You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@eagle.apache.org by Julian Hyde <jh...@apache.org> on 2016/07/15 17:07:03 UTC
Why?
I am seeing a few JIRA cases which are basically just check-in comments. They are created just before the commit, they explain what was done in the commit, do not explain why, do not link to any previous or future work.
An example of this is EAGLE-378. It arrives a couple of days after I had a conversation with Michael [1] about cleaning up included jars, yet it seems to be doing exactly the opposite.
Is the Eagle project operating commit-then-review or review-then-commit? It seems to be operating commit-then-review, but if so, there’s not enough information in the public record for people to review what is happening.
As my math teacher used to say: don’t just write down the answer, you need to show your working!
Julian
[1] https://issues.apache.org/jira/browse/EAGLE-378 <https://issues.apache.org/jira/browse/EAGLE-378>
[2] https://issues.apache.org/jira/browse/EAGLE-377 <https://issues.apache.org/jira/browse/EAGLE-377>
Re: Why?
Posted by Michael Wu <mc...@gmail.com>.
Sure, Julian, I'll take notice of your reminding. Thank you very much!
Michael
On Mon, Jul 18, 2016 at 4:16 PM, Julian Hyde <jh...@apache.org> wrote:
> Thanks for the explanation. But I hope you took note of my main point:
> code changes should be accompanied by sufficient explanation so that the
> community can review/understand them. We shouldn’t have to ask for an
> explanation.
>
> Making code changes transparent helps to build community.
>
> > On Jul 17, 2016, at 7:26 PM, Michael Wu <mc...@gmail.com> wrote:
> >
> > @Julian,
> >
> > it does seem confusing in this special case, let me explain why:
> > 1. at the time we had the discussion, 0.4.0-incubating PPMC vote had not
> > finished, and therefore I was not quite sure if it's the appropriate
> moment
> > to create the patch for .jar files. So, that's why EAGLE-378 came later
> > than EAGLE-377.
> > 2. conventionally, we used to make a pull request bound to a jira ticket,
> > so that's why I created the ticket just before the commit. If this is
> not a
> > good manner, I will try to avoid it in the future.
> > 3. EAGLE-377 was created based on the result of our discussion, and
> > pointing to 0.5.0 as its "fix version", which could be seen as a bug-fix
> > plan for 0.5.0.
> >
> > Thanks for pointing the problem out.
> >
> > Michael
> >
> > On Sat, Jul 16, 2016 at 1:28 AM, Edward Zhang <yonzhang2012@apache.org
> <ma...@apache.org>>
> > wrote:
> >
> >> I feel the same. Eagle project today needs more discussion in Eagle dev
> DL.
> >> I do see many discussions and code reviews within individual emails
> instead
> >> of going through Eagle dev DL. And some users also ask questions to
> >> individual email directly :-)
> >>
> >> Could I suggest Eagle committers and community please discuss important
> >> plans and issues in Eagle dev DL to have public record for people to
> review
> >> at any time?
> >>
> >> Thanks
> >> Edward
> >>
> >>
> >> On Fri, Jul 15, 2016 at 10:07 AM, Julian Hyde <jhyde@apache.org
> <ma...@apache.org>> wrote:
> >>
> >>> I am seeing a few JIRA cases which are basically just check-in
> comments.
> >>> They are created just before the commit, they explain what was done in
> >> the
> >>> commit, do not explain why, do not link to any previous or future work.
> >>>
> >>> An example of this is EAGLE-378. It arrives a couple of days after I
> had
> >> a
> >>> conversation with Michael [1] about cleaning up included jars, yet it
> >> seems
> >>> to be doing exactly the opposite.
> >>>
> >>> Is the Eagle project operating commit-then-review or
> review-then-commit?
> >>> It seems to be operating commit-then-review, but if so, there’s not
> >> enough
> >>> information in the public record for people to review what is
> happening.
> >>>
> >>> As my math teacher used to say: don’t just write down the answer, you
> >> need
> >>> to show your working!
> >>>
> >>> Julian
> >>>
> >>> [1] https://issues.apache.org/jira/browse/EAGLE-378 <
> >>> https://issues.apache.org/jira/browse/EAGLE-378 <
> https://issues.apache.org/jira/browse/EAGLE-378>>
> >>>
> >>> [2] https://issues.apache.org/jira/browse/EAGLE-377 <
> https://issues.apache.org/jira/browse/EAGLE-377> <
> >>> https://issues.apache.org/jira/browse/EAGLE-377 <
> https://issues.apache.org/jira/browse/EAGLE-377>>
>
>
Re: Why?
Posted by Julian Hyde <jh...@apache.org>.
Thanks for the explanation. But I hope you took note of my main point: code changes should be accompanied by sufficient explanation so that the community can review/understand them. We shouldn’t have to ask for an explanation.
Making code changes transparent helps to build community.
> On Jul 17, 2016, at 7:26 PM, Michael Wu <mc...@gmail.com> wrote:
>
> @Julian,
>
> it does seem confusing in this special case, let me explain why:
> 1. at the time we had the discussion, 0.4.0-incubating PPMC vote had not
> finished, and therefore I was not quite sure if it's the appropriate moment
> to create the patch for .jar files. So, that's why EAGLE-378 came later
> than EAGLE-377.
> 2. conventionally, we used to make a pull request bound to a jira ticket,
> so that's why I created the ticket just before the commit. If this is not a
> good manner, I will try to avoid it in the future.
> 3. EAGLE-377 was created based on the result of our discussion, and
> pointing to 0.5.0 as its "fix version", which could be seen as a bug-fix
> plan for 0.5.0.
>
> Thanks for pointing the problem out.
