untitled.send :title, 'Hello!'

thisissgchineseprivilege:

bankuei:

The people you have to handhold, lead, and try to plead with about your humanity are not “potential allies”. They are far from allies.

You cannot win respect from people who demand you demean yourself or they will harm you. Respect is not in their equation at all.

For the minorities who constantly want to ‘be nice’ to their Chinese friends, or they won’t listen to you.

Reblogged from: thisissgchineseprivilege

thisissgchineseprivilege:

catstin:

dorothyparkerwashere:

catstin:

thisissgchineseprivilege:

Dear Spotify,

I am NOT in China. Please for the love of God, STOP with your Mandarin advertisements.

Chinese Privilege is having even international companies cater to you and your language despite it being streamed in a multiracial country.

Isn’t one of the official languages of Singapore Chinese??? That’s like getting upset when you hear announcements in Flemish & French in Belgium? I’m sincerely asking, not trying to be a jackass

Tamil and Malay are also official languages. In face, Malay is the NATIONAL LANGUAGE. So why aren’t there Malay ads?

It’s so convenient that the ads just happen to be in Mandarin.

You might not be trying to be a jackass, but your dismissive question sure made you sound like one.

Maybe the Chinese market is bigger? Idk.
Romansch is an official language of Switzerland but hardly any of its media is in it because it has such a small market.

The fact that you insist on continuing to defend Chinese language supremacy in Singapore really says everything. You can’t even be bothered to try to understand why this is upsetting to minorities, or that your ridiculous excuses have been used to erase and deny us representation for decades. You have nothing new or original to say here, you are actually spouting rhetoric that racists do all the time.

When you sound similar to racists, maybe it’s time you stop and think about who you really are.

If Singapore isn’t in China yet, you see the signs all around saying “We’re working on it”.

This last GE really was about whether or not we wanted Singapore as a multicultural, democratic, independent state, or whether we were happy with our continuing acceleration into the suffocating embrace of the PRC.

Those of us of a certain age recall when the God-King LKY cried on television at a singularly opportune time. He may have lost his chance to be God-King of a unified Malaya, but for his dynasty, a to-be-unified China awaits.

Is this why you did your NS? I didn’t think so.

Reblogged from: thisissgchineseprivilege

And that’s a Very Good Thing™ for both language communities. Yes, it’s good to know both languages and to use each where they make sense. But They’re Not The Same Thing. Obvious, yes?

Introduction

Abe Voelker wrote a blog post in which he sang the praises of the Kleisli Gem to, in effect, let you write Haskell code in your Ruby. He started out by demonstrating a “naive” validation object for validating an uploaded image according to four criteria, verifying that the image:

1. Has a size of less than one megabyte;
2. Is a valid image format;
3. Is of type PNG, JPEG, or GIF; and
4. Has dimensions of less than 5,000 by 5,000 pixels.

In addition, he states a requirement for “a JSON-friendly version of the validation return value as the website is uploading the image via Ajax”.

Fine. Easy enough, without resorting to exotic “Reese’s Pieces” language concoctions.

Hasty Rewrite

The Gist copied at the end of this post is what I’d consider a first draft of a rewritten AvatarValidator class. It’s a somewhat more readable version of the original; both are still massive SRP violations. They do too much; they have leaky abstractions through and through. But this is still an improvement:

Command Pattern and State Machine

The original code made was clearly a classic command pattern, but implemented in a massive nested-if furball. You had to squint to see that it was implementing a stepwise state machine, with subsequent states being implemented as ever-more-deeply-nested if blocks.

In the first replacement, the #validate method (inside the guard clause) is a sequence of calls to private methods with intention-revealing names that (should) make it easier to find what you’re looking for, or to understand the overall process.

Error Detection and Message Output

The original code has inline, nested if statements whose else clauses each build and return literal strings, more than doubling the length of the #validate method (lines 10-33) compared with the new implementation. Changes to the original output messages do not originate in a “single source of truth”, which complicates and adds risk to change; are you sure you’ve done everything and only what you set out to do? (Remember, there are no tests yet.)

In the new code, any validation error detected will raise a RuntimeError whose message encodesan identifier for the error, along with a single relevant data item. The handler for that RuntimeError calls a private method that centralises all text strings for error messages, making it easy to change later if needed (e.g, for i18n).

