Final reminder that our CFP (Call For Presentations) ends on Friday 9/15.
This just needs to be a sentence or two about what you plan to present and
how much time you need. You can send your proposal to
summ...@trafficcontrol.incubator.apache.org. Let me know if you have any
questions.
Thanks!
D
+1 but the fact that only committers can be assigned to issues is a big
problem in my opinion.
On Tue, Aug 29, 2017 at 7:16 PM, Eric Friedrich (efriedri) <
efrie...@cisco.com> wrote:
> Thanks Leif-
> I’ll check with the Incubator gurus. We already pulled the trigger, so
> at this point its more
There has been quite a bit of discussion about how to move forward with the
Traffic Ops Rewrite in terms of Golang dependencies. Currently there is
only one dependency for Mocking out the database for unit testing called
https://github.com/DATA-DOG/go-sqlmock. Another that we want to evaluate is
h
After looking at the code, and given the work I've been doing with rewriting
the config file endpoints, I have to say sqlx all the way. What's involved in
the maintenance?
Derek
On Sep 12, 2017, at 8:28 PM, Dewayne Richardson
mailto:dewr...@gmail.com>> wrote:
There has been quite a bit of di
It would be maintaining about a 1500 line codebase (excluding tests with ~70%
coverage), it uses reflection and tag introspection so it isn’t the simplest go
code but it does seem to be well commented.
On 9/12/17, 6:36 PM, "Gelinas, Derek" wrote:
After looking at the code, and given the wo
I am a pretty big -1 on sqlx. Those PRs are extremely deceptive.
Those lines are entirely unnecessary.
I have created an example PR at https://github.com/apache/incu
bator-trafficcontrol/pull/1165
The relevant commits are
https://github.com/apache/incubator-trafficcontrol/pull/1165
/commits/6fc7
I'm also -1 on sqlx.
That said, the "single line" in the `sql` version is 674 characters long. I
think we can agree that it's one heck of a line.
We're focusing on the Scan call vs. the ScanStruct call, which I think is
the right place to look. Here's what `sql` requires:
- One identifier per col
@dewayne - you said "performance was +/-5% depending on the cloud resources
that were active". How many milliseconds difference (average) are we
talking about for ONE call to /api/1.2/servers (using sqlx and not using
sqlx)? I'm assuming it is so small that we can rule out performance as a
factor w
> Why the fear of using dependencies?
There's no fear of using dependencies. There's a recognition that
dependencies aren't free and a balancing of the costs they entail. In this
case, looking at the costs associate with `sqlx`, I don't think the
benefits justify it. The way we're using it, we cou
As one ready to jump in and add more endpoints, I'm a strong +1 on
using sqlx. I agree that adding a new dependency should not be done
without consideration, but I find the sqlx version much more readable
and easier to approach than either your or dew's version of non-sqlx
and would be much easi
@dan, how do you feel about the middle-ground I proposed: a
reflection-based function that would look like this:
rows.Scan(FieldsOf(&server)...) ?
On Tue, Sep 12, 2017 at 9:05 PM, Dan Kirkwood wrote:
> As one ready to jump in and add more endpoints, I'm a strong +1 on
> using sqlx. I agree tha
I’m +1 on using sqlx. We’ve done this dance, what 3 times now already? Let’s
just use it and move on.
Just my $0.02.
Rgds,
JvD
> On Sep 12, 2017, at 9:08 PM, Chris Lemmons wrote:
>
> @dan, how do you feel about the middle-ground I proposed: a
> reflection-based function that would look like
12 matches
Mail list logo
