feed.nomath.org

astatusbar can be started with the option `--icons` and will replace sensor labels with appropriate icons. The user is expected to have Nerd Fonts, in particular IosevkaTerm Nerd Font, installed.

It appears as if lightyear inputs are supposed to be keys that are being held down. This abstraction does not fit well with our inputs, which are one-time actions. In particular, I've found inputs being missed. We hope to improve upon this, by sending inputs as lightyear messages via a reliable, oredered channel.

Previously, we tracked re-opening of issues by advancing the commit hash for scrambles that do not witness an issue. Advancing the commit hash for scrambles not witnessing an issue is obviously incorrect. To see that, consider a top/earlier commit that creates a new file with an issue in it. None of the bottom/later scrambles witness the issue. Hence, it would have been logged as being created in the bottom-most/latest commit instead of the top-most/earliest commit.

Requesting review from @kirchner for inclusion into `main`.

Select champions using the `--champion` flag. Valid values are `ranged` and `meele` (default).

Reviewed branch feature/review at commit 07aacbbe55f4793daa0c9893efe4a64951575081.

Reviewed branch feature/review at commit 07aacbbe55f4793daa0c9893efe4a64951575081.

Can be disabled by ending the quasi quotation template with a single backslash.

lgtm, possible splitting some non-business-logic stuff into separate commits would be nice commit 9a49ec0dcd6f75736949350844f85d80fe48a662

lgtm, possible splitting some non-business-logic stuff into separate commits would be nice commit 9a49ec0dcd6f75736949350844f85d80fe48a662

Prototype of the `review` command, cf. `anissue review -h`. Also adds the `status` command.

Prototype of the `review` command, cf. `anissue review -h`. Also adds the `status` command.

StateT should not be an instance of `MonadUnliftIO`. This reverts commit d441ab33582975ff9eda162db952f2ec5da59223.

Changes the heuristic for the probability map.

Tags can be removed by setting their value to "-".

This naive implementation scores tags based on the words that correlate to a tag being set. It then uses that score to determine the highest-scoring value for a tag based on the words in a document.

Disables caching.

This commit replaces the AST querying with manual tree traversal. @prerequisite-for add-languages-elm-shell-nix

This commit replaces the AST querying with manual tree traversal. @prerequisite-for add-languages-elm-shell-nix

This replaces the tree-sitter bindings with a call to a single C function that traverses the AST. We expect the query API to be slower than manually traversing the tree for this particular use case. This will be addressed in an upcoming commit. @prerequisite-for add-languages-elm-shell-nix

This replaces the tree-sitter bindings with a call to a single C function that traverses the AST. We expect the query API to be slower than manually traversing the tree for this particular use case. This will be addressed in an upcoming commit. @prerequisite-for add-languages-elm-shell-nix

Drop haskell-tree-sitter in favor of custom bindings to tree-sitter. haskell-tree-sitter is outdated and seems unmaintained. The implementation add low-level bindings to tree-sitter and traverses the AST in Haskell. We suspect that many FFI calls are more expensive than performing just a single API call to a C function that does the traversal. This will be addressed in upcoming commits. @prerequisite-for add-languages-elm-shell-nix

Drop haskell-tree-sitter in favor of custom bindings to tree-sitter. haskell-tree-sitter is outdated and seems unmaintained. The implementation add low-level bindings to tree-sitter and traverses the AST in Haskell. We suspect that many FFI calls are more expensive than performing just a single API call to a C function that does the traversal. This will be addressed in upcoming commits. @prerequisite-for add-languages-elm-shell-nix

This file is unused since the Haskell rewrite.

This file is unused since the Haskell rewrite.

This facilitates full-text search via: ``` anissue list --filter '@rawText /any.*text/' ```

This facilitates full-text search via: ``` anissue list --filter '@rawText /any.*text/' ```

Example: ``` anissue list --filter '@title /foo/' ```

Example: ``` anissue list --filter '@title /foo/' ```

This reverts commit 53623436f68e548b1a840ac0f7e1e4472270613d.

This reverts commit 53623436f68e548b1a840ac0f7e1e4472270613d.

Regresses in that we only support Haskell for now, as Elm, Nix or Bash are not available as tree-sitter-* Haskell packages.

Regresses in that we only support Haskell for now, as Elm, Nix or Bash are not available as tree-sitter-* Haskell packages.

Initial cache generation is slower, as we are losing out on parallelism.

Initial cache generation is slower, as we are losing out on parallelism.

This makes it more convenient to use `anissue` from within this project, as it always refers to the current build. Note that we don't use `alias` whithin `shellHook` for the same purpose, because direnv does not support aliases.

This makes it more convenient to use `anissue` from within this project, as it always refers to the current build. Note that we don't use `alias` whithin `shellHook` for the same purpose, because direnv does not support aliases.

Closed issues can be shown, but listing closed issues requires passing `--closed`.

Closed issues can be shown, but listing closed issues requires passing `--closed`.

`getCommitHashes` now returns at least one commit, and reverses commits by default (ie. oldest to newset).

`getCommitHashes` now returns at least one commit, and reverses commits by default (ie. oldest to newset).

