profile
viewpoint

andreypopp/autobind-decorator 1405

Decorator to automatically bind methods to class instances

Andarist/react-textarea-autosize 1239

<textarea /> component for React which grows with content

andreypopp/backbone.projections 112

backbone.projections is a set of projections for Backbone.Collection

andreypopp/backbone.viewdsl 41

Declarative view technology for Backbone

andreypopp/babel-plugin-ast-literal 11

Babel Plugin AST Literal

andreypopp/backbone.viewevents 8

Events for Backbone.View which can bubble up through view hierarchy.

andreypopp/backbone.module 5

Spine.Module but for Backbone

andreypopp/backbone.avgrund 3

Avgrund modal concept by @hakimel packaged as Backbone.View

andreypopp/asyncomplete-ale.vim 2

LSP completion source (via ALE) for asyncomplete.vim

andreypopp/akka 1

Akka Transactors

startedsteveruizok/perfect-arrows

started time in 2 days

startedarminbiere/kissat

started time in 5 days

pull request commentprometheusresearch/baseline-codebase

Baseline/rapid tests

Some weird failure on CI, triggered a re-run.

Oize

comment created time in 7 days

PR opened prometheusresearch/baseline-codebase

Reviewers
rex.sms: retry on 429 Too Many Requests

We probably should try an exponential backoff instead but 1 seconds default retry delay is ok, I think.

+20 -1

0 comment

2 changed files

pr created time in 7 days

PR opened prometheusresearch/baseline-codebase

rex.widget: fix autocomplete widget lookup

title default to id expression and if this id expression is not a text then ~ operator will fail.

+1 -1

0 comment

1 changed file

pr created time in 9 days

startedc-cube/poc-modular-io

started time in 10 days

startedutkarshkukreti/reaml

started time in 12 days

Pull request review commentprometheusresearch/baseline-codebase

core: Added ability to use sys_prefix and cwd substitutions in !include tags

 def include_path(self, node):                     "expected a file name, but found an empty node",                     node.start_mark)         basename = getattr(self.stream, 'name', None)-        filename = node.value+        filename = node.value.format(+            sys_prefix=sys.prefix,+            cwd=os.getcwd(),

nit: maybe hoist os.getcwd() to the top scope? (this is what we do in other places like PathVal.

jayclassless

comment created time in 15 days

startedocamllabs/vscode-ocaml-platform

started time in 16 days

startedzombodb/pgx

started time in 21 days

PR opened prometheusresearch/baseline-codebase

Reviewers
deploy: add support for managing SQL views

This PR takes a stab at implementing support for managing SQL views with rex.deploy.

The example usage:

- view: persons
  definition: |
    select name, email
    from user

- view: persons
  definition: |
    select name, email
    from user
    union
    select name, email
    from patient

- view: persons
  present: false

I've been trying to follow the code for tables but of course making this specific to the views.

Some notes/design questions appeared:

  • How things are introspected has changed, previously both tables and views were loaded as TableImage objects. Now we load views as ViewImage objects and we DON'T load view columns. I suspect this will make a different when querying views with HTSQL?

  • When building a model out of facts we need to check for name clashes between tables and views as they share the same namespace in the database.

  • Should we raise an error in case the same named view appears with different definitions? Or should we instead override the former with the latter (dangerous)?

  • Right now we load view definition with pg_get_viewdef function but I'm not sure if it is suitable for comparison with view definition in facts as db can probably mangle the original SQL. We should probably store view definition in a COMMENT in a database.

  • Do we need to support view defined with HTSQL? That would be cool.

  • Would be nice to be able to specify a pk for a view (required for HTSQL to query such view).

  • We need to track dependencies:

    As we use SQL for view definitions we might need to use SQL parser for that. I'm thinking of using https://pglast.readthedocs.io/en/latest/ which wraps PostgreSQL's parser.

    If we want to add HTSQL based view then we can use HTSQL to infer the dependencies.

    • view -> table, why?

      • So we can require to drop view before we drop the table. It is implemented for linked tables? Would be nice as it will allow to validate migrations w/o accessing a database.
    • view -> view, why?

      • so we can require to drop dependent views before we drop the view

      • so we can change definition of a view with dependents (in this case we need to drop and re-create a view and its dependents)

+441 -4

0 comment

7 changed files

pr created time in a month

push eventprometheusresearch/baseline-codebase

Andrey Popp

commit sha 07eb3712d8a11d7f0781f760576911f035df2c12

deploy: add support for managing SQL views This commit takes a stab at implementing support for managing SQL views with rex.deploy. The example usage: ``` - view: persons definition: | select name, email from user - view: persons definition: | select name, email from user union select name, email from patient - view: persons present: false ``` I've been trying to follow the code for tables but of course making this specific to the views. Some notes/design questions appeared: - How things are introspected has changed, previously both tables and views were loaded as TableImage objects. Now we load views as ViewImage objects and we DON'T load view columns. I suspect this will make a different when querying views with HTSQL? - When building a model out of facts we need to check for name clashes between tables and views as they share the same namespace in the database. - Should we raise an error in case the same named view appears with different definitions? Or should we instead override the former with the latter (dangerous)? - Right now we load view definition with `pg_get_viewdef` function but I'm not sure if it is suitable for comparison with view definition in facts as db can probably mangle the original SQL. We should probably store view definition in a COMMENT in a database. - Do we need to support view defined with HTSQL? That would be cool. - Would be nice to be able to specify a pk for a view (required for HTSQL to query such view). - We need to track dependencies: As we use SQL for view definitions we might need to use SQL parser for that. I'm thinking of using https://pglast.readthedocs.io/en/latest/ which wraps PostgreSQL's parser. If we want to add HTSQL based view then we can use HTSQL to infer the dependencies. - view -> table, why? - So we can require to drop view before we drop the table. It is implemented for linked tables? Would be nice as it will allow to validate migrations w/o accessing a database. - view -> view, why? - so we can require to drop dependent views before we drop the view - so we can change definition of a view with dependents (in this case we need to drop and re-create a view and its dependents)

view details

push time in a month

create barnchprometheusresearch/baseline-codebase

branch : baseline/deploy-view

created branch time in a month

startedgrain-lang/grain

started time in a month

issue openedperseas/Pyrseas

Should "dbtoyaml -t TABLE" extract enum types as well

I understand that dbtoyaml is modelled after pg_dump and pg_dump doesn't extract enums if you supply -t but I'm wondering if dbtoyaml should diverge here? What do you think.

My use case is that I have a tool to build new pg databases out of existing ones and one of the features I want is to selectively to dump/restore (selectively as by specifying a subset of tables). pg_dump doesn't work for me (for the reason above) but I was wondering if I can use dbtoyaml instead for that.

created time in a month

startedwillmcgugan/rich

started time in a month

push eventesy/esy

Paul Shen

commit sha 687232c0ac9038cc902f32c92a1aa92ecea4c887

fix Makefile version message

view details

push time in a month

PR merged esy/esy

Fix Makefile required version message

When running make bootstrap, this was showing the version that I had installed, not the version I needed.

+5 -4

1 comment

1 changed file

paulshen

pr closed time in a month

pull request commentesy/esy

Fix Makefile required version message

Oops, good catch. Thank you!

paulshen

comment created time in a month

push eventesy/esy.github.io

Andrey Popp

commit sha 405689d1a5cb0cbafc328c70fecd0fcde2bde4c0

Deploy website Deploy website version based on 4ebdb24a51f1b8ff6abc4ad3ee135676c5902959

view details

push time in a month

created tagesy/esy

tagv0.6.6

package.json workflow for native development with Reason/OCaml

created time in a month

push eventesy/esy

Andrey Popp

commit sha 4ebdb24a51f1b8ff6abc4ad3ee135676c5902959

v0.6.6

view details

push time in a month

push eventesy/esy

EduardoRFS

commit sha 99388fcb7c908b9254b05a0f017948da21045ab0

fix concurrency on windows * also added an ENV var to override it

view details

EduardoRFS

commit sha b2ec500562e4d9d8c428698413d2bc1097f07404

added build concurrency option * also uses NUMBER_OF_PROCESSORS only on Windows

view details

EduardoRFS

commit sha dd7b05739d3dc9db3bb02054bac572d804218b14

get --fetch-concurrency also from an env var

view details

Andrey Popp

commit sha e3b1ebe4f497b3d44b9e29bd246d36e7d02a27d8

Better build/fetch concurrency config (#1129) - Allow to specify fetch concurrency via `ESY__FETCH_CONCURRENCY` env var. - Allow to specify build concurrency via `ESY__BUILD_CONCURRENCY` env var and `--build-concurrency` command line option - Fix detecting number of CPUs on Windows (outside of cygwin).

view details

push time in a month

PR merged esy/esy

fix concurrency on windows

Problems

  • Windows doesn't support getconf on powershell,
  • There is no way to limit the number of parallel processes during build

