weixin_39772566的博客

dispatch_set_context looks as if it is available in MacOS X 10.6, unlike dispatch_queue_set_specific, which mandated 10.7+.

And the same applies to colSpan and rowSpan in the internal views, having the chance to set them once the view is created but before it is added rather than having to modify the static XML attribute.

Definitely, even if it required a following invalidate-like method. You can have both pointing to the same field regardless, and allow definition to the user.

It's worth saying that whatever the case, most of my SpannableGrids (in their current incarnation based off ListView) have their size/ratio values assigned at runtime due to remote configuration, so keeping it as an attribute requires a bit more work from my part.

If you can move it out of an attribute and into a configurable field, I'd go with exact dimension because as a user I can calculate the measures myself. Otherwise, aspect ratio as an attribute.

Thank you for the fast response.

Shamelessly bumping this feat request because my project is on hold until it arrives.

What is the ETA on Sonatype update? Thank you.

When do you expect to release this guy?

Please approve this guy!!!!!

PLEASE!!!!!!!!

Apologies for the late response, I seem to have notifications disabled. Enabled for this issue.

The location where the Exception is thrown is very isolated so you'd have to let it throw a subclass of OAuthException, catch it in OAuth10aServiceImpl and throw another subclass of OAuthException in order to provide users access to the response/response code.

I was thinking about replacing the RequestTuner with a listener which can do more than some very limited tuning. (like adding headers before adding a body). What do you think?

Fixed :grin:

Seriously? I pushed it that same day. Will check when I get to a computer. On Wed, Feb 25, 2015 at 8:13 PM Steve Klabnik notifications.com wrote:

How odd, I thought that Bundler handled that for me.

— Reply to this email directly or view it on GitHub https://github.com/rails-api/active_model_serializers/issues/786#issuecomment-76082307 .

yay. Good to know.

Merged :)

Thanks for pointing that out. I forgot to push. Will do once I'm back to a computer. On Wed, Jan 21, 2015 at 8:19 PM Benjamin Fleischer notifications.com wrote:

See v0.9.2...0.9.3 https://github.com/rails-api/active_model_serializers/compare/v0.9.2...0.9.3

(i.e. it's 0.9.3 not v0.9.3).

and 0-9-stable...0.9.3 https://github.com/rails-api/active_model_serializers/compare/0-9-stable...0.9.3

cc https://github.com/kurko

ref: https://rubygems.org/gems/active_model_serializers/versions/0.9.3

— Reply to this email directly or view it on GitHub https://github.com/rails-api/active_model_serializers/issues/786.

it will depend on the program. In the 10 or so programs I'm involved with, the program reads a configuration file (JSON) and may optionally read/write a different state file (JSON). The rest of the time, the program is servicing requests via nanomsg, websockets or small http servers - which also uses cJSON when possible 😄.

I've written file io bindings for FatFs + Freertos. Non-trivial, which is why I think this group of features/functions should be configurable with compile switches (Ideally to simply not include the file).

There's already sprintf() dependency, which (having implemented both file io & sprintf() in embedded environments) is much more work.

I've been finding that I write lots of boiler plate code that is duplicated when I use cJSON for 1 extremely common pattern: read a json file into a json struct, then later write a struct out.

Even if this set of helper functions went into a different header (cJSON_Utils.h for example) & was conditionally compiled out (for systems that don't want/need it), it would be a benefit to the community at large as it helps users adopt cJSON with less friction. This also provides a path to support streaming files later.

My comment is meant to apply to #104, too.

Best, Wes

Having the same issue. But added to that when i installed OJ, serialization times doubled. Which was due to OJ not being set for rails_optimize. But then can't activate Rails optimizations as OJ (3.5) segfaults with that when encountering Time objects before 1970. So i have either wrongly formatted date strings in the JSON output, horribly slow serialization or a segfaulting ruby.

You would need to upgrade to Rails (~> 5.0) first as ActiveSupport versions are tied to the Rails version you're running.

Short: no.

I would love for someone to create this, but we don't use gradle at FB, so we've got no expertise to do it at the moment.

Just for curiosity, what do you guys use at FB instead of Gradle? 😮

Same issue on Mac OS X 10.10.5

It works for me..., maybe restarting the app ?

Thank you for your comment! :)

I assume developers would want to define a only-for-test activity inside test package, and if their project is a multi-module project, then they would want to put all the test related classes inside test-specific module and access the classes inside those module only when testing, with testImplementation.

After reading: https://github.com/robolectric/robolectric/issues/4640#issuecomment-464155200, I thought there should be a way to make the build pick up the AndroidManifest in the test package somehow. I'd love to know if there is any way to do that.

Hi, Maybe this is unnecessary, but I have made a repository for reproducing the NPE as the issue mentions, using Robolectric 4.2.

I hope it helps in some ways, and also it would be great if you let me know if I miss or misunderstand something. https://github.com/shoheikawano/Robolectric4.2Sample/tree/master/app/src/test

Thought it might be helpful if someone confirmed they'd got it working :tongue:

Thx for the formula I've got a working build using it

has good points. Once those are addressed, this is a +2

LGTM

needs rebase

As I mentioned here: https://github.com/robolectric/robolectric/issues/4640#issuecomment-468939893, I understand and agree with the idea that robolectric follows its behavior as application running on the device, but I also would like to know if there is a way to put test related classes(including for-test-only activity) inside test package, which requires loading AndroidManifest.xml inside test package when testing, or other ways.

Hi, I have made a sample repo for reproducing this NPE. It would be great if I misunderstand something or misplace the manifest. Thanks in advance!

https://github.com/shoheikawano/Robolectric4.2Sample/tree/master/app/src/test/java/com/shoheikawano/robolectricsample

CLA signature looks good :+1:

CLA signature looks good :+1:

Ok, this is good, but I'm thinking something else. Should we default to json_api adapter so this would never be an issue?

CLA signature looks good :+1:

CLA signature looks good :+1:

CLA signature looks good :+1: