi,
>
> This is the vote to release Apache MXNet (incubating) version 1.2.1. Voting
> will start now and close Thursday June 21st 7:00 PM PDT.
>
> Link to release candidate 1.2.1.rc0:
>
>
I appreciate the effort and understand the motivation. However, I'm
concerned that it basically means merging operator PRs becomes sequential.
Developers who work on operators have to update their PR every time a new
operator is merged to master, the burden becomes significant if there're 20
Hi Macro,
Thanks for your feedback! I believe this should not be a block for contributor
in most of the cases.
Firstly, this would only be triggered if there is an operator changes (Only
general operators).
Secondly, it is simple to go through. Just need to change 1 line of code would
make
I had a discussion yesterday with Jun Wu (wujun@gmail.com) to get a
better understanding about the concerns raised, that users might get
confused and maintenance efforts.
I agree with Jim that building and fostering the community is important.
First of all, I suggest we should be open minded
17 matches
Mail list logo