Solutions

  • Windows has an env variable called NUMBER_OF_PROCESSORS for that
  • added the env variable called ESY_BUILD_CONCURRENCY
+43 -13

0 comment

5 changed files

EduardoRFS

pr closed time in a month

Pull request review commentesy/esy

fix concurrency on windows

 let version =   };  let concurrency = {-  /*** TODO: handle more platforms, right now this is tested only on macOS and Linux */-  let cmd = Bos.Cmd.(v("getconf") % "_NPROCESSORS_ONLN");-  switch (Bos.OS.Cmd.(run_out(cmd) |> to_string)) {-  | Ok(out) =>-    switch (out |> String.trim |> int_of_string_opt) {-    | Some(n) => n-    | None => 1+  let to_int =+    Option.bind(~f=n_str => n_str |> String.trim |> int_of_string_opt);+  let env_esy_build_concurrency =+    Sys.getenv_opt("ESY_BUILD_CONCURRENCY") |> to_int;+  let env_number_of_processors =+    Sys.getenv_opt("NUMBER_OF_PROCESSORS") |> to_int;++  let getconf_nprocessors = {+    let cmd = Bos.Cmd.(v("getconf") % "_NPROCESSORS_ONLN");+    switch (Bos.OS.Cmd.(run_out(cmd) |> to_string)) {+    | Ok(n_str) => n_str |> String.trim |> int_of_string_opt+    | Error(_) => None+    };+  };+  switch (env_esy_build_concurrency) {+  | Some(n) => n+  | None =>+    switch (getconf_nprocessors, env_number_of_processors) {

Hm... in this case we'll read NUMBER_OF_PROCESSORS on non Windows as well, do we want that? I think it's better to keep it only on Windows given that we are going to have a dedicated env var for configuring that.

EduardoRFS

comment created time in a month

startedConservatory/s7

started time in a month

Pull request review commentesy/esy

Docs: Explanations for missing sections

 It is advised to commit this file to a project's repository so that the build environment is reproducible and doesn't depend on the current state of package registries (either npm or opam). -## Environment+## Stores  -## Build environment+Esy maintains three kinds of artifacts - sources, builds and installed artfiacts - in three different locations. -## Command environment+### Sources -## Test environment+Source entries in the esy store are source code files and the recipies that build and install them bundled together. Change any of them, and esy creates a different entry for the source of a package. -## Package exported environment+If, for instance, build instructions of a package from opam are overridden (and the source files are not necessarily touched), esy places them in a different location. -## Build store+Conceptually, a source entry in the esy store is both it's source files and it's recipie.

nit: typo

Conceptually, a source entry in the esy store is both its source files and its recipe.
prometheansacrifice

comment created time in a month

Pull request review commentesy/esy

Docs: Explanations for missing sections

 It is advised to commit this file to a project's repository so that the build environment is reproducible and doesn't depend on the current state of package registries (either npm or opam). -## Environment+## Stores  -## Build environment+Esy maintains three kinds of artifacts - sources, builds and installed artfiacts - in three different locations. -## Command environment+### Sources -## Test environment+Source entries in the esy store are source code files and the recipies that build and install them bundled together. Change any of them, and esy creates a different entry for the source of a package.

nit: typo

Source entries in the esy store are source code files and the recipes that build and install them bundled together. Change any of them, and esy creates a different entry for the source of a package.
prometheansacrifice

comment created time in a month

Pull request review commentesy/esy

Docs: Explanations for missing sections

 installed: esy x Hello ``` +### Where is the binary executable?+When a compiled binary is run with `esy x Hello`, esy creates a local install sandbox with `/bin`, `/lib`, `/etc` and other directories found globally where binaries are meant to be installed. If you're curious, you could peek into them, running++```shell+esy echo #{self.install}+```++and inspect the contents yourself. You'll find the binaries in the `bin` directory.++However, they are not meant to be run directly as they could be missing the necessary [exported environment](./concepts.md) - it could be possible that the binary you created needs a dependency during the runtime. `esy x <your project binary>` is the recommended way to run them.++Checkout [concepts](./concepts.md) for more information.+ ## Rebuild the project  Hack on project's source code and rebuild the project: +```shell+esy +```++This will take care of updating dependencies as well as building the project. Same as running `esy install` and `esy build` sequentially.++If you're sure, dependencies haven't been updated, simply run `esy build`

I think we can remove this suggestion, esy install is really fast for no-op and if esy install is a no-op (dependencies has changed) then esy build will fail. So it's always better just to run esy.

prometheansacrifice

comment created time in a month

Pull request review commentesy/esy

Docs: Explanations for missing sections

 It is advised to commit this file to a project's repository so that the build environment is reproducible and doesn't depend on the current state of package registries (either npm or opam). -## Environment+## Stores  -## Build environment+Esy maintains three kinds of artifacts - sources, builds and installed artfiacts - in three different locations. -## Command environment+### Sources -## Test environment+Source entries in the esy store are source code files and the recipies that build and install them bundled together. Change any of them, and esy creates a different entry for the source of a package. -## Package exported environment+If, for instance, build instructions of a package from opam are overridden (and the source files are not necessarily touched), esy places them in a different location. -## Build store+Conceptually, a source entry in the esy store is both it's source files and it's recipie. -## Global build store+### Build store -## Local build store+Build store is simply location where build artifacts are stored. They can be local to a project or global (so that they can be reused). -## Package build location+Build artifacts (object files, bytecode files etc) stored as separate entries dependending on the instructions that build and install them, environment variables present during its build and of course the sources (source code + recipie).

nit: typo

Build artifacts (object files, bytecode files etc) stored as separate entries dependending on the instructions that build and install them, environment variables present during its build and of course the sources (source code + recipe).
prometheansacrifice

comment created time in a month

pull request commentocaml/ocaml-lsp

Fix #206: Select markdown language id based on document syntax

@giltho @opam/reason package should have it

bryphe

comment created time in a month

pull request commentocaml/dune

Add test for finalize/fork behavior

Anyway, @rgrinberg suggested that we plan a meeting to discuss all this and that seems like a good idea to me. I created a calendar event and invited you, let me know if the time doesn't suit you.

Sounds good. But I don't see the invite anywhere.

rgrinberg

comment created time in a month

startedhannesm/patch

started time in a month

pull request commentocaml/dune

Add test for finalize/fork behavior

To better illustrate my suggestion with making Future.t a ('a, exn) result — I'm attaching a gist https://gist.github.com/andreypopp/ac5533142c91be6756d2fecbbc497442 of a fork of fiber (along with a small test suite) I'm using in my app.

Few notes:

  • In the gist each forked fiber has its own execution context (EC). This is because EC has its own on_error which fills the ivar with Error exn in case of failure. This is needed so the fiber which suspends on waiting for future will terminate in case of failure (doesn't happen currently in dune's fibers).

    Though It's possible to implement parallel_iter (and possibly some others) in a more efficient way by sharing a single EC between forked fibers there, I didn't do it yet.

  • What I like about the current dune's fibers is that with_error_handler allows to attach a callback for errors issued from many forked fibers and this callback will be called promptly (as soon as an error happens, this is important, for example, for error reporting).

    Here in the gist this is lost (as you need to get the value out of the Future.t to re-raise the error). But I "recovered" this property in parallel_iter, see on_error callback there which will be called promptly on errors and allows to recover from them if needed.

  • See also

    val try_with : (unit -> 'b t) -> ('b, exn) Result.t t
    

    which I find useful and I'm not sure how to implement this "right" with dune's fibers now.

  • In general I find the case when a fiber has 1:1 correspondence with its result/failure easier to reason about than with the current dune's fibers where an error handler can collect multiple errors. But that maybe this is me not getting dune's fibers semantics.

  • As I understand that current dune's fibers allow more freely express concurrency (you can fork multiple fibers in any way you like and then wait for them to finish) while in my fork you need to issue Future.wait calls / use pre-defined primitives like parallel_iter to structure concurrency patterns.

rgrinberg

comment created time in a month

push eventesy/esy.github.io

Andrey Popp

commit sha cb1bcd01d7d42b5668f13e95ade5fa3b635265ba

Deploy website Deploy website version based on 217b38d4811e0392c974a11dfd9051d0e0190b4c

view details

push time in a month

created tagesy/esy

tagv0.6.5

package.json workflow for native development with Reason/OCaml

created time in a month

push eventesy/esy

Andrey Popp

commit sha 217b38d4811e0392c974a11dfd9051d0e0190b4c

v0.6.5

view details

push time in a month

push eventesy/esy

Andrey Popp

commit sha 96989292e92093318141280c43bc237d822a4612

site: add 0.6.4

view details

push time in a month

pull request commentocaml/dune

Add test for finalize/fork behavior

However, I found another issue: exceptions raised by forked fibers don't seem to be collected. So something else is going on.