What’s Our Input?

The original code passes in an uploaded_file parameter, which is assigned to an instance variable (at line 7). That ivar is then used, duck-type style, to find its size (by calling the #size method at line 12) and by calling the path method of the return object from uploaded_file’s #tempfile method at line 15).

In contrast, the first whack at a replacement AvatarValidator has a nested class, AvatarValidator::UploadedFile, which reports the #size (needed for validation) and produces a MiniMagick::Image instance of the data in the uploaded file (however that file was specified, whether as a blob of data, a filename or an instance of AvatarValidator::UploadedFile itself).

Known Problems

1. There are no tests. Even if there were, it’s tedious/dangerous to attempt to test individual validation steps in isolation.
2. The image-dimension validation now separately checks height and width, reporting different error messages when invalid; the original code (see lines 18-19) did not.
3. AvatarValidator still has repetitive code. The internal validation-step methods (#verify_file_size, #validate_image_format, #verify_allowed_format, and (to a slightly lesser degree) #verify_image_dimensions) each follow the same pattern: bail out if we’re valid, or call #fail_with if we’re not.
4. This code is unloved by RuboCop, a widely-used Ruby style checker and our shop standard. In particular, the offence triggered by the #verify_image_dimensions method, Assignment Branch Condition size for verify_image_dimensions is too high. [16.16/15] is one I have learned to take seriously. RuboCop’s default trigger value of 15 is in fact higher than what you should see if, say, you were following Sandi Metz’ rules.

Coming Soon

A new iteration of the rewrite will soon address these problems, and produce code that is neither a straw man to provide a convenient excuse to switch languages, nor code that would budge the WTF-O-Meter® at your team’s code review.

The First Draft

Change is good. Change is a sign of life. Change that moves you towards your goal is a Good Thing™.

Change that makes you run around the same circle over and over again is not such a Good Thing.

OpenSSL has had security Issues for quite some time. Recently, like in the last year or so, it’s been getting updated more regularly as problems get addressed.

“Good”, you say. “Secure data is what the Internet relies on, isn’t it?” Well, no; not exactly. And closer to home, every Ruby developer dreads seeing that message from their package manager of choice saying “Your openssl version has been updated”.

Because, if you’re using a version manager like RVM or rbenv, that means you get to go reinstall every Ruby version you need to run on your system.

Here’s basically what I do. I’m using rbenv with a current production Ruby of 2.1.4 on a Mac; your setup will differ based on how different your system is. (Translation: Don’t come crying to me if this doesn’t work for you. You’re smart enough to use Ruby; you’re smart enough to fix it, with a little help from DuckDuckGo):

Step 1: Run rbenv versions. This presently gives me output of

  system
  2.0.0-p481
* 2.1.4 (set by /usr/local/rbenv/version)
  2.2.0-preview1

Step 2: Nuke from orbit. (Note that I’m running on OS X 10.8.5… yeah, yeah…)

$ rbenv global system
$ rbenv rehash
$ ruby --version
ruby 1.8.7 (2012-02-08 patchlevel 358) [universal-darwin12.0]
$ rm -rf $RBENV_ROOT/versions/*

Step 3: Reinstall “main” Ruby version and systemwide Gems. The `say “All done” is so Daniel tells me when it’s done, so I can go do other things and come back.

$ rbenv install 2.1.4 && say "All done"
$ rbenv global 2.1.4
$ gem install rails rspec simplecov pry-{bloodline,doc,docmore,byebug,stack_explorer,rescue,theme,pretty-numeric,highlight,editline,git} awesome_print bond jist jeweler lotusrb lotus-{model,helpers} padrino padrino-{assets,flash} rubycritic && say "all done"
$ rbenv rehash
$ gem cleanup

Step 4: Lather; rinse; repeat for every other Ruby version needed on the system. Remember that openssl is statically linked into Ruby and it has to match the current (latest security-update-driven) version on the system.

$ rbenv global system && rbenv rehash
# ...
$ gem cleanup
$ rbenv global 2.1.4 && rbenv rehash

Now you can get started on what you set out to do this morning… after tea break.