profile
viewpoint
Lily Ballard lilyball Twitch San Francisco, CA https://blog.eridi.us iOS Developer at Twitch. Open source contributor. Programming language aficionado. She/her.

apple/swift 50650

The Swift Programming Language

apple/swift-evolution 10801

This maintains proposals for changes and user-visible enhancements to the Swift Programming Language.

apple/swift-package-manager 7535

The Package Manager for the Swift Programming Language

apple/swift-corelibs-foundation 3763

The Foundation Project, providing core utilities, internationalization, and OS independence

apple/swift-corelibs-libdispatch 1733

The libdispatch Project, (a.k.a. Grand Central Dispatch), for concurrency on multicore hardware

apple/swift-corelibs-xctest 791

The XCTest Project, A Swift core library for providing unit test support

apple/swift-llbuild 748

A low-level build system, used by Xcode and the Swift Package Manager

apple/swift-lldb 642

This is the version of LLDB that supports the Swift programming language & REPL.

issue commentsoc/directories-rs

Config dir on macOS should not be ~/Library/Preferences

Apple is tightening up system stuff, including the root volume, but the only restrictions they've placed on what a user can do in their own home folder is adding explicit permissions to accessing certain sensitive data (such as contacts database). There's no motivation for them to prevent users from creating dotfiles and I would be shocked if they ever did it.

As for NSUserDefaults / CFPreferences, if you want to use those then use those. That's not related to directories-rs though. Those aren't APIs for storing arbitrary configuration, they're APIs for a specific key-value preferences store.

I'm not going to sign up to build stuff that guarantees that people will come running my way when things break a few years down the road, because my lib happens to show up last in their stacktrace.

Stuffing arbitrary stuff into ~/Library/Preferences is pretty much the only behavior being discussed that has any chance of breaking a few years down the road, because locking down that folder to just cfprefsd is pretty much the only applicable restriction I could see Apple implementing. Using ~/Library/Application Support is what we should be doing if we're using Apple's folder structure (and will always be supported), and using XDG is what we probably should actually be doing for cross-platform CLI apps because that's the unix way, and both approaches can be expected to be supported for the indefinite future.

lilyball

comment created time in 2 days

Pull request review commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

 let     '';      patches = [-      # Fixes "Integer 243 in '243 (243)' followed by non-digit" error with systemctl completion.-      # https://github.com/fish-shell/fish-shell/issues/5689 (will be included in fish 3.1.0)-      (fetchpatch {-        url = "https://github.com/fish-shell/fish-shell/commit/c6ec4235136e82c709e8d7b455f7c463f9714b48.patch";-        sha256 = "02m6pkhhx6y21csydznsxkbpnwhcpzyz99xgd9ryh7s03v7wbigw";-      })+      # Fixes compilation on old Apple SDKs+      ./old-sdk-compilation.patch

The PR has been merged (and is slated for 3.1.1) so fetching the commit is definitely safe (fetching the PR is safe at this point too, but fetching the commit is just as good).

cole-h

comment created time in 3 days

delete branch lilyball/nixpkgs

delete branch : tokei

delete time in 4 days

Pull request review commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

 let       sed -r "s|command grep|command ${gnugrep}/bin/grep|" \           -i "$out/share/fish/functions/grep.fish"       sed -i "s|which |${which}/bin/which |"               \-              "$out/share/fish/functions/type.fish"+             "$out/share/fish/functions/type.fish"       sed -e "s|\|cut|\|${coreutils}/bin/cut|"             \           -i "$out/share/fish/functions/fish_prompt.fish"       sed -e "s|gettext |${gettext}/bin/gettext |"         \           -e "s|which |${which}/bin/which |"               \           -i "$out/share/fish/functions/_.fish"       sed -e "s|uname|${coreutils}/bin/uname|"             \           -i "$out/share/fish/functions/__fish_pwd.fish"   \-            "$out/share/fish/functions/prompt_pwd.fish"+             "$out/share/fish/functions/prompt_pwd.fish"       sed -e "s|sed |${gnused}/bin/sed |"                  \           -i "$out/share/fish/functions/alias.fish"        \-            "$out/share/fish/functions/prompt_pwd.fish"+             "$out/share/fish/functions/prompt_pwd.fish"       sed -i "s|nroff |${groff}/bin/nroff |"               \-            "$out/share/fish/functions/__fish_print_help.fish"+             "$out/share/fish/functions/__fish_print_help.fish"       sed -i "s|/sbin /usr/sbin||" \

As previously mentioned the line was modified in 3.1.0 to be

set -lx -a PATH /usr/local/sbin /sbin /usr/sbin

So this is going to leave it as

set -lx -a PATH /usr/local/sbin

To that end, this should be modified as

      sed -i "s|/usr/local/sbin /sbin /usr/sbin||" \
cole-h

comment created time in 4 days

Pull request review commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

 let             "$out/share/fish/functions/prompt_pwd.fish"       sed -i "s|nroff |${groff}/bin/nroff |"               \             "$out/share/fish/functions/__fish_print_help.fish"-      sed -i "s|/sbin /usr/sbin||" \-            "$out/share/fish/functions/__fish_complete_subcommand_root.fish"

This particular sed in question is mucking with the completions for sudo and doas. I wish I knew why Fish thought it had to muck with the PATH in general for completions of those two commands.

