inaka / sheldon Goto Github PK

for markdown adapter use emarkdown

Hi Team,

I have tested this sheldon spell checker utility. I did see, the the hyphen words are not suggested by the spell checker.
for example:
self-ownership
self-paid
self-painter
self-pampered
self-pampering
self-panegyric
self-parasitism
self-parricide
self-partiality
self-paying
self-peace

Is there something i am missing to validate this?.

Testing with spellingCI I saw 5 seconds is not good enough. We can increase the default timeout or/and make that number configurable from config file

I missed to remove some Elixir references

After #40

I see that the only existing dictionary is English, but I'd like to use it for Portuguese-BR. Seems like it would only need a bazinga.txt and a dictionary.txt, correct?

If I can provide these, would you merge the PR?

Thanks!

Adding Travis CI

rebar3_edoc_extensions

rpc:pmap/3 can throw exit(badrpc)

in this code

suggestions(MisspelledWords, Lang) ->
  try rpc:pmap({?MODULE, add_suggestions}, [Lang], MisspelledWords) of
    Result -> Result
  catch
    exit:badrpc:Stacktrace ->
      error_logger:error_msg( "~p:~p >> Error: badrpc~n\tStack: ~p"
                            , [?MODULE, ?LINE, Stacktrace]
                            ),
      MisspelledWords
  end.

we can remove the catch since the rpc:pmap/3 doesn't throw an exception

sheldon version

0.3.0

OS version

Windows

Steps to reