#51 Lint methods called without colon 22 days ago

Ticket created by ~andreyorst on ~xerool/fennel-ls

When declaring methods with self as the first argument, I sometimes forget that I need to call them with a : instead of .:

(local Foo {})

(fn Foo.new [self x]
  (setmetatable {:x x} self)

(Foo.new 42) ;; warning: Foo.new called without self argument, use (Foo:new 42)

A warning would be helpful in this scenario.

#259 fennelview doesn't respect multiline for associative tables with one key 26 days ago

Comment by ~andreyorst on ~technomancy/fennel

Pushed a v2 patch: https://lists.sr.ht/~technomancy/fennel/patches/58761

#259 fennelview doesn't respect multiline for associative tables with one key 28 days ago

Comment by ~andreyorst on ~technomancy/fennel

This particular case is no longer possible:

>> {:sequence ["fooooooooooooooooo" "baaaaaaaaaaaaaaaaar" "baaaaaaaaaaaaaaaaaz" "quuuuuuuuuuuuuuuuux"]}
{:sequence ["fooooooooooooooooo"
            "baaaaaaaaaaaaaaaaar"
            "baaaaaaaaaaaaaaaaaz"
            "quuuuuuuuuuuuuuuuux"]}

Previously, it would violate the line-length rule if the key was too long:

{:seeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeequenceeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee ["fooooooooooooooooo"
                                                                             "baaaaaaaaaaaaaaaaar"
                                                                             "baaaaaaaaaaaaaaaaaz"
                                                                             "quuuuuuuuuuuuuuuuux"]}

#259 fennelview doesn't respect multiline for associative tables with one key 28 days ago

Comment by ~andreyorst on ~technomancy/fennel

More examples for the context.

Before this change, associative tables that had only one key, and a really long value would not be split into multiple lines despite the :line-length setting in the view options:

>> {:foo "baaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaar"}
{:foo "baaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaar"}

After this change, a newline is properly included:

>> {:foo "baaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaar"}
{:foo
 "baaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaar"}

This slightly changes how nested data structures are printed:

>> {:sequence ["fooooooooooooo" "baaaaaaaaaaaaar" "baaaaaaaaaaaaaz" "quuuuuuuuuuuuux"]}
{:sequence
 ["fooooooooooooo" "baaaaaaaaaaaaar" "baaaaaaaaaaaaaz" "quuuuuuuuuuuuux"]}

Here, the sequence itself doesn't pass the threshold, but the overall map does.

Here, however, both the map and the sequence are above the threshold:

>> {:sequence ["fooooooooooooooooo" "baaaaaaaaaaaaaaaaar" "baaaaaaaaaaaaaaaaaz" "quuuuuuuuuuuuuuuuux"]}
{:sequence
 ["fooooooooooooooooo"
  "baaaaaaaaaaaaaaaaar"
  "baaaaaaaaaaaaaaaaaz"
  "quuuuuuuuuuuuuuuuux"]}

This means there would be less indentation in the nested data structures that was previously introduced by long keys.

#259 fennelview doesn't respect multiline for associative tables with one key 28 days ago

Comment by ~andreyorst on ~technomancy/fennel

The patch is here: https://lists.sr.ht/~technomancy/fennel/patches/58734

#259 fennelview doesn't respect multiline for associative tables with one key 28 days ago

Ticket created by ~andreyorst on ~technomancy/fennel

Fennelview breaks into multiple lines if the line is too long by default, howerver it doesn't work for tables with single key:

>> {:a {:b {:c {:d {:e {:f {:g {:h {:i {:j {:k {:l {:m {:n {:o {:p {:q {:r {:s {:t {:u {:v {:w {:x {:y {:z 42}}}}}}}}}}}}}}}}}}}}}}}}}}
{:a {:b {:c {:d {:e {:f {:g {:h {:i {:j {:k {:l {:m {:n {:o {:p {:q {:r {:s {:t {:u {:v {:w {:x {:y {:z 42}}}}}}}}}}}}}}}}}}}}}}}}}}

Instead it would be more appropriate to insert a newline after the key to respect the line length:

>> {:a {:b {:c {:d {:e {:f {:g {:h {:i {:j {:k {:l {:m {:n {:o {:p {:q {:r {:s {:t {:u {:v {:w {:x {:y {:z 42}}}}}}}}}}}}}}}}}}}}}}}}}}
{:a
 {:b
  {:c
   {:d
    {:e
     {:f
      {:g
       {:h
        {:i
         {:j
          {:k
           {:l
            {:m
             {:n
              {:o {:p {:q {:r {:s {:t {:u {:v {:w {:x {:y {:z 42}}}}}}}}}}}}}}}}}}}}}}}}}}

I have a patch ready for that.

#256 ,find command doesn't work with long filenames a month ago

Comment by ~andreyorst on ~technomancy/fennel

Fixed in 0bb9837

REPORTED RESOLVED CLOSED

#257 Special forms arglist should list the arguments individually a month ago

Comment by ~andreyorst on ~technomancy/fennel

I think the first one makes more sense, though it is easy to misinterpret it as destructuring. The second one kinda implies that the form should be written as (each k v (pairs tbl)) which is not right

#256 ,find command doesn't work with long filenames a month ago

Ticket created by ~andreyorst on ~technomancy/fennel

the ,find REPL command uses sourcemap, which stores a shortened path, resulting in incorrect path information when source is found:

>> (local fennel (require :fennel))
nil
>> ,find fennel.list
/data/data/com.termux/files/home/.local/bin/fennel:5733
>> (local uuid (require :io.gitlab.andreyorst.uuid))
nil
>> ,find uuid.uuid?
...34eaf90b70ddb0d14ea254/src/io/gitlab/andreyorst/uuid.fnl:96

This breaks editor integration, as the editor can't find the ...34eaf90b70ddb0d14ea254/src/io/gitlab/andreyorst/uuid.fnl path, which should be .deps/git/io.gitlab.andreyorst/uuid.fnl/209ff4f3a70ba8354034eaf90b70ddb0d14ea254/src/io/gitlab/andreyorst/uuid.fnl

#254 ,reload unloads macros from macro-loaded when using both init.fnl and init-macros.fnl 2 months ago

Comment by ~andreyorst on ~technomancy/fennel

So the main issue here is:

When we reload a macro module that was loaded previously by loading a non-macro module, it is removed from the macro-loaded table by the else branch of reload. When the module name is init.fnl and macros are stored in the same directory in init-macros.fnl, it makes sense to remove it from macro-loaded as well, but I think it should be done prior to reloading the module, so it would reappear after reloading, as reloading a module will reload the macro module too.

So the bug is that it doesn't reappear. I've sent a patch that fixes that, and also restores the module in case of an error when reloading the runtime module: https://lists.sr.ht/~technomancy/fennel/patches/57968