cole-h

comment created time in 4 days

pull request commentfish-shell/fish-shell

Work around compilation failure with old Apple SDKs

@floam I don't know which SDK added the #include <sys/time.h> to <sys/proc.h> so I don't feel comfortable making it conditional.

lilyball

comment created time in 4 days

Pull request review commentfish-shell/fish-shell

Work around compilation failure with old Apple SDKs

 #include <procfs.h> #endif #if __APPLE__+#include <sys/time.h> // Required to build with old SDK versions

I'm not set up to run IWYU so I don't know if it would even flag this. We do use something from this file, it's just that in newer SDKs it's redundant because <sys/proc.h> includes it.

lilyball

comment created time in 4 days

push eventlilyball/nixpkgs

volth

commit sha 1649f3eb78b8b84370f2396d9f499d4ed8e08fbf

[cpan2nix] perlPackages.DBIxClassHelpers: 2.034001 -> 2.034002

view details

volth

commit sha 08bea27f1c78dc6f50ecc8e26e44918540f9f2e9

[cpan2nix] perlPackages.NetAmazonS3: 0.86 -> 0.87

view details

volth

commit sha 7bbdd029d77b4d4b9d82b204eb35deedd38a226d

[cpan2nix] perlPackages.TaskFreecellSolverTesting: cleanup

view details

volth

commit sha c8f26a956cece0fe7964e72cd1badd961badafd4

fix tests

view details

volth

commit sha 923878af0dcaef5153fa8ea8333b913a8ab629b5

perlPackages.XMLLibXSLT: 1.96 -> 1.97

view details

volth

commit sha 77260174832ea1b2bb9cbbb87943f95b4bafc6a3

[cpan2nix] perlPackages.CDB_File: 0.99 -> 1.01

view details

volth

commit sha 46fd509d8741728b5aa9d84cee6d4a8b9afcafad

[cpan2nix] perlPackages.CPANPerlReleases: 5.20191220 -> 5.20200120

view details

volth

commit sha 9056ad0fa2d6d55c01e771f6ae91451a9c7eb125

[cpan2nix] perlPackages.Error: 0.17028 -> 0.17029

view details

volth

commit sha 8f7cf15296ea51bb75cb95b95d53d513316ddb23

[cpan2nix] perlPackages.ExtUtilsMakeMaker: 7.42 -> 7.44

view details

volth

commit sha f2c7af88ba1c93a8c147b716e4b7747295091b10

[cpan2nix] perlPackages.IO: 1.39 -> 1.42

view details

volth

commit sha 881b5a42f456c11e5e3c202aca735a606c55af98

[cpan2nix] perlPackages.IOStringy: 2.112 -> 2.113

view details

volth

commit sha c9da52270fe603a51ee2be3866d6891e24c7e456

[cpan2nix] perlPackages.IOTty: 1.12 -> 1.14

view details

volth

commit sha b3d3481853117c8d29d74274a0e9c16fa61e8900

[cpan2nix] perlPackages.IPCSystemSimple: 1.25 -> 1.26

view details

volth

commit sha ea8e6a8b1922d15a5fd0eadf5642fdbcbd82cbca

[cpan2nix] perlPackages.Mojolicious: 8.30 -> 8.32

view details

volth

commit sha 761f82683f6598ba6c6df21f1221b571f089f66f

[cpan2nix] perlPackages.NetIdent: 1.24 -> 1.25

view details

volth

commit sha ee8cf0a62d2189e0ba2a06c7d2857d84e1656830

[cpan2nix] perlPackages.PathTiny: 0.108 -> 0.110

view details

volth

commit sha 386412fc82dba8329016451211a119033057af7b

[cpan2nix] perlPackages.RegexpParser: 0.22 -> 0.23

view details

volth

commit sha 6e70fad7ec2f9913096dea558d05cf800b991d86

[cpan2nix] perlPackages.SetObject: 1.39 -> 1.40

view details

volth

commit sha aaf5b22a9ef949e59a7d675a17f10445f9fe8219

[cpan2nix] perlPackages.TemplateToolkit: 3.003 -> 3.007

view details

volth

commit sha 165f39a31af87ef0977b1c14252de57d522dd5e4

[cpan2nix] perlPackages.TestSimple13: 1.302170 -> 1.302171

view details

push time in 4 days

create barnchlilyball/nixpkgs

branch : tokei

created branch time in 4 days

PR opened NixOS/nixpkgs

tokei: 10.1.1 -> 10.1.2
Motivation for this change

https://github.com/XAMPPRocky/tokei/releases/tag/v10.1.2

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [X] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+3 -3

0 comment

1 changed file

pr created time in 4 days

Pull request review commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

 let             "$out/share/fish/functions/prompt_pwd.fish"       sed -i "s|nroff |${groff}/bin/nroff |"               \             "$out/share/fish/functions/__fish_print_help.fish"-      sed -i "s|/sbin /usr/sbin||" \-            "$out/share/fish/functions/__fish_complete_subcommand_root.fish"

It looks like __fish_complete_subcommand_root.fish got effectively inlined into share/completions/doas.fish and share/completions/sudo.fish, so instead of deleting this, we should retain it but run it on those two files instead.

