nuclide-logging - npm Package Versions

0.0.35

• Improvements to prevent spawning many Mercurial processes.
• Diff View:
  • Performance Optimizations (Caching and avoid expensive calls when Diff View is closed).
  • Add a 'Navigation Bar' for easier (add/delete/change) line change scrolling
• Hyperclick / Go to definition Improvements:
  • Go to definition doesn't collide with multi-cursor feature (not end up with multiple cursors).
  • Hack: Go to definition shows a navigable suggestion list when hh_server can't determine the exact location of the (e.g. dynamically-typed method calls).
  • Hack: XHP go to definition highlights XHP tags correctly (while it still jumps to the right place).
  • UX: Fix cursor indicator for loading and result-ready.
• Support formatting of const requires in format-js

0.0.34

npm-only release. Our Node packages were released so that external packages could consume them, but we did not update any Atom packages.

0.0.32

• Bug Fix

  • Exclude nuclide-debugger-node from nuclide-installer until we can fix the issue with its native dependencies getting built properly. Fixes #193.
  • Hack and Flow diagnostics are now removed appropriately when the corresponding project folder is removed.

• Documentation

  • Added README for nuclide-diagnostics-store.

0.0.31

• New Feature

  • New Diagnostics UI Previously, we encouraged Nuclide users to install the third-party linter package in order to see errors from our Flow and Hack packages, but now we include our own packages for displaying diagnostics as part of Nuclide. We divided this functionality across two Atom packages: nuclide-diagnostics-store and nuclide-diagnostics-ui. The former is the model/controller component where diagnostics are reported and broadcast out to registered listeners. The latter is the view component that subscribes to the model and presents diagnostics to the user via editor gutters and a table view. The nuclide-diagnostics-store is capable of consuming providers that conform to the linter API in addition to its own API. This means that even if you disable the linter package, you will still see all of the same errors and warnings from packages that adhere to the linter API in Nuclide.
  • Buck We now support building with Buck in Nuclide. If the active editor corresponds to a file in a Buck project (which means it has a .buckconfig file in its ancestor directory), then the Buck toolbar will be displayed at the top of your Atom window. From there, you can enter a Buck build target or alias, and then choose to build, run, or debug that target. (Note that like most features in Nuclide, this also works if you are working with remote files.) We also syntax highlight BUCK files and make it possible to cmd-click on build targets to navigate between BUCK files. As Buck supplies all sorts of information about the structure of your project to Nuclide, you can expect us to provide even more integration with Buck going forward.
  • iOS support This release builds on top of the new Buck support to also provide support for iOS. The major caveat is that your iOS project must be configured to build with Buck in order to get support for iOS development in Nuclide. Once you have made this investment, common IDE features, such as autocomplete, click-to-symbol, and diagnostics reporting will be available for your iOS code in Nuclide. You can also build, run, and yes, debug your iOS app from Nuclide. As is common in most IDEs, you can click in the gutter of an editor to set a breakpoint. When running your app in debug mode and you encounter a breakpoint, you can inspect it via the familiar Chrome DevTools UI. Note that the Console in the DevTools is wired up to the LLDB REPL, so you can issue LLDB commands such as thread backtrace.
  • cxx support In many ways, iOS support is just a special case of supporting the C-family of languages. However, the same caveat still applies: your cxx project needs to build with Buck in order to get support for autocomplete, click-to-symbol, and diagnostics.
  • Hack improvements:
    • Use hh_client for the diagnostics result if available for better accuracy and consistency and eliminate the false positives of cases not known for the local Hack model.
    • Prioritize Hack's context-sensitive autocompletions over grammar snippets.
    • Autocomplete of static and instance methods does not require a typed prefix (to be more predictable).
    • Updated hh_ide.js to the latest HHVM build.
  • Flow improvements:
    • Display all Flow errors in the entire project, rather than just those in files that have been recently opened.
    • Fix Flow autocomplete, and use snippets to pre-fill function arguments.
