Feature indent depth check #3395

yubaoquan · 2017-10-26T11:09:31Z

PR checklist

Addresses an existing issue: Indent rule not reporting or fixing indent size violations #2814
Enhancement
- Includes tests
Documentation update

Overview of change:

Add ability to check the indent depth.

For example:

function foo() {
    var a = 1 // 1 indent depth
    var b = function() {
        var c = 2 // 2 indent depth
    } // 1 indent depth
} // 0 indent depth

Is there anything you'd like reviewers to focus on?

Maybe not all use cases in the test files are covered.

CHANGELOG.md entry:

[new-fixer]
[enhancement]

ajafff

Just a quick first review. The code is a bit difficult to follow. I will do a more thorough review in the next days (hopefully tomorrow).

ajafff · 2017-10-26T20:00:12Z

src/rules/indentRule.ts

+}
+
+function recursiveGetNodeAt(root: ts.Node, pos: number): ts.Node | undefined {
+    let result: ts.Node | undefined = root.getChildAt(pos);


this does not what you intended. it gets the nth child, not the child at position n

It just works. Not found a test case can prove this usage is wrong.

ajafff · 2017-10-26T20:11:49Z

src/rules/indentRule.ts

+            if (node.getFullStart() < pos && node.getEnd() >= pos) {
+                result = recursiveGetNodeAt(node, pos);
+            }
+            ts.forEachChild(node, cb);


you won't find the node if you recurse here, because the parent is already out of the range

ok, code removed

ajafff · 2017-10-26T20:14:51Z

src/rules/indentRule.ts

+    const currentNodePos: number = pos + indentEnd;
+    let currentNode: ts.Node | undefined = recursiveGetNodeAt(sourceFile, currentNodePos);
+    let depth = 0;
+    if (currentNode !== undefined && inTsx(sourceFile)) {


Is there a reason to do nothing at all in .tsx files? And should the same apply to .jsx?
If there is a good reason, you could consider moving this check to the top of walk to avoid unnecessary work.

Added the check for jsx syntax.

ajafff · 2017-10-26T20:16:12Z

src/rules/indentRule.ts

+    indentEnd: number;
+    line: string;
+    sourceFile: ts.SourceFile;
+}): boolean {


consider moving this function inside walk so it closes over the variables like ctx, tabs, sourceFile and so on.

ajafff · 2017-10-26T20:20:13Z

src/rules/indentRule.ts

+        ts.SyntaxKind.TemplateHead,
+        ts.SyntaxKind.TemplateMiddle,
+        ts.SyntaxKind.TemplateTail,
+    ].indexOf(node.kind) > -1;


node.kind >= ts.SyntaxKind.FirstTemplateToken && node.kind <= ts.SyntaxKind.LastTemplateToken

As you wish

ajafff · 2017-10-26T20:21:57Z

src/rules/indentRule.ts

+    let result = 0;
+    let parent: ts.Node | undefined = node.parent;
+    const blockTypes: ts.SyntaxKind[] = [
+        ts.SyntaxKind.Block, // 207


these numbers are subject to change with future versions of typescript. you can remove them, because they do not add any value

ajafff · 2017-10-26T20:23:35Z

src/rules/indentRule.ts

+        ts.SyntaxKind.CaseBlock, // 235
+        ts.SyntaxKind.EnumMember, // 264
+        ts.SyntaxKind.CaseClause, // 257
+        // ts.SyntaxKind.IfStatement, // Have no idea on how to check whether a node is in if paren


you probably also need ForStatement, ForOfStatement, ForInStatement, WhileStatement, DoStatement and WithStatement but only if they do not contain a block

Added another function to check text in parens of these statements

ajafff · 2017-10-26T20:24:40Z

docs/_data/formatters.json

@@ -0,0 +1,83 @@
+[


if you merge latest master these generated files should be gone

for some strange resons the files still show up in the diff. please git rm docs/formatters/stylish/index.html docs/_data/formatters.json

adidahiya · 2017-10-26T20:55:28Z

Note that there is already another open PR that tackles the same issue: #3300

1. add jsx syntax lint 2. add check for text in `if`/`for`/`while` parens

yubaoquan · 2017-10-30T05:54:21Z

How's the progress?