Though looking further, those files also gained an inclusion of /usr/local/sbin as well. Or more specifically, both files have the following line:

set -lx -a PATH /usr/local/sbin /sbin /usr/sbin

I admit I'm not clear as to why these completions are modifying the PATH, or why we're then un-modifying it.

cole-h

comment created time in 4 days

pull request commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

Well, I suppose fetching the PR as a patch opens us up to hash changes. I'm not sure what the best solution is prior to the PR being merged; we could fetch the commit, but then if the PR's history is edited the commit isn't guaranteed to stay.

Or we could just copy the patch locally.

cole-h

comment created time in 4 days

pull request commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

I've filled a PR with Fish as https://github.com/fish-shell/fish-shell/pull/6602. We should be able to just fetch and apply https://github.com/fish-shell/fish-shell/pull/6602.patch as a patch.

cole-h

comment created time in 4 days

PR opened fish-shell/fish-shell

Work around compilation failure with old Apple SDKs

When building fish-shell with the macOS 10.12 SDK, <sys/proc.h> does not include <sys/time.h> but references struct itimerval. This causes a compilation failure if we don't import <sys/time.h> ourselves.

This was previously masked by an import of <sys/sysctl.h>, which was removed in fc0c39b6fdb6.

This fixes a build failure when building Fish using Nix, as Nix is currently on the macOS 10.12 SDK.

+1 -0

0 comment

1 changed file

pr created time in 4 days

create barnchfish-shell/fish-shell

branch : old_sdk_compilation

created branch time in 4 days

create barnchlilyball/fish-shell

branch : old_sdk_compilation

created branch time in 4 days

push eventlilyball/fish-shell

Fabian Homborg

commit sha 7d3bec12973e9e1839337a05d3088b22cbadd5f1

test.fish: Notice littlecheck failures This failed to propagate the exit status. Sorry!

view details

Fabian Homborg

commit sha d775c584a56333c6ac744ec139d9df9860de32f7

Port realpath tests to littlecheck I'm gonna add more tests to this and I don't want to touch the old stuff. Notice that this needs to have the output of the complete_directories test adjusted because this one now runs later. That's something we should take into account in future.

view details

Fabian Homborg

commit sha 4ae18951ee5e5a0a645b8f9521a0d9c4548d6e1d

Add NetBSD's error to the bad-option test I'm kinda close to just checking `.*fish: .*option.*Z.*`

view details

Fabian Homborg

commit sha 8013686d17c13d9e3363ffa1ccd55f96cc39b2cb

delete-or-exit: Only exit if the commandline is empty This exitted if the cursor was at the end of the line as well (i.e. if delete-char failed). That's a bit too eager. Also documentation, which should have already been included.

view details

Fabian Homborg

commit sha 87b05d3111c726f19c975e5efca5c90ecc201021

Port eval tests to littlecheck

view details

Fabian Homborg

commit sha f1fa8d5f5b3f997deda944f09f7a1efa40ae07aa

Port complete tests to littlecheck

view details

Fabian Homborg

commit sha 65eb4cd9d340f40e054997464907ce80959e2224

eval tests: Fix path I tested this manually (`littlecheck.py -s fish=fish tests/checks/eval.fish`) from the base directory, which means I got "tests/checks/eval", while the real test gets "checks/eval". I then reran `make test_fishscript`, but that didn't pull in the updated test - we should really handle that better.

view details

Leonard Hecker

commit sha ca5427865d0ad7be5564d41a16173d177a51bb02

Fixed invalid assignments in fish_print_git_action

view details

ridiculousfish

commit sha 8282369f45c92f01ed49107d159657a2f0b4fbec

child_setup_process to stop passing the process child_setup_process only cares about whether we are in a forked child, not the entire process structure. Narrow the parameter.

view details

ridiculousfish

commit sha b1a1b617f15655168709c8a69df663320af5fc75

child_setup_process to accept new termowner directly Soon we will have more complicated logic around whether to call tcsetpgrp. Prepare to centralize the logic by passing in the new term owner pgrp, instead of having child_setup_process perform the decision.

view details

ridiculousfish

commit sha bc0329f77588988d2a852f6ccbe9c154349cc303

Add wcstolower Converts a string to lowercase. Eliminates some loops that did this explicitly.

view details

ridiculousfish

commit sha 0e4fe53718a2396c6ebf4c1d2858a4b18700ccac

clang-format some files

view details

tomKPZ

commit sha 42f38bb806e2f278cf28b36412e1ff5d8ae91f65

Fix status code coloring for nim prompt

view details

Peter Lithammer

commit sha 9d47c46a7a6950331b4adf31bdfc488d0f625361

Suppress shell function lookup for ssh during scp completion This prevents functions or aliases to run during tab completion of remote files for the `scp` command.

view details

ridiculousfish

commit sha cdcea0f00dfc286f8397eac8018ba526bd0f01a5

Port setenv tests to littlecheck

view details

Mahmoud Al-Qudsi

commit sha 7ee675afcf486609caf03018b54d1ccd20efec07

Update `aws` and `aws s3` completions Improve handling of switches interjected between the commands, and support for completing buckets automatically for secondary+ parameters to `aws s3`.

view details

ridiculousfish

commit sha d577eb4aaa3d25bb46984287e352ecffb7ac6fe7

Only use the global fish_complete_path and fish_function_path Prior to this fix, fish would attempt to react if a local fish_complete_path or fish_function_path were set. However this has never been very well tested and will become impossible with concurrent execution. Always use the global values.