The important bit is that collect_errors doesn't return before all the forked fibers have terminated.

That's exactly the problem — if some of forked fibers fails then one that waits on the corresponding future won't wake up and thus won't terminate — hence on_release won't be called.

rgrinberg

comment created time in a month

startedkeigoi/ocaml-mpst

started time in a month

startedUcombinator/parsing-with-zippers

started time in a month

startedmichaelforney/samurai

started time in a month

startedtummychow/git-absorb

started time in a month

startedliuchengxu/vista.vim

started time in 2 months

issue commentesy/esy

Esy is unable to resolve dependencies for Bap

Seems like another bug with constructing package universe for solver. cc @zindel

bbenne10

comment created time in 2 months

pull request commentesy-ocaml/libffi

add LD_LIBRARY_PATH and DYLD_LIBRARY_PATH to exportedEnv

@glennsl I think we should export $LD_LIBRARY_PATH still, what do you think?

glennsl

comment created time in 2 months

PR opened hamza0867/todo-backend-reasonml

Update $LD_LIBRARY_PATH for libffi

This fixes build on Linux. The proper fix would be to fix this at libff, looking into it.

+3 -0

0 comment

1 changed file

pr created time in 2 months

push eventandreypopp/todo-backend-reasonml

Andrey Popp

commit sha 3bd9201492f1d0062c70dd82988495aa366acab3

Update $LD_LIBRARY_PATH for libffi Proper fix would be to fix this at libffi.

view details

push time in 2 months

startedEduardoRFS/reason-css-parser

started time in 2 months

pull request commentocaml/dune

Add test for finalize/fork behavior

I wonder if the fix could be to make the Future.t to be

type 'a t = ('a, Exn_with_backtrace.t) Result.t Ivar.t

and then make it re-raise the exception (if any) on Future.wait.

rgrinberg

comment created time in 2 months

pull request commentocaml/dune

Add test for finalize/fork behavior

The test case:

let%expect_test "finalize/fork behavior (failure)" =
  let fiber =
    let log s = Fiber.return (print_endline s) in
    let open Fiber.O in
    Fiber.finalize
      ~finally:(fun () -> log "finally")
      (fun () ->
        let* f = Fiber.fork (fun () ->
          print_endline "fork";
          failwith "fork failure")
        in
        let* () = log "after fork" in
        let* () = Fiber.Future.wait f in
        log "fiber finished")
  in
  test unit fiber;
  [%expect
    {|
    fork
    after fork
    finally
    () |}]
rgrinberg

comment created time in 2 months

pull request commentocaml/dune

Add test for finalize/fork behavior

After applying patch the success case works as expected but if something within fork fails then finalize masks the exception:

let () =
  let open Fiber in
  let open Fiber.O in
  let log s = Fiber.return (print_endline s) in
  let finally () = log "finally" in
  let () =
    match run (finalize ~finally (fun () ->
      let* future = fork (fun () -> failwith "fork failure") in
      let* () = log "after fork" in
      let* () = Future.wait future in
      log "finished"
    )) with
    | None -> assert false
    | Some () -> ()
  in
  ()

This fails with Assert_failure. If I remove finalize then it shows the uncaught exception.

rgrinberg

comment created time in 2 months

starteddarrenldl/sandboxing

started time in 2 months

startedandrewchambers/hermes

started time in 2 months

startedhercules-team/augeas

started time in 2 months

startedjamii/imp

started time in 2 months

startedodis-labs/streaming

started time in 2 months

delete branch prometheusresearch/baseline-codebase

delete branch : baseline/ups

delete time in 2 months

push eventprometheusresearch/baseline-codebase

Andrey Popp

commit sha e202634561e638d210de6c547f31bb269d0eda16

rex-graphql: backport changes

view details

Andrey Popp

commit sha 8e4e69467499a2a6e92fd79c6e8d423d3edad829

rex-graphql-codegen: updates

view details

Andrey Popp

commit sha 1658cde3765d6a2c04e8a65ebb28b31d65e2a049

Remove rex-demo-ui code Leave only dummy page

view details

Andrey Popp

commit sha d9522304bad29bff0765defeddde90b10913b03c

flow-typed: update

view details

Andrey Popp

commit sha a445bc28bae606b65dc437821523dc87836d39f1

rex-ui: remove Router It wasn't used anywhere.