yubaoquan · 2017-10-31T02:38:37Z

Any problems with this pr? @ajafff

ajafff

Besides my comments, please add some more comments to describe what the code does and why it needs to do it. As it is now, the code is pretty hard to follow.

ajafff · 2017-10-31T18:55:36Z

docs/_data/formatters.json

@@ -0,0 +1,83 @@
+[


for some strange resons the files still show up in the diff. please git rm docs/formatters/stylish/index.html docs/_data/formatters.json

ajafff · 2017-10-31T18:58:50Z

src/rules/indentRule.ts

+            const currentNodePos: number = pos + indentEnd;
+            let currentNode: ts.Node | undefined = recursiveGetNodeAt(sourceFile, currentNodePos);
+            let depth = 0;
+            if (currentNode !== undefined && currentNode.getStart() === currentNodePos) {


consider passing sourceFile as parameter to .getStart() for performance reasons

ajafff · 2017-10-31T19:10:43Z

src/rules/indentRule.ts

+                const passIndentDepthCheck = inStringTemplate(currentNode) || (line.trim().charAt(0) === "*");
+
+                if (!passIndentDepthCheck && expectedIndentation !== actualIndentation) {
+                    failure = Rule.FAILURE_STRING(tabs ? "tab" : size === undefined ? "space" : `${depth * size} space`);


size can never be undefined in this block

ajafff · 2017-10-31T19:18:15Z

src/rules/indentRule.ts

+    const result: ts.Node | undefined = root.getChildAt(pos);
+    if (result === undefined) {
+        ts.forEachChild(root, function cb(node: ts.Node): void {
+            ts.forEachChild(node, cb);


this is actually a no-op, because the callback never modifies result

ajafff · 2017-10-31T19:22:24Z

src/rules/indentRule.ts

@@ -121,6 +159,130 @@ function walk(ctx: Lint.WalkContext<Options>): void {
    }
 }

+function recursiveGetNodeAt(root: ts.Node, pos: number): ts.Node | undefined {
+    const result: ts.Node | undefined = root.getChildAt(pos);


I still don't think this is right. Since root is always the SourceFile, getChildAt will return a SyntaxList with all statements when pos === 0 and EndOfFileToken when pos === 1. In all other cases it just returns undefined

Updated. The recursiveGetNodeAt and recursiveGetNodeAt are not suitable names with the logic, and the inner logic is not clearly, so I refactor theses two function to one single function recursiveGetNodeAt.
The function just find the node of current line. If the pos is between the getFullStart and getEnd, that means the line is the node or is child of the node.

ajafff · 2017-10-31T19:56:16Z

src/rules/indentRule.ts

+                const expectedIndentation: string = tabs ? "\t".repeat(depth)
+                : " ".repeat(size * depth);
+                const actualIndentation = line.slice(0, indentEnd);
+                const passIndentDepthCheck = inStringTemplate(currentNode) || (line.trim().charAt(0) === "*");


(line.trim().charAt(0) === "*") is used to check for multiline JSDoc comments?
If yes, this needs to be updated to also support multiline comments without a leading asterisk. You can simply use isPositionInComment from the tsutils package

ajafff · 2017-10-31T19:57:58Z

src/rules/indentRule.ts

@@ -121,6 +159,130 @@ function walk(ctx: Lint.WalkContext<Options>): void {
    }
 }

+function recursiveGetNodeAt(root: ts.Node, pos: number): ts.Node | undefined {


for your use case you can probably just use getTokenAtPosition from the tsutils package

Tried, the replacement produce wrong results. The getTokenAtPosition will always return a token, which is not expected. In recursiveGetNodeAt, the getChildAt will return undefined on some positions, and then the function should return the root node. If replace to getTokenAtPosition, the returned value will be unexpected. Anyway I have refactored this function.

ajafff · 2017-10-31T20:17:27Z

src/rules/indentRule.ts

+            result++;
+        }
+        if (inStatementParen(parent)) {
+            result ++;


some special cases to consider:

// multiline condition, works well for `if`, but probably doesn't for `while` while(foo() && bar() && baz()) { doStuff; } // might not be worth fixing, because it can be reformatted to while( foo() && bar() && baz() ) { doStuff; } // statement continuation on the next line let foo = bar ? baz : bas; foo = "some" + "really" + "long" + "string"; // multiline property access foo .bar .baz();

OK, add these into test cases. Test passed.

ajafff · 2017-10-31T20:18:25Z

src/rules/indentRule.ts

+    if (
+        node.parent !== undefined
+        && node.parent.kind === ts.SyntaxKind.CallExpression
+        && /^\s*\./.test(line)


please add a comment explaining what this condition does. it took me a while to figure it out

ajafff · 2017-10-31T20:18:52Z

src/rules/indentRule.ts

+    );
+}
+
+function nodeAtOutside(node: ts.Node): boolean {


please add a comment to describe what this function does

ajafff · 2017-10-31T20:24:24Z

@yubaoquan sorry for the delay, I got distracted. Unfortunately writing new code is a lot more fun than reviewing someone else's code.

yubaoquan · 2017-11-01T06:26:16Z

Updated the code as you said. Thank you for your review and advices!

yubaoquan · 2017-11-01T06:32:49Z

Something is wrong with CI, how to fix it? Should I merge the master branch into this branch?

error Couldn't find the binary npm-run-all -p compile:core compile:test -s compile:scripts

ajafff · 2017-11-01T07:39:40Z

Something is wrong with CI, how to fix it? Should I merge the master branch into this branch?

Exactly. This is fixed on master, you just need to merge

yubaoquan · 2017-11-05T04:11:20Z

how's it going?

yubaoquan · 2017-11-13T12:06:48Z

Do we have an estimate of when this might be merged? @ajafff @adidahiya

ajafff · 2017-11-13T19:40:27Z

@yubaoquan before jumping right back to reviewing the code, I tried how it works. There are several false positives that need to be addressed:

multiline arguments:

foo(
    bar, // 0 space indentation expected
    baz, // 0 space indentation expected
);

new Foo(
    bar, // 0 space indentation expected
    baz, // 0 space indentation expected
);

continuation on the next line:

let someReallyLongVariableName =
    anotherReallyLongVariableName; // 0 space indentation expected

return foo
    && bar // 0 space indentation expected
    && baz; // 0 space indentation expected

someArray.filter((element) =>
    doStuff(element)); // 0 space indentation expected
// or
const someFn = (element) =>
    doStuff(element); // 0 space indentation expected

multiline named imports:

import {
    foo, // 0 space indentation expected
    bar, // 0 space indentation expected
} from "foobar";

// probably the same for 'export'
export {
    foo, // 0 space indentation expected
    bar, // 0 space indentation expected
};

else-if:

if (foo) {
    foo;
} else if (bar) {
    bar; // 8 space indentation expected
} // 4 space indentation expected

type literals:

export interface TestResult {
    directory: string;
    results: {
        [fileName: string]: TestOutput | SkippedTest; // 4 space indentation expected
    };
}

ParenthesizedExpression:

return (
    foo // 0 space indentation expected
);

template expressions:

`foo ${
   foo // 0 space indentation expected
}`

template literals:

foo(
    `foo`, // this line is valid
// next line should have an error, currently it is ignored
  `
bar
`,
// the two lines above are valid, because the indent is inside the template string
);

indent wrongly calculated at start of statement:

// next line starts with one space
 for(;;) {} // 4 space indentation expected

Of course everything mentioned above is open for discussion.

yubaoquan · 2017-11-14T15:24:43Z

All cases fixed except two below:

foo(
    `foo`, // this line is valid
// next line should have an error, currently it is ignored
  `
bar
`,
// the two lines above are valid, because the indent is inside the template string
);

I don't know what error should we get.
I removed indent checking in string template because the indent of the first variable in string template isn't credible (the indent could be part of the string).

What's more, if you got a ~ [1] in a string template, how do you make sure whether it is an error mark or content of the string?

ajafff · 2017-11-14T19:39:24Z

IMO the opening backtick of a template string should be checked like every other token. We just ignore the indent inside the template string, because that could be part of the content.

What's more, if you got a ~ [1] in a string template, how do you make sure whether it is an error mark or content of the string?