view details

ridiculousfish

commit sha 8181883111f17aa053b1668efcd9f1b1aa1dd1eb

Minor refactoring of logic around when a job wants to claim the terminal Introduce should_claim_terminal() which encapsulates an && exprsesion which was previously repeated a lot.

view details

Mahmoud Al-Qudsi

commit sha 57508b46d2ddc994e8a9fe9f17df9956155f7ba1

[docs] Fix build warning plus the rendering of the escaped space Use a `:code:` role for the escaped space, it avoids the nastiness of the whitespace mangling in inline literals.

view details

Mahmoud Al-Qudsi

commit sha 90882ca9eeddcf740be2ec99e70e65eb9f2d1a4f

[docs] Clarify behavior of `--no-scope-shadowing` and variable lifetimes

view details

push time in 4 days

created taglilyball/fish-shell

tag3.1.0

The user-friendly command line shell.

created time in 4 days

created taglilyball/fish-shell

tag3.1b1

The user-friendly command line shell.

created time in 4 days

pull request commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

The failure was introduced by fc0c39b6fdb685ee64121ad58aa68194b3b35588, which removed an import of <sys/sysctl.h>, which was indirectly importing <sys/time.h>.

cole-h

comment created time in 4 days

pull request commentNixOS/nixpkgs

fish: 3.0.2 -> 3.1.0

Interestingly, sys/proc.h on my macOS 10.15 system has #include <sys/time.h> (struct itimerval is defined in that header) but /nix/store/0xqs5h1n9avficlxbhh299bphpmnpbhk-Libsystem-osx-10.12.6/include/sys/proc.h is missing that include. That missing include is the proximate cause for the error, so perhaps we can work around it by inserting an #include <sys/time.h> at the right spot, but I'll continue investigating.

cole-h

comment created time in 5 days

delete branch lilyball/nixpkgs

delete branch : cocoapods-beta

delete time in 10 days

create barnchlilyball/nixpkgs

branch : cocoapods-beta

created branch time in 11 days

PR opened NixOS/nixpkgs

cocoapods-beta: 1.9.0.beta.2 -> 1.9.0.beta.3
Motivation for this change

https://github.com/CocoaPods/CocoaPods/releases/tag/1.9.0.beta.3

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [X] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+19 -19

0 comment

2 changed files

pr created time in 11 days

pull request commentpostmates/PMKVObserver

Updating to Swift 5.

I really wish CocoaPods could mark up its generated pbxproj files to suppress those warnings.

In any case I'll revisit this soon. I just had a baby so I'm low on time at the moment. I want to make sure I'm comfortable with the versioning strategy when I do this (whether I decide to bump the major version or just declare that Swift version changes a sufficient time after Xcode has been updated are safe).

emisvx

comment created time in 11 days

issue commentCanop/broot

This shouldn't use ~/Library/Preferences on macOS

If soc/directories isn't going to update then it would be a good idea to do it here. I hope they're willing to make the change in the library though.

lilyball

comment created time in 11 days

issue commentfish-shell/fish-shell

Parser should accept `echo {foo`

[ 1 -eq 1 ] isn't literal text though, it's a different syntactic structure entirely.

Something else to consider: should it also be the case that a sole } is treated as a literal character?

I would say yes. Braces should be treated just like other literal text unless they contain a comma or variable expansion. A lone } can't contain anything.

I wrote in the issue text that echo {foo, should still be an error as that's attempted parameter expansion. At this point I'm on the fence and could be swayed either way about that. But echo {foo and echo foo} should both be legal.

lilyball

comment created time in 17 days

issue commentfish-shell/fish-shell

Parser should accept `echo {foo`

I can't think of any other case where a single job comprised entirely of literal text becomes invalid if you delete a character off the end.

And besides, square brackets don't always have to come in pairs. I'm not actually sure why echo foo[bar isn't legal, I didn't realize that and think it should be changed, but echo [foo is perfectly legal.

lilyball

comment created time in 17 days

issue commentapollographql/apollo-tooling

Installed apollo-codegen-swift doesn't match version at time of apollo publish

We're doing things from the CLI directly. We use npm install apollo@2.21.0 because we have a custom script that deals with it.

lilyball

comment created time in 18 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.4.0

 nodePackages // {     name = "bitwarden-cli-${drv.version}";   }); +  insect = nodePackages.insect.override {+    nativeBuildInputs = [ pkgs.psc-package pkgs.purescript nodePackages.pulp ];

Ok, but we don't want to step on its toes if buildNodePackage is ever updated in the future to start using nativeBuildInputs.

teto

comment created time in 18 days

issue openedrust-lang/rust

rust/RELEASES.md doesn't list the Box<T> FFI change for 1.41.0

The announcement post for Rust 1.41.0 talks about Box<T> now being ABI-compatible with C's T* type, but I can't find any mention of that in the detailed release notes.

created time in 19 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.4.0

 nodePackages // {     name = "bitwarden-cli-${drv.version}";   }); +  insect = nodePackages.insect.override {+    nativeBuildInputs = [ pkgs.psc-package pkgs.purescript nodePackages.pulp ];

I'm not sure if node2nix sets up any default nativeBuildInputs but just in case this really should be

  insect = nodePackages.insect.override (drv: {
    nativeBuildInputs = drv.nativeBuildInputs or [] ++ [ pkgs.psc-package pkgs.purescript nodePackages.pulp ];
  });
teto

comment created time in 20 days

issue openedapollographql/apollo-tooling

Installed apollo-codegen-swift doesn't match version at time of apollo publish

The version of the apollo-codegen-swift dependency doesn't match the version documented in the CHANGELOG for a given apollo release. Or rather, the CHANGELOG actually seems to be documenting the wrong versions of dependencies (at least, of apollo-codegen-swift) to begin with, but even taking that into account, when I install apollo@2.21.0 today I get a different version of apollo-codegen-swift than I did when version 2.21.0 was released, and what I get matches the version expected from apollo@2.21.3.