  • hg blame From the initial launch of Nuclide, one of our most unique features compared to most editors is tight integration with Mercurial. In this release, we take things one step further by adding support for hg blame. If your current file belongs to an Hg repository, you can right-click in the editor and choose Show Blame from the context menu to see the author and changeset ID for each line in a custom gutter. Right-click again and choose Hide Blame to remove the blame information. (Note that as stated in the README, this has been tested with Hg 3.4 and above.)
  • Diff View (works only for mercurial repositories): The Diff View now offers more help discovering the uncommitted source control changes by showing a highlighted up-to-date file status changes in a list on the right and changing the right pane editor to be editable and savable to allow previewing the changes in realtime and saving the contents to the filesystem, when ready. The Diff View is not yet able to detect and diff against the fork base of the source control to allow amend commit diffs iterating on the same commit. Usage (Open, Navigate, Edit and Save) is as follows:
    • For a certain file, either choose Open in Diff View from the context menu or Packages > Open Diff View from the Atom's menu bar.
    • Double-clicking changed files in the change list will load the diff for that file.
    • Edit the right pane editor with an updated change and see the diff in real-time.
    • (cmd-s) or File > Save will save the new contents to the filesystem.
  • Connection Status: We now display an icon in the status bar to indicate whether a file is local or remote, and whether the server is connected or disconnected.
  • nuclide-file-tree-deux We have heard your complaints about nuclide-file-tree loud and clear, so we have embarked on a rewrite of our file tree package named nuclide-file-tree-deux that uses proper React/Flux design patterns to ensure the tree is performant for the largest of directory structures. The new package has not reached feature parity with the existing nuclide-file-tree quite yet, so it is available for dogfooding, but it is not enabled by default. Once we have achieved feature parity, we will replace the code for nuclide-file-tree with that of nuclide-file-tree-deux and will delete the nuclide-file-tree-deux package.
  • New service framework One of the primary features of Nuclide is support for developing remote files. This is more than just the ability to edit files on a remote machine: it also means surfacing information from developer tools that are running locally on that machine, such as Buck, Mercurial, Clang, Hack, and Flow. To that end, we design our integration with these tools to use asynchronous APIs that are agnostic to where the tools are running. Our pattern for doing this has been continuously evolving, but our most recent implementation focuses on definition files written in JavaScript/Flow with RPCs that return promises or observables. Our nuclide-buck-base package is the first to take advantage of this new framework, but we plan to retrofit the rest of our codebase to use this soon. We also intend to document it so that you can define your own remote services that plug into nuclide-server.
  • New package: nuclide-move-pane The nuclide-move-pane package overrides the default behavior of cmd/ctrl-k arrow to move the pane rather than split it. If you want the original behavior, we encourage you to disable this package.
  • New package: nuclide-format-js This package is meant to help format JavaScript code in order to speed up development. Currently, it has support for automatically adding, removing, and sorting requires. It also understands Flow types and will promote or demote a require() when appropriate. This feature is still in an experimental phase: please see the README for the heuristics involved and how best to integrate it into your workflow. In the future we will be adding more AST based transformations that can be configured on and off.

• Bug Fix

  • Improve nuclide-ocaml behavior when ocamlmerlin is missing or errors. Previously this could cause a crash.

0.0.30

• Bug Fix
  • #163

0.0.29

• Bug Fix
  • Fixed the transpilation issue identified in #162.

0.0.28

• New Feature
  • Babel files in Node and Atom packages are now pre-transpiled prior to publishing. This should reduce initial startup time for nuclide-server as well as for Atom, post-Nuclide install. Both nuclide-server and Atom have schemes to cache their transpilations, but

0.0.26

• Bug Fix
  • #157

0.0.25

• New Feature
  • Was long overdue for a release. There were 222 commits since 0.0.24.
  • New Atom package: nuclide-arcanist.
  • New Atom package: nuclide-diagnostics-store.
  • New Atom package: nuclide-diagnostics-ui.
  • New Atom package: nuclide-find-references.
  • New Atom package: nuclide-ocaml.

0.0.24

• Bug Fix
  • Back out Facebook-specific code.