The *.ts.lint files are preprocessed to remove all error markers before parsing the file content as TypeScript. That means you can have valid typescript code mistakenly treated as error marker (very unlikely), but not the other way around.

ajafff · 2017-11-14T20:24:44Z

Here we go with the next round of false positives:

        const linter = new Linter({
            fix: false,
            formatter: "prose",
        }); // 12 spaces indentation expected - currently 8

still a problem with continuation on the next line:

        const expectedResult =
            getFailureString(TEST_FILE, 1,  1, "first failure", "first-name") + // 8 space indentation expected - currently 12
            getFailureString(TEST_FILE, 2, 12, "mid failure", "mid-name") + // 8 space indentation expected - currently 12
            getFailureString(TEST_FILE, 9,  2,  "last failure", "last-name"); // 8 space indentation expected - currently 12

chained method calls:

fs.createReadStream("test/files/multiple-fixes-test/multiple-fixes.test.ts")
    .pipe(fs.createWriteStream(tempFile))
    .on("finish", () => {
        doStuff(); // 4 spaces indentation expected
    }); // 0 spaces indentation expected

blocks in CaseClause and DefaultClause

            case ts.SyntaxKind.BinaryExpression: {
                break; // 20 spaces indentation expected - currently 16
            } // 20 spaces indentation expected - currently 12

multiline arrow function as argument:

    return find(sym.declarations, (decl) =>
        isClassLikeDeclaration(decl) ? decl.typeParameters : undefined); // 12 - currenly 8

property initializer on the next line:

class Foo {
    public static FAILURE_STRING_TRAILING =
        "Did not expect single-line type literal to have a trailing ';'."; // 4 spaces indentation expected -
 currently 8
}

something with TypeAlias:

export type Location =
    | ts.PrefixUnaryExpression // 0 spaces indentation expected (for all of them)
    | ts.IfStatement
    | ts.WhileStatement
    | ts.DoStatement
    | ts.ForStatement
    | ts.ConditionalExpression
    | ts.BinaryExpression;

something with abstract class:

// only reproducible if the class declaration is not on the first line
abstract class Foo {} // 4 spaces indentation expected

multiline type arguments:

type T = Map<
    string, // 0 space indentation expected
    string // 0 space indentation expected
>;

I like how you handle nested conditional expressions:

return foo
    ? foo
    : bar
        ? bar
        : baz;

Though others prefer the following, which is now invalid. I don't have a strong opinion for one of them.

return foo ? foo
    : bar ? bar
    : baz;

I actually feel kind of sorry for you, because your first contribution turns out to be that difficult. I really appreciate the effort you put into this PR.

yubaoquan · 2017-11-16T05:49:49Z

My logic of calculating the depth of code after chaining call is not good. I need time to refactor that part. Recently pretty busy, so it will take weeks before committing next pr.

refactor chaining call check and add other test cases

yubaoquan · 2017-11-21T16:16:49Z

Pretty hard to cover all the code styles. Pretty hard to debug this time. Anyway, fixed all cases you mentioned last time, and added the corresponding test cases.

As for the conditional expressions, I have no idea, should it to be an configurable option in this rule?

…ure-indent-size-check Fix circleCI error

yubaoquan · 2017-11-21T16:25:39Z

test/rules/trailing-comma/singleline-always/test.ts.lint: Failed!
It failed on a test which is not of indent check, is it matters?
I have already merged the master branch, how to fix it now?

ajafff · 2017-11-21T16:30:17Z

You cannot fix the failing test, see #3494 (comment)

adidahiya

Pretty hard to cover all the code styles. Pretty hard to debug this time.