I'm not quite sure how this happens, this repo uses file references for the dependencies but it seems when they're published to the registry it turns into pinned versions with no package lock, and so apollo@2.21.0 seems to have a declared dependency on "apollo-codegen-swift": "^0.35.12", which then resolves (today) to version 0.35.15 at install time.

This is a problem because it causes inconsistent behavior when running apollo between different members of my team based on when they installed the tool. We probably should just vendor the apollo tool into our repo, but for the time being we simply require everyone to install the same version (and verify this version in our tooling), but that doesn't protect us from apollo-codegen-swift changes that affect existing source, such as adding type annotations (https://github.com/apollographql/apollo-tooling/pull/1638).

My expectation is that since apollo is a tool rather than a library, that installing it should lock all of its dependencies to precise versions (or at the very least all of the apollo dependencies).

created time in 22 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

  {pkgs ? import <nixpkgs> {     inherit system;-  }, system ? builtins.currentSystem, nodejs ? pkgs."nodejs-13_x"}:+  }, system ? builtins.currentSystem, nodejs ? pkgs."nodejs-8_x"}:

This needs to be fixed again.

teto

comment created time in 22 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

 let         sha512 = "nne9/IiQ/hzIhY6pdDnbBtz7DjPTKrY00P/zvPSm5pOFkl6xuGrGnXn/VtTNNfNtAfZ9/1RtehkszU9qcTii0Q==";       };     };-    "ajv-6.10.2" = {

We probably don't need to commit changes to the v12 or v13 package sets when you're just trying to update the v10 one.

teto

comment created time in 22 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

 nodePackages // {     '';   }; +  insect = nodePackages.insect.override {+    buildInputs = [ nodePackages.pulp nodePackages.bower pkgs.purescript ];+  };
  insect = nodePackages.insect.override (drv: {
    nativeBuildInputs = drv.nativeBuildInputs or [] ++ [ pkgs.psc-package pkgs.purescript nodePackages.pulp ];
  });
teto

comment created time in 24 days

Pull request review commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

 , "hueadm" , "imapnotify" , "indium"+, "insect"

Update this to v5.4.0, which gets rid of the bower dependency

teto

comment created time in 24 days

pull request commentNixOS/nixpkgs

swiftformat: init at 0.40.11

This passes nix-review on my macOS 10.14.6 machine.

https://github.com/NixOS/nixpkgs/pull/67226
1 package built:
swiftformat
bdesham

comment created time in 24 days

Pull request review commentNixOS/nixpkgs

swiftformat: init at 0.40.11

+{ stdenv, lib, fetchFromGitHub, runCommand }:++# This derivation is impure: it relies on an Xcode toolchain being installed+# and available in the expected place. The values of buildSymlinks,+# sandboxProfile, and hydraPlatforms are copied pretty directly from the MacVim+# derivation, which is also impure.++let+  buildSymlinks = runCommand "swiftformat-build-symlinks" {} ''+    mkdir -p $out/bin+    ln -s /usr/bin/xcodebuild $out/bin

Do we actually need this? You're not invoking some Makefile that references xcodebuild or anything like that, you're just invoking xcodebuild directly in buildPhase, so I would imagine you could just invoke /usr/bin/xcodebuild instead.

bdesham

comment created time in 24 days

pull request commentNixOS/nixpkgs

swiftformat: init at 0.40.11

The PR title needs to be updated

bdesham

comment created time in 24 days

pull request commentNixOS/nixpkgs

swiftformat: init at 0.40.11

You may also want to consider borrowing macvim's sandboxProfile as well, which means you can build it locally with sandbox = relaxed. The sandboxProfile in this case will still block network access, and is written to block access to /usr/local (to avoid accidental dependencies on Homebrew).

bdesham

comment created time in 25 days

pull request commentNixOS/nixpkgs

swiftformat: init at 0.40.11

The build relies on Xcode, so yes, it will fail in the sandbox.

@bdesham You can set meta.hydraPlatforms = []; to avoid having hydra waste time building it. This is what macvim does.

bdesham

comment created time in 25 days

pull request commentpostmates/PMKVObserver

Updating to Swift 5.

I've been a bit leery to do this because Carthage has no way to support multiple swift versions and so I feel like making a change like this requires a major version bump. I may just be overly cautious though.

emisvx

comment created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

@truh The instructions are higher up in this thread, but basically, your /etc/synthetic.conf should have a line just consisting of /nix (nothing else), your volume should have the label "Nix", and your /etc/fstab should have a line instructing the system to mount that volume at /nix. I don't remember the precise formulation of the line, but it's in this thread somewhere.