>
> Michael
>
> On Sat, Jul 16, 2016 at 1:28 AM, Edward Zhang <yonzhang2012@apache.org <ma...@apache.org>>
> wrote:
>
>> I feel the same. Eagle project today needs more discussion in Eagle dev DL.
>> I do see many discussions and code reviews within individual emails instead
>> of going through Eagle dev DL. And some users also ask questions to
>> individual email directly :-)
>>
>> Could I suggest Eagle committers and community please discuss important
>> plans and issues in Eagle dev DL to have public record for people to review
>> at any time?
>>
>> Thanks
>> Edward
>>
>>
>> On Fri, Jul 15, 2016 at 10:07 AM, Julian Hyde <jhyde@apache.org <ma...@apache.org>> wrote:
>>
>>> I am seeing a few JIRA cases which are basically just check-in comments.
>>> They are created just before the commit, they explain what was done in
>> the
>>> commit, do not explain why, do not link to any previous or future work.
>>>
>>> An example of this is EAGLE-378. It arrives a couple of days after I had
>> a
>>> conversation with Michael [1] about cleaning up included jars, yet it
>> seems
>>> to be doing exactly the opposite.
>>>
>>> Is the Eagle project operating commit-then-review or review-then-commit?
>>> It seems to be operating commit-then-review, but if so, there’s not
>> enough
>>> information in the public record for people to review what is happening.
>>>
>>> As my math teacher used to say: don’t just write down the answer, you
>> need
>>> to show your working!
>>>
>>> Julian
>>>
>>> [1] https://issues.apache.org/jira/browse/EAGLE-378 <
>>> https://issues.apache.org/jira/browse/EAGLE-378 <https://issues.apache.org/jira/browse/EAGLE-378>>
>>>
>>> [2] https://issues.apache.org/jira/browse/EAGLE-377 <https://issues.apache.org/jira/browse/EAGLE-377> <
>>> https://issues.apache.org/jira/browse/EAGLE-377 <https://issues.apache.org/jira/browse/EAGLE-377>>
Re: Why?
Posted by Michael Wu <mc...@gmail.com>.
@Julian,
it does seem confusing in this special case, let me explain why:
1. at the time we had the discussion, 0.4.0-incubating PPMC vote had not
finished, and therefore I was not quite sure if it's the appropriate moment
to create the patch for .jar files. So, that's why EAGLE-378 came later
than EAGLE-377.
2. conventionally, we used to make a pull request bound to a jira ticket,
so that's why I created the ticket just before the commit. If this is not a
good manner, I will try to avoid it in the future.
3. EAGLE-377 was created based on the result of our discussion, and
pointing to 0.5.0 as its "fix version", which could be seen as a bug-fix
plan for 0.5.0.
Thanks for pointing the problem out.
Michael
On Sat, Jul 16, 2016 at 1:28 AM, Edward Zhang <yo...@apache.org>
wrote:
> I feel the same. Eagle project today needs more discussion in Eagle dev DL.
> I do see many discussions and code reviews within individual emails instead
> of going through Eagle dev DL. And some users also ask questions to
> individual email directly :-)
>
> Could I suggest Eagle committers and community please discuss important
> plans and issues in Eagle dev DL to have public record for people to review
> at any time?
>
> Thanks
> Edward
>
>
> On Fri, Jul 15, 2016 at 10:07 AM, Julian Hyde <jh...@apache.org> wrote:
>
> > I am seeing a few JIRA cases which are basically just check-in comments.
> > They are created just before the commit, they explain what was done in
> the
> > commit, do not explain why, do not link to any previous or future work.
> >
> > An example of this is EAGLE-378. It arrives a couple of days after I had
> a
> > conversation with Michael [1] about cleaning up included jars, yet it
> seems
> > to be doing exactly the opposite.
> >
> > Is the Eagle project operating commit-then-review or review-then-commit?
> > It seems to be operating commit-then-review, but if so, there’s not
> enough
> > information in the public record for people to review what is happening.
> >
> > As my math teacher used to say: don’t just write down the answer, you
> need
> > to show your working!
> >
> > Julian
> >
> > [1] https://issues.apache.org/jira/browse/EAGLE-378 <
> > https://issues.apache.org/jira/browse/EAGLE-378>
> >
> > [2] https://issues.apache.org/jira/browse/EAGLE-377 <
> > https://issues.apache.org/jira/browse/EAGLE-377>
>
Re: Why?
Posted by Edward Zhang <yo...@apache.org>.
I feel the same. Eagle project today needs more discussion in Eagle dev DL.
I do see many discussions and code reviews within individual emails instead
of going through Eagle dev DL. And some users also ask questions to
individual email directly :-)
Could I suggest Eagle committers and community please discuss important
plans and issues in Eagle dev DL to have public record for people to review
at any time?
Thanks
Edward
On Fri, Jul 15, 2016 at 10:07 AM, Julian Hyde <jh...@apache.org> wrote:
> I am seeing a few JIRA cases which are basically just check-in comments.
> They are created just before the commit, they explain what was done in the
> commit, do not explain why, do not link to any previous or future work.
>
> An example of this is EAGLE-378. It arrives a couple of days after I had a
> conversation with Michael [1] about cleaning up included jars, yet it seems
> to be doing exactly the opposite.
>
> Is the Eagle project operating commit-then-review or review-then-commit?
> It seems to be operating commit-then-review, but if so, there’s not enough
> information in the public record for people to review what is happening.
>
> As my math teacher used to say: don’t just write down the answer, you need
> to show your working!
>
> Julian
>
> [1] https://issues.apache.org/jira/browse/EAGLE-378 <
> https://issues.apache.org/jira/browse/EAGLE-378>
>
> [2] https://issues.apache.org/jira/browse/EAGLE-377 <
> https://issues.apache.org/jira/browse/EAGLE-377>