Yep, which is why I personally prefer to leave such formatting to tools like prettier :)

adidahiya · 2017-11-28T22:38:27Z

src/rules/indentRule.ts

+    return result;
+}
+
+// function isSameLineWithConditionalExpr(sourceFile: ts.SourceFile, nodePos: number, parent: ts.Node): boolean {


remove commented code

adidahiya · 2017-11-28T22:38:43Z

src/rules/indentRule.ts

+foo()
+.then()
+In this case, `.then` should have one more indent.
+Maybe this should depend on a configuration.


use proper JSDoc formatting please

also, is this last line a question for code reviewers? I don't think this is a helpful comment without more elaboration. Either remove it or make the "maybe" part more clear.

adidahiya · 2017-11-28T22:40:22Z

src/rules/indentRule.ts

+ *      // xxx
+ *  }
+ *  In this case, the `Foo` is regarded as should have the same indent depth with `class`.
+ *  Should it have one more depth? I'm not sure.


use proper JSDoc formatting -- there is an extra whitespace after each * here

…ure-indent-size-check merge master

ajafff · 2017-11-29T19:07:03Z

@adidahiya

I personally prefer to leave such formatting to tools like prettier :)

Is that a -1 on this PR? If so, we can stop here and instead point people to ter-indent from tslint-eslint-rules

adidahiya · 2017-11-29T19:21:27Z

Sorry, no, it's not an explicit -1 on this PR.

There's a lot of demand for this feature in core TSLint, so it makes sense to add it, but we (project collaborators) might provide a lower amount of support for it than other core features. It might even deserve a callout in the docs that says the fixer for this rule is "experimental" (defer to your judgement here -- I haven't looked very closely at the CR progress and the false positives identified here).

I think in the future we should break up the core rule set into multiple packages, one of which would be formatting rules. Users of TSLint + Prettier would be able to disable the formatting rules all at once and we could set users' expectations of support accordingly. But of course that's out of scope for this PR.

yubaoquan · 2017-12-05T15:01:32Z

When can this be merged? Or lets think more test cases? @ajafff

ajafff · 2017-12-05T16:21:17Z

Ok, let's continue:

const expectedResult = 
    `foo`; // 0 spaces expected

return execRunner(
    {
        foo,
    }, // 0 spaces expected
);

const options = {
    foo; // 8 spaces expected
}; // 4 spaces expected

type Options = Record< // 4 spaces expected
    "branch" | "decl" | "operator" | "module" | "separator" | "restSpread" | "type" | "typecast" | "typeOperator" | "preblock",
    boolean>;
type Foo = boolean; // 4 spaces expected

return (a.dotDotDotToken !== undefined) === (b.dotDotDotToken !== undefined) &&
        (a.questionToken !== undefined) === (b.questionToken !== undefined); // 0 spaces expected

const couldBeSwitch = everyCase(node, (expr) => {
    casesSeen++;
    if (switchVariable !== undefined) {
        return nodeEquals(expr, switchVariable, sourceFile);
    } else {
        switchVariable = expr; // 12 spaces expected
        return true;
    }
});

doStuff([
    foo, // 8 spaces expected
]); // 4 spaces expected

I think that's it for now.

I think we better add a new option for checking the indent level. This new behavior could make a lot users insane when enabled by default.

…ure-indent-size-check

yubaoquan · 2017-12-09T13:21:55Z

Done. New test cases added, bug fixed.
The cases are growing bigger and bigger.

yubaoquan · 2017-12-09T13:31:37Z

I think a better way to check indent is making a tree, each line of the source file corresponds to a leaf on the tree. The depth of the leaf is the indent depth of the line. So the whole thing would be much easier. But I'm not good at making such a tree.

cyberhck · 2018-01-11T01:31:26Z

is it almost ready to merge now? any estimation on when this can get merged?

yubaoquan · 2018-01-11T08:23:34Z

@ajafff It's time to find new bugs, need your report! thanks :)

erikkemperman · 2018-05-04T08:41:08Z

@yubaoquan @ajafff I'd really love to see this feature released, so for what it's worth:

I've fetched this branch and merged current master into it;
Ran yarn install --pure-lockfile && yarn run validate (all tests still pass);
Used yarn link to use this version in my project;
Verified that it finds various indentation errors I intentionally added;
Verified that it fixes indentation errors I intentionally added;

Of course there are many variations of nested blocks, and statement continuations, and so on -- thus I can't try everything, but as far as I did try this seems to work well!

Tslint indent

giladgray · 2018-07-28T00:39:47Z

@yubaoquan please update this branch so we can review it, or close it if no longer relevant. we will close this if we do not hear from you in two weeks.

I am inclined to reject this feature for two reasons:

soooo much code is hard to review and trust.
Prettier handles this feature elegantly and has fixer support.

johnwiseheart · 2018-08-22T18:30:58Z

Closing due to age and inactivity.

yubaoquan added 3 commits October 25, 2017 22:06

init

beb2de3

Indent depth check

d6849b7

Merge branch 'master' into feature-indent-size-check

93b1d87

ajafff reviewed Oct 26, 2017

View reviewed changes

update

bf0e47d

1. add jsx syntax lint 2. add check for text in `if`/`for`/`while` parens

yubaoquan changed the title ~~Feature indent size check~~ Feature indent depth check Oct 27, 2017

ajafff reviewed Oct 31, 2017

View reviewed changes

refactor get node function, add comments and test cases.

1776b78

yubaoquan added 3 commits November 1, 2017 17:09

Merge branch 'master' into feature-indent-size-check

c9a1b3f

Add new test cases

e95367e

Add check for interface and namespace declarations

3dd9b07

add more test cases

c099298

remove !for number

71c4da2

update: bugfix

0d9e233

refactor chaining call check and add other test cases

Merge branch 'master' of https://github.com/palantir/tslint into feat…

4c056d5

…ure-indent-size-check Fix circleCI error

adidahiya reviewed Nov 28, 2017

View reviewed changes

yubaoquan added 2 commits November 29, 2017 11:16

Merge branch 'master' of https://github.com/palantir/tslint into feat…

5e970e5

…ure-indent-size-check merge master

format commnents

74986da

adidahiya mentioned this pull request Nov 29, 2017

docs update - Add hasFix metadata for the indent rule #3529

Merged

4 tasks

yubaoquan added 2 commits December 9, 2017 21:11

Update the ident rule logic, fix some bugs

648ca33

Merge branch 'master' of https://github.com/palantir/tslint into feat…

f0f2031

…ure-indent-size-check

This was referenced Dec 18, 2017

Split out formatting rules into a separate package #3592

Closed

Add fixer for max-line-length rule #3585

Closed

aervin mentioned this pull request Dec 19, 2017

[enhancement] fixer for newline-before-return rule #3584

Closed

4 tasks

Merge branch 'master' into tslint-indent

e181237

Add space count to indent rules in 'all' config

895fd6c

erikkemperman mentioned this pull request May 4, 2018

Tslint indent yubaoquan/tslint#1

Merged

Merge pull request #1 from erikkemperman/tslint-indent

3fa1c1f

Tslint indent

aervin mentioned this pull request May 11, 2018

Indent rule not reporting or fixing indent size violations #2814

Closed

adidahiya mentioned this pull request Aug 10, 2018

Add check for indent size violations #3300

Closed

4 tasks

johnwiseheart closed this Aug 22, 2018

Feature indent depth check #3395

Feature indent depth check #3395

Conversation

yubaoquan commented Oct 26, 2017

PR checklist

Overview of change:

Is there anything you'd like reviewers to focus on?

CHANGELOG.md entry:

ajafff left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

adidahiya commented Oct 26, 2017

yubaoquan commented Oct 30, 2017

yubaoquan commented Oct 31, 2017

ajafff left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ajafff commented Oct 31, 2017

yubaoquan commented Nov 1, 2017

yubaoquan commented Nov 1, 2017

ajafff commented Nov 1, 2017

yubaoquan commented Nov 5, 2017

yubaoquan commented Nov 13, 2017

ajafff commented Nov 13, 2017

yubaoquan commented Nov 14, 2017 • edited Loading

ajafff commented Nov 14, 2017

ajafff commented Nov 14, 2017

yubaoquan commented Nov 16, 2017

yubaoquan commented Nov 21, 2017

yubaoquan commented Nov 21, 2017

ajafff commented Nov 21, 2017

adidahiya left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ajafff commented Nov 29, 2017

adidahiya commented Nov 29, 2017

yubaoquan commented Dec 5, 2017

ajafff commented Dec 5, 2017

yubaoquan commented Dec 9, 2017

yubaoquan commented Dec 9, 2017

yubaoquan commented Nov 14, 2017 •

edited

Loading

erikkemperman commented May 4, 2018 •

edited

Loading