With these instructions, if your volume is encrypted then GUI apps will start to restore before it finishes mounting (this is a problem if Terminal.app is one of these apps and you use a Nix-installed shell as your login shell). If your volume is unencrypted then it will mount prior to the GUI session logging in. That said, I assume this caveat applies to your current setup as well (where you're symlinking to /Volumes/NixStore).

mroi

comment created time in a month

issue openedsharkdp/insect

Can't install without bower

bower is marked as a devDependency but this package has a postinstall script that invokes it. So if I don't have bower installed on my system and I run npm install insect it fails at the postinstall command.

Is the terminal install of this tool supposed to rely on bower, or is that just for the web version?

created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

I've been running with the separate volume approach for a while now with zero problems (though I keep the volume unencrypted so it mounts prior to GUI initialization; I don't use it to build work source, just for tooling support). As I understand it the only real holdup with getting this into the installer is testing it on a wide variety of preconditions.

mroi

comment created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

@emptyflask The particular workaround you linked to is the symlink approach, which appears to break a lot of packages. I strongly recommend using the separate volume approach instead (documented many times in this thread).

mroi

comment created time in a month

issue commentlilyball/rust-lua

Travis badge points to old repository

Thanks, I just fixed it. This repo is pretty old so I'm surprised anyone is even still looking at it.

palfrey

comment created time in a month

push eventlilyball/rust-lua

Lily Ballard

commit sha 2233b2444930a31c8448bf34fe3fbd33d2fea524

Update my name and email address Fixes #31.

view details

push time in a month

issue closedlilyball/rust-lua

Travis badge points to old repository

The badge in the README points to https://travis-ci.org/kballard/rust-lua not https://travis-ci.org/lilyball/rust-lua

closed time in a month

palfrey

issue commentNixOS/nix

adding support for a nix.fish script along side $HOME/.nix-profile/etc/profile.d/nix.sh

(If you don't use a fish plugin manager you can just copy the script into your own ~/.config/fish/conf.d folder as well

cartazio

comment created time in a month

issue commentNixOS/nix

adding support for a nix.fish script along side $HOME/.nix-profile/etc/profile.d/nix.sh

@golddranks AFAIK the other approaches just get you the nix environment and don't get you the rest of the fish stuff (e.g. completions, vendor scripts, etc). My plugin was designed for precisely your setup: https://github.com/lilyball/nix-env.fish

cartazio

comment created time in a month

issue commentrust-lang/cargo

Features cannot enable/disable features of dependencies

This particular issue was about controlling features of indirect dependencies. I do not believe there's any way to disable features in general (except for the flag to avoid enabling default features). I'm actually surprised that I can't find an issue specifically about this. I too have run into this problem, where I'd like to say "give me default features except for foo" but instead I have to say "give me bar and baz" and hope that the list of default features doesn't change without my notice.

lilyball

comment created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

The login hook appears to execute before the GUI session is initialized (or rather, before GUI apps are launched). I haven't tried it myself but that's what I infer based on people using it to force-mount the encrypted partition.

AFAIK LaunchDaemons do not block GUI app initialization.

mroi

comment created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

@nycnewman This approach means the disk isn't mounted prior to user login session setup. We already have that problem with an encrypted Nix partition, but an unencrypted partition is mounted prior to GUI app restoration, and there's a workaround using a login hook for forcing the encrypted partition to mount first. The DMG approach doesn't seem to support this at all, which is rather unfortunate as it causes problems with GUI app restoration depends on stuff installed from Nix (such as restoring Terminal.app when your login shell is from Nix).

mroi

comment created time in a month

pull request commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

I noticed there's a bower2nix command but it seems like it's supposed to be used as part of a larger derivation so I don't know how to package up the output.

teto

comment created time in a month

pull request commentNixOS/nixpkgs

nodesPackages.insect: init at 5.3.0

I'd love to see insect be packaged but I don't understand anything about bower or how to make this work.

teto

comment created time in a month

issue commentogham/exa

Add option to print current colors in LS_COLORS and EXA_COLORS format

@alivefromupstairs I'm not sure what you mean. If you're using LS_COLORS already you can just set EXA_COLORS to the same thing, and then start adding in any exa customizations you want. If you don't have either variable set, consult the manpage to see what the defaults are.

dandavison

comment created time in a month

delete branch lilyball/nixpkgs

delete branch : cocoapods-beta

delete time in a month

startedsharkdp/insect

started time in a month

issue commentkoalaman/shellcheck

Add support for #!/usr/bin/env nix-shell

I would really love to see shellcheck gain native support for this, so I don't end up with errors in VS Code every time I edit a shell script that uses nix-shell. In my own scripts I can add shell directives, but I'm often opening scripts I don't control.

coretemp

comment created time in a month

issue commentNixOS/nixpkgs

Update default attributes to latest version available

nodejs should be updated. It's currently at 10.x but the recommended LTS is 12.x, and 10.x goes to maintenance mode in a few months.

matthewbauer

comment created time in a month

issue closedNixOS/nixpkgs

nodejs-8.15.0: missing binary download for darwin

Issue description

The nodejs-8.15.0 package doesn't seem to have a binary download for darwin. Both of my machines are currently building nodejs from scratch right now and this is very surprising.

Steps to reproduce

Install nodejs-8.15.0

In my particular case, I had nodejs-8.14.1 installed and I ran nix-env -u, but I would imagine a fresh install of nodejs would also do the trick.