view details

Andrey Popp

commit sha 8181576ac2d3cd1de67fcf5a14da88a68af78e92

rex-ui: updates

view details

Andrey Popp

commit sha 1c1b2d2ef91d6649b00eeb20c71c52e066ba453b

rex-widget: updates due to rex-ui changes

view details

Andrey Popp

commit sha 0d15a9bee3b909c73a1a0a52434fb077a3665b56

js: fix package.json

view details

Andrey Popp

commit sha 5d64be508817ef7d27e3a70597f3d405e2115cc5

flow-typed: fix mui.grid props

view details

Andrey Popp

commit sha 8b5f73da6f77f5d915e89b4a385e5131fd4a2086

rex-ui: add missing deps

view details

Andrey Popp

commit sha ca4a047cf0b8f5a5b65de70952b2ca75996f6dd9

Updates to JS packages and rex.graphql #31

view details

push time in 2 months

PR merged prometheusresearch/baseline-codebase

Updates to JS packages and rex.graphql

This also removes most of rex-demo-ui as the code used for this was rotten bby recent updates and it uses a custom router we don't want to maintain.

+7538 -4762

0 comment

112 changed files

andreypopp

pr closed time in 2 months

PR opened prometheusresearch/baseline-codebase

Updates to JS packages and rex.graphql

This also removes most of rex-demo-ui as the code used for this was rotten bby recent updates and it uses a custom router we don't want to maintain.

+7538 -4762

0 comment

112 changed files

pr created time in 2 months

create barnchprometheusresearch/baseline-codebase

branch : baseline/ups

created branch time in 2 months

startedjserot/hocl

started time in 2 months

pull request commentneovim/neovim

[WIP] Add nvim_add_autocommand API

Unfortunately don’t have time to work on that right now, so if anyone could pick this up — that’d be great, the feature is very useful indeed.

andreypopp

comment created time in 2 months

startedcwida/duckdb

started time in 2 months

delete branch prometheusresearch/baseline-codebase

delete branch : baseline/webpack-config-updates

delete time in 3 months

push eventprometheusresearch/baseline-codebase

Konstantin Ryzhov

commit sha c5f8ce501eac3ad14329bb732b31a951990f3512

js: flow version updated

view details

Konstantin Ryzhov

commit sha 29bcb60aa9b5b6484532924cd7c8c5184a3d380a

webpack-config: Added imagemin and bundle analyzer plugins

view details

Konstantin Ryzhov

commit sha d3e0bd917214631f440a55731049f62658b75563

js: Reverted flow update

view details

Andrey Popp

commit sha 06c22a77ef4a818ce3f1f90c0ada1253b8eec5a7

Update webpack config with analyzer/image optimization #28

view details

push time in 3 months

PR merged prometheusresearch/baseline-codebase

Reviewers
Baseline/webpack config updates
  • [x] flow version updated
  • [x] Added imagemin and bundle analyzer plugins
+1357 -81

0 comment

3 changed files

theaklair

pr closed time in 3 months

startedarchseer/snippets.nvim

started time in 3 months

issue commentesy/esy

Unable to install esy on Ubuntu 20.04

Try sudo npm install -g —unsafe-perm esy Invocation. The unsafe-perm flag is needed when installing with sudo.

wesleybits

comment created time in 3 months

startedfacebookexperimental/Recoil

started time in 3 months

push eventprometheusresearch/baseline-codebase

Andrey Popp

commit sha 4d0231f108c3dedfb30f08a60084b8033b8ac8db

rex.graphql: backport changes

view details

push time in 3 months

startedleonerd/libtickit

started time in 3 months

issue commentmarlonrichert/zsh-autocomplete

Closing termnial window when pressing arrow up

@marlonrichert no, I'm not. Still a bit confusing why it closes the shell on "Up" in that case.

DanDobrick

comment created time in 3 months

issue commentmarlonrichert/zsh-autocomplete

Closing termnial window when pressing arrow up

Have fzf installed and in PATH but it happens for me as well, if I comment a line which binds an up arrow in this plugin then it works ok.

DanDobrick

comment created time in 3 months

startedhchbaw/auto-fu.zsh

started time in 3 months

pull request commentocaml/ocaml-lsp

Remove the dune hack

@giltho what's the failure with esy?

giltho

comment created time in 3 months

issue openedmarlonrichert/zsh-autocomplete

zsh-autocomplete overrides fzf configuration

First of all, thanks for the amazing completion experience with this plugin!

I'm trying this out now and found out it exports $FZF_* environment variables which override my customisations to fzf. I'm wondering why is it doing so and if it's needed — maybe it can be scoped to the plugin's usage of fzf somehow?

created time in 3 months

more