At some point, we noticed a performance drop when generating the history. It turns out that per-file granularity is not performant anymore, presumably since we're analizing changed files. This restores performance by switching to per-commit granularity instead.

At some point, we noticed a performance drop when generating the history. It turns out that per-file granularity is not performant anymore, presumably since we're analizing changed files. This restores performance by switching to per-commit granularity instead.

This regression had been introduced in d9400635bcb28674c8510d71aa6eed94194bf669.

This regression had been introduced in d9400635bcb28674c8510d71aa6eed94194bf669.

Check for `../anissue\n` having been created by invocations prior to this fix.

Check for `../anissue\n` having been created by invocations prior to this fix.

We interpret the following types: - ISO8601-formatted strings as dates, - integer-formatted strings as integers, - all other strings as strings. If the filter value matches a format, it is cast into the respective data type. For comparison-based filtering, the tag's value is then cast into the same data type and comparison is performed type-aware. If the tag's value is not castable, we consider it not matching.

We interpret the following types: - ISO8601-formatted strings as dates, - integer-formatted strings as integers, - all other strings as strings. If the filter value matches a format, it is cast into the respective data type. For comparison-based filtering, the tag's value is then cast into the same data type and comparison is performed type-aware. If the tag's value is not castable, we consider it not matching.

The following filter expressions are now additionally valid: ``` --filter @tag <=VALUE --filter @tag <VALUE --filter @tag >=VALUE --filter @tag >VALUE ``` Note that negation needs some re-work. Currently, only `--filter !@tag <VALUE` is valid, and `--filter @tag !<VALUE` is not. The first version may not behave as you would expect it. A tracking issue has been created. Note that without typing tag values, all comparisons perform lexicographically on `String`s.

The following filter expressions are now additionally valid: ``` --filter @tag <=VALUE --filter @tag <VALUE --filter @tag >=VALUE --filter @tag >VALUE ``` Note that negation needs some re-work. Currently, only `--filter !@tag <VALUE` is valid, and `--filter @tag !<VALUE` is not. The first version may not behave as you would expect it. A tracking issue has been created. Note that without typing tag values, all comparisons perform lexicographically on `String`s.

The underlying problem had been truncating `String` to `ByteString` (via `IsString(fromString))`, which is unsafe.

The underlying problem had been truncating `String` to `ByteString` (via `IsString(fromString))`, which is unsafe.

Adds type-trickery akin to `Text.printf` to provide - (1) an abstraction over `fromString (printf ".." ..)` for `ProcessConfig`s, - (2) have arguments be quoted automatically. As string is the only argument type that Shell knows, the formatting character is simply "%".

Adds type-trickery akin to `Text.printf` to provide - (1) an abstraction over `fromString (printf ".." ..)` for `ProcessConfig`s, - (2) have arguments be quoted automatically. As string is the only argument type that Shell knows, the formatting character is simply "%".

Ditches `ghci` in favor of `cabal repl`.

Ditches `ghci` in favor of `cabal repl`.

Requires [$EDITOR integration](https://haskell-language-server.readthedocs.io/en/latest/configuration.html#configuring-your-editor).

Requires [$EDITOR integration](https://haskell-language-server.readthedocs.io/en/latest/configuration.html#configuring-your-editor).

The `hoogle` command is useful, among other things, for searching the used libraries for functions, ie: ```console hoogle 'a -> a' GHC.Prelude id :: a -> a .. ``` You can also start a web-server: ```console hoogle server ```

The `hoogle` command is useful, among other things, for searching the used libraries for functions, ie: ```console hoogle 'a -> a' GHC.Prelude id :: a -> a .. ``` You can also start a web-server: ```console hoogle server ```

Invoking `ghcid` allows us to watch for compilation errors while developing.

Invoking `ghcid` allows us to watch for compilation errors while developing.

The underlying problem is that tree-sitter does not allow for matching sibling nodes. So any issues comprised of line comments (vs. block comments) suffered from this issue, independent of the language (shell). Thus, we manually merge such sibling matches after running tree-grepper.

The underlying problem is that tree-sitter does not allow for matching sibling nodes. So any issues comprised of line comments (vs. block comments) suffered from this issue, independent of the language (shell). Thus, we manually merge such sibling matches after running tree-grepper.

The `list` command can be omitted and is the default. The `show` command retains past output. ```console $ anissue -h Usage: anissue anissue list anissue show ``` The implementation uses docopts, the shell variant of docopt [1]. Because of that, wrapping with `--argv0 ''` is necessary when packaging. [1] http://docopt.org/

The `list` command can be omitted and is the default. The `show` command retains past output. ```console $ anissue -h Usage: anissue anissue list anissue show ``` The implementation uses docopts, the shell variant of docopt [1]. Because of that, wrapping with `--argv0 ''` is necessary when packaging. [1] http://docopt.org/

This had been an oversight of 17f8148945bf5d636eb0ab2b90e3f0d1c57ecf3f.

This had been an oversight of 17f8148945bf5d636eb0ab2b90e3f0d1c57ecf3f.

its functionality has been refactored into extract.sh and helpers

its functionality has been refactored into extract.sh and helpers