Technical details

  • system: "x86_64-darwin"
  • host os: Darwin 18.2.0, macOS 10.14.2
  • multi-user?: no
  • sandbox: no
  • version: nix-env (Nix) 2.1.3
  • channels(kevin): "nixpkgs-19.03pre167241.2a858b3f766"
  • nixpkgs: /Users/kevin/.nix-defexpr/channels/nixpkgs

closed time in a month

lilyball

issue commentNixOS/nixpkgs

nodejs-8.15.0: missing binary download for darwin

I haven't noticed this happening in a while.

lilyball

comment created time in a month

issue commentNixOS/nix

adding support for a nix.fish script along side $HOME/.nix-profile/etc/profile.d/nix.sh

Actually one caveat: If the $NIX_PROFILES env var is somehow already set by the time the shell starts, it isn't necessary. I don't run with multi-user nix myself so I don't know all the details of it, but I assume it doesn't have a way to set up the environment prior to the shell's execution.

cartazio

comment created time in a month

issue commentNixOS/nix

adding support for a nix.fish script along side $HOME/.nix-profile/etc/profile.d/nix.sh

To be precise, I think non-NixOS systems that install nix with --daemon (multi-user mode) might not need this either?

They do. The mechanism that means this isn't necessary is the presence of /etc/fish/nixos-env-preinit.fish, a file that sources the Nix environment (and in particular sets up the $NIX_PROFILES env var).

This file is set up by the nixos fish module, not by the fish package (which makes sense because the fish package can't write to the system /etc dir).

cartazio

comment created time in a month

push eventlilyball/nix-env.fish

Lily Ballard

commit sha cf99a2e6e8f4ba864e70cf286f609d2cd7645263

Update README.md Tweak the wording a bit.

view details

push time in a month

issue commentNixOS/nix

adding support for a nix.fish script along side $HOME/.nix-profile/etc/profile.d/nix.sh

@analyticd You're right, the README is slightly misleading. What I was trying to say is that Nix-installed Fish on NixOS doesn't need this, but it does everywhere else, including Nix-installed Fish on systems other than NixOS. I'll tweak the wording.

cartazio

comment created time in a month

issue commentsoc/directories-rs

Config dir on macOS should not be ~/Library/Preferences

@squarism That line is the data dir. Look one line up, the config dir is Library/Preferences.

lilyball

comment created time in a month

issue commentsoc/directories-rs

Config dir on macOS should not be ~/Library/Preferences

I agree that XDG-style paths are preferable for CLI apps, though there would need to be some way to get the desired macOS-style paths for GUI apps.

lilyball

comment created time in a month

issue commentsoc/directories-rs

Config dir on macOS should not be ~/Library/Preferences

When this fix is made, there will also need to be some API that provides access to the old path, so tools can use it to migrate data to the correct location.

lilyball

comment created time in a month

issue commentCanop/broot

This shouldn't use ~/Library/Preferences on macOS

I filed the upstream issue as https://github.com/soc/directories-rs/issues/62

lilyball

comment created time in a month

issue openedsoc/directories-rs

Config dir on macOS should not be ~/Library/Preferences

This library is vending ~/Library/Preferences as the config dir on macOS. This is a bad idea. Apple's own documentation explicitly states

You should not create files in this directory yourself. Instead, use the NSUserDefaults class or CFPreferences API to get and set preference values for your app.

Any configuration files written by the app without using one of those two APIs belongs in ~/Library/Application Support instead.

created time in a month

issue commentsoc/directories-rs

Support for using ~/.config on Mac for command line apps

Cross-platform CLI apps that write to ~/Library on macOS feel wrong to me. I would really like my CLI apps to use ~/.config even on macOS. Using ~/Library/Caches is much more reasonable since, as a user, I don't have to care about that, and it won't get backed up by Time Machine. But configuration and user data should go into the appropriate CLI location.

lavifb

comment created time in a month

issue openedCanop/broot

This shouldn't use ~/Library/Preferences on macOS

Describe the bug After running broot, it wrote out data to a folder ~/Library/Preferences/org.dystroy.broot/. This is very surprising. The ~/Library/Preferences/ folder contains cfprefsd plists, and it's not somewhere that tools should be writing to directly. I'm also surprised to see it doing this at all since this is a command-line application, not a GUI application, and so it should just use the XDG Base Directory Specification locations instead.

If you really want to write to a Mac-like directory, then ~/Library/Application Support/org.dystroy.broot would be the place to use. But I'd really prefer it to just use the XDG dirs instead, e.g. ~/.config/broot/.

To Reproduce Steps to reproduce the behavior:

  1. Run broot

Expected behavior Any configuration it writes out should go to a sane location.

Configuration (please complete the following information):

  • OS: macOS 10.15.2 (19C57)
  • Version 0.11.4

Additional context It looks like this directory is coming from the directories library. Apple's documentation on ~/Library/Preferences very explicitly says you should not be creating your own files in that directory but only using NSUserDefaults or CFPreferences, and that ~/Library/Application Support is the right place to put configuration files. I will be filing an issue against directories as well, but in the meantime broot should switch to using data_dir() instead of config_dir() on macOS.

