Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix splice list insertion #84

Merged
merged 13 commits into from
Jun 6, 2024

Conversation

kekavc24
Copy link
Contributor

This PR:

@jonasfj Everything works fine and as expected. However, there is a test in randomTest.dart that fails because the string generated is invalid and is never parsed correctly from the start.

It's the only test failing.


  • I’ve reviewed the contributor guide and applied the relevant portions to this PR.
Contribution guidelines:

Note that many Dart repos have a weekly cadence for reviewing PRs - please allow for some latency before initial review feedback.

- true
- - false
---
- [splice, [key, 1], 0, 1, ['test']]
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Missing newline at end of the line.

Consider making a configuration change to your editor.

@jonasfj
Copy link
Member

jonasfj commented May 31, 2024

@jonasfj Everything works fine and as expected. However, there is a test in randomTest.dart that fails because the string generated is invalid and is never parsed correctly from the start.

Can we change it such that it doesn't generate invalid cases?

We need to either skip or fix tests, if they are unrelated to this issue, I don't mind skipping and filing an issue to fix them separately.

@kekavc24
Copy link
Contributor Author

kekavc24 commented May 31, 2024

@jonasfj Everything works fine and as expected. However, there is a test in randomTest.dart that fails because the string generated is invalid and is never parsed correctly from the start.

Can we change it such that it doesn't generate invalid cases?

We need to either skip or fix tests, if they are unrelated to this issue, I don't mind skipping and filing an issue to fix them separately.

It's the 20th iteration. Same random string (doesn't feel random) each time. An issue will be fine. I've skipped it for now. I can tackle it separately. I had to modify the Generator class slightly to get the current iteration.

@kekavc24 kekavc24 requested a review from jonasfj May 31, 2024 15:55
Copy link
Member

@jonasfj jonasfj left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nits only really :D

Looks like a solid solution to me.

Comment on lines 205 to 206
/// Anytime our '-' is non-existent, use '\n' as source of truth. Also, if
/// the new line is closer
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What does this mean, when does this happen? Can it happen?

Suggested change
/// Anytime our '-' is non-existent, use '\n' as source of truth. Also, if
/// the new line is closer
/// If '\n' comes before '-' then we're not inside a nested block list.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, it does since we are scanning backwards. I think you covered it in the last comment. Basically if:

  1. Index of - is -1 or less than \n index (even if \n index is -1), we always assume we inserting it i.e. +1.
  2. Otherwise, we insert after the last - i.e. `+2 inclusive of space after if nested in list.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: Should we update the comment or not?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll make it clearer.

Comment on lines 198 to 199
/// Checks if the [YamlNode]'s list is directly nested within another list
(bool isNested, int offset) _isNestedInBlockList(int start, String yaml) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are we sure it's a good idea to make this a function? are we going to use it elsewhere?

I don't mind, but if we do, we probably should document it clearly.

Like what is start what is yaml and what is the list being talked about.

The documentation has to explain that start is expected to be the offsets of the - indicator for an element in a list formatting in block-mode, and yaml is the entire YAML document.

And this method aims to find out if:

  • The element that has a start offset at start is inside a nested block list
  • offset for...

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Since this will probably only be used inside _insertInBlockList we could consider inlining it. Also given how hard it is to explain what start really is.

But add documentation comments that explains what it does is also fine.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I extracted it to make _insertInBlockList readable. Even though the function is doing a relatively simple thing, it can be hard to follow if leave it inside due to the many yaml.lastIndeOf calls we make

Comment on lines 180 to 196
final start = yaml.lastIndexOf('\n', currNodeStart) + 1;

return SourceEdit(start, 0, formattedValue);
var currSequenceOffset = yaml.lastIndexOf('-', currNodeStart - 1);

final (isNested, offset) = _isNestedInBlockList(currSequenceOffset - 1, yaml);

/// We have to get rid of the left indentation applied by default
if (isNested && index == 0) {
// Give the previous first element its indent
formattedValue = formattedValue.trimLeft() + indent;
}

return SourceEdit(offset, 0, formattedValue);
}
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So when we're inserting into a block list, we:

  • Find the node that is currently at the index we want to insert at.
  • Find the - indicator in front of that node.
  • Determine if we're in a nested block list or not
    • If so, ...?
  • else we....

I think what you're doing is probably correct in both cases, it's just a bit weird.

If I have a nested list:

 - - foo

and I'm prepending bar to the inner list, then I get a diff that looks like:

 - - foo
   ^
   Insert: "- bar\n   "

Where as if I have:

 - foo

and I prepend bar I get a diff that says:

 - foo
^
Insert: " - bar\n"

So in the case with no nested lists, we're inserting a full new line.
The - indicator that belongs to foo moves down.

Where as, if we have nested block lists, we insert into the middle of the existing line.

I guess this makes sense because in - - the first - belongs to the outer list, and we're not modifying the outer list. We're modifying the inner list.

Just typing this out so I think I understand it, this seems solid.

I'd suggest we either inline _isNestedInBlockList or improve the documentation so that it can stand on its own -- otherwise, it's a bit hard to understand what it does when someone reads it.


I guess technically we could in the second case have done:

 - foo
 ^
 Insert: "- bar\n "

That would be more consistent with the first case, and we wouldn't need to know if it's a nested list or not. But the diff is much better when we're inserting a full line, and that's also the common case after all, nested lists in block mode is fairly weird. So I totally understand why we need to know if it's inside a nested block list.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

😅 yaml syntax is quirky making everything weird. The existing code works well and assumes people will be civil and use

- 
  - value

instead of

- - value

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

oh, I didn't even realize that:

- 
  - value

was an option... can we add a few tests 🤣

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

oh, I didn't even realize that:

- 
  - value

was an option... can we add a few tests 🤣

Added these in 5978c20, I think.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added additional docs

@kekavc24 kekavc24 requested a review from jonasfj June 4, 2024 06:39
@kekavc24 kekavc24 mentioned this pull request Jun 4, 2024
1 task
Copy link
Member

@jonasfj jonasfj left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

just nits

@@ -132,7 +132,8 @@ SourceEdit _appendToBlockList(
}

/// Formats [item] into a new node for block lists.
String _formatNewBlock(YamlEditor yamlEdit, YamlList list, YamlNode item) {
(String formatted, String indent) _formatNewBlock(
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are you sure it wouldn't make more sense to return listIndentation ? Instead of returning the actual string.

Copy link
Contributor Author

@kekavc24 kekavc24 Jun 4, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'll investigate that. I did that for convenience. I didn't want to change so many things in the existing code.

/// Since we CANT'T/SHOULDN'T manipulate the next element to get rid of the
/// space it has, we remove the padding (if any is present) from the indent
/// itself.
indent = indent.replaceFirst(padding, '');
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Couldn't we use indent.substring()...

But also if we made _formatNewBlock return listIndentation instead of a string, then we can just do ' ' * (listIndentation - leftPad), right?

Or am I wrong here?

I'm just wondering, if maybe it's smarter to return integers instead of strings, when we want to communicate indentation. We return an integer from getListIndentation.

Honestly, we could probably also just call getListIndentation(yaml, list); again, rather than change _formatNewBlock, but I don't mind returning both from _formatNewBlock :D

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This actually makes sense. The _formatNewBlock can be tweaked and the other reference updated.

@kekavc24 kekavc24 requested a review from jonasfj June 6, 2024 08:59
@jonasfj jonasfj merged commit 08a146e into dart-lang:main Jun 6, 2024
10 checks passed
@jonasfj
Copy link
Member

jonasfj commented Jun 6, 2024

@kekavc24 by the way, feel free to file a PR with a new bullet point for the CHANGELOG.md. it might be nice for package consumers to know what we fixed :D

@kekavc24 kekavc24 deleted the fix-splice-list-insertion branch June 6, 2024 12:21
copybara-service bot pushed a commit to dart-lang/sdk that referenced this pull request Jun 6, 2024
…annel, webdev, yaml_edit

Revisions updated by `dart tools/rev_sdk_deps.dart`.

crypto (https://github.com/dart-lang/crypto/compare/0a10171..7a9428a):
  7a9428a  2024-06-04  Kevin Moore  Bump lints dep (dart-lang/crypto#172)

dartdoc (https://github.com/dart-lang/dartdoc/compare/45627f9..ddb8fb4):
  ddb8fb44  2024-06-03  Konstantin Scheglov  Use package:analyzer/source/source.dart (dart-lang/dartdoc#3780)

http (https://github.com/dart-lang/http/compare/7bfbeea..a3567f8):
  a3567f8  2024-06-05  Brian Quinlan  Prepare to release cronet_http 1.3.1 (dart-lang/http#1228)
  7addc61  2024-06-05  Hossein Yousefi  [cronet_http] Upgrade jnigen to 0.9.2 to close `#1213` (dart-lang/http#1225)
  eb87a60  2024-06-04  Anikate De  pkgs/ok_http: Close the client (override `close()` from `BaseClient`) (dart-lang/http#1224)
  9f022d2  2024-06-04  Anikate De  pkgs/http_client_conformance_tests: add boolean flag `supportsFoldedHeaders` (dart-lang/http#1222)
  90837df  2024-06-03  Anikate De  pkgs/ok_http: Condense JNI Bindings to `single_file` structure, and add missing server errors test (dart-lang/http#1221)

logging (https://github.com/dart-lang/logging/compare/73f043a..dbd6829):
  dbd6829  2024-06-04  Sarah Zakarias  Update `topics` in `pubspec.yaml` (dart-lang/logging#165)

test (https://github.com/dart-lang/test/compare/2464ad5..83c597e):
  83c597e5  2024-06-05  Jacob MacDonald  enable asserts in dart2wasm tests (dart-lang/test#2241)
  60353804  2024-06-04  Nate Bosch  Remove an unnecessary `dynamic` (dart-lang/test#2239)
  43ad4cd8  2024-06-03  Kevin Moore  random cleanup (dart-lang/test#2232)
  18db77c3  2024-06-02  Kevin Moore  prepare to publish test_api (dart-lang/test#2238)
  cd72e8d8  2024-06-02  Kevin Moore  Prepare to publish (dart-lang/test#2237)
  9c6adaa7  2024-06-01  Kevin Moore  fixes (dart-lang/test#2235)
  0110a80b  2024-06-01  dependabot[bot]  Bump the github-actions group with 2 updates (dart-lang/test#2236)
  b1b2c029  2024-06-01  Martin Kustermann  Fix `dart run test -p chrome -c dart2wasm` (dart-lang/test#2233)

tools (https://github.com/dart-lang/tools/compare/86b3661..4321aec):
  4321aec  2024-06-05  Andrew Kolos  [unified_analytics] Suppress any `FileSystemException` thrown during `Analytics.send` (dart-lang/tools#274)
  e5d4c8b  2024-06-03  dependabot[bot]  Bump actions/checkout from 4.1.5 to 4.1.6 in the github-actions group (dart-lang/tools#271)

web_socket_channel (https://github.com/dart-lang/web_socket_channel/compare/45b8ce9..bf69990):
  bf69990  2024-06-03  Thibault Chatillon  bump web_socket dependency 0.1.3 -> 0.1.5 (dart-lang/web_socket_channel#370)
  5b428dd  2024-06-02  dependabot[bot]  Bump actions/checkout from 4.1.5 to 4.1.6 in the github-actions group (dart-lang/web_socket_channel#371)
  afd1e3c  2024-05-23  Brian Quinlan  Remove `--fatal-infos` from `dart pub downgrade` analysis (dart-lang/web_socket_channel#367)
  cb20b71  2024-05-23  Sarah Zakarias  Add `topics` to `pubspec.yaml` (dart-lang/web_socket_channel#362)
  8514229  2024-05-22  Kevin Moore  Bump and fix lints (dart-lang/web_socket_channel#366)

webdev (https://github.com/dart-lang/webdev/compare/a97c2a1..9ada46f):
  9ada46fc  2024-06-05  Nicholas Shahan  Hide more temporary variables when debugging (dart-lang/webdev#2445)

yaml_edit (https://github.com/dart-lang/yaml_edit/compare/963e7a3..08a146e):
  08a146e  2024-06-06  Kavisi  Fix splice list insertion (dart-lang/yaml_edit#84)
  561309e  2024-06-01  dependabot[bot]  Bump actions/checkout from 4.1.5 to 4.1.6 in the github-actions group (dart-lang/yaml_edit#85)
  b582fd5  2024-05-31  Kavisi  Fix error thrown when inserting keys (dart-lang/yaml_edit#80)

Change-Id: I36acbc26fe97d8a1e3fdfa5f4855cc4be7d84dd7
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/370080
Commit-Queue: Devon Carew <devoncarew@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Splice list in a nested block list fails!
2 participants