I'm not sure what the migration path for existing configuration will end up being if directories fixes their code, since you'll need a way to find the old path in order to migrate stuff. Hopefully the directories library can figure this out.

created time in a month

issue commentNixOS/nix

/nix will not be writable on macOS Catalina

Using a DMG like that caps the size of the store though, doesn't it? It looks like your script is capping it at 50GB. Using a separate APFS volume means the store can grow as long as there's unused disk space.

mroi

comment created time in a month

starteddtolnay/cxx

started time in a month

delete branch lilyball/nixpkgs

delete branch : jazzy

delete time in a month

PR opened NixOS/nixpkgs

jazzy: 0.13.0 -> 0.13.1
Motivation for this change

https://github.com/realm/jazzy/releases/tag/v0.13.1

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [X] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+18 -18

0 comment

2 changed files

pr created time in a month

push eventlilyball/Tomorrowland

Lily Ballard

commit sha d3ac9ca65f2d478b0821abb6e846d78a8548f14c

Update Makefile comment

view details

push time in a month

create barnchlilyball/nixpkgs

branch : jazzy

created branch time in a month

Pull request review commentNixOS/nix

nix search: allow `-a` for attribute-only searches

 struct CmdSearch : SourceExprCommand, MixJSON                 throw SysError("cannot rename '%s' to '%s'", tmpFile, jsonCacheFileName);         } -        if (results.size() == 0)-            throw Error("no results for the given search term(s)!");+        if (!json) {+            if (results.size() == 0)+                throw Error("no results for the given search term(s)!"); -        RunPager pager;-        for (auto el : results) std::cout << el.second << "\n";+            RunPager pager;+            for (auto el : results) std::cout << el.second << "\n";

After taking another look at the code, it appears the real answer is "when json == true, results is always empty".

Ma27

comment created time in a month

startedrust-crdt/rust-crdt

started time in a month

Pull request review commentNixOS/nix

nix search: allow `-a` for attribute-only searches

 struct CmdSearch : SourceExprCommand, MixJSON                 throw SysError("cannot rename '%s' to '%s'", tmpFile, jsonCacheFileName);         } -        if (results.size() == 0)-            throw Error("no results for the given search term(s)!");+        if (!json) {+            if (results.size() == 0)+                throw Error("no results for the given search term(s)!"); -        RunPager pager;-        for (auto el : results) std::cout << el.second << "\n";+            RunPager pager;+            for (auto el : results) std::cout << el.second << "\n";

Don't we need to run this line when json == true as well?

Ma27

comment created time in a month

pull request commentNixOS/nix

nix search: allow `-a` for attribute-only searches

Commit message and PR title reference -a but the implementation uses -A.

Ma27

comment created time in a month

pull request commentNixOS/nix

[WIP] install: configure and bootstrap synthetic.conf on darwin

Oh sorry, I thought you meant you got a failure running these instructions.

LnL7

comment created time in a month

pull request commentNixOS/nix

[WIP] install: configure and bootstrap synthetic.conf on darwin

@brightredchilli What's the failure?

LnL7

comment created time in a month

delete branch lilyball/nixpkgs

delete branch : ffsend

delete time in 2 months

PR opened NixOS/nixpkgs

ffsend: 0.2.57 -> 0.2.58

Motivation for this change

https://github.com/timvisee/ffsend/releases/tag/v0.2.58

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [X] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+3 -3

0 comment

1 changed file

pr created time in 2 months

create barnchlilyball/nixpkgs

branch : ffsend

created branch time in 2 months

pull request commentstarship/starship

feat: Add module for showing exit codes of shell pipeline

It looks like the empty line has 2 stray spaces in it, which rustfmt wants to delete.

nrabulinski

comment created time in 2 months

delete branch lilyball/nixpkgs

delete branch : tokei

delete time in 2 months

PR opened NixOS/nixpkgs

tokei: 10.1.0 -> 10.1.1
Motivation for this change

https://github.com/XAMPPRocky/tokei/releases/tag/v10.1.1

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [X] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+3 -3

0 comment

1 changed file

pr created time in 2 months

create barnchlilyball/nixpkgs

branch : tokei

created branch time in 2 months

pull request commentNixOS/nixpkgs

Python package set major updates

What's the status of this? git-revise is currently broken because of importlib-metadata.

FRidh

comment created time in 2 months

CommitCommentEvent

issue commentapollographql/apollo-ios

Naming error when generating mutation in Apollo Swift Codegen

  1. Right, but the only call to singularization I could find was the same place where it was being pascal-cased. Unless the graphql library is also doing singularization itself, which would be kind of odd.
andre-felipe

comment created time in 2 months

delete branch lilyball/nixpkgs

delete branch : bat

delete time in 2 months

PR opened NixOS/nixpkgs

bat: insert less into PATH
Motivation for this change

The default pager for bat is less -RF --no-init. Overriding this to less -RF in order to enable alternate screen behavior (e.g. page up/down keys) causes the -F flag to behave badly in the version of less shipped on macOS. The simplest way to fix this is to simply put Nix's less into the PATH so it takes precedence.

Things done

<!-- Please check what applies. Note that these are not hard requirements but merely serve as information for reviewers. -->

  • [ ] Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
  • Built on platform(s)
    • [ ] NixOS
    • [X] macOS
    • [ ] other Linux distributions
  • [ ] Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • [ ] Tested compilation of all pkgs that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review wip"
  • [X] Tested execution of all binary files (usually in ./result/bin/)
  • [ ] Determined the impact on package closure size (by running nix path-info -S before and after)
  • [X] Ensured that relevant documentation is up to date
  • [X] Fits CONTRIBUTING.md.
+10 -3

0 comment

1 changed file

pr created time in 2 months

create barnchlilyball/nixpkgs

branch : bat

created branch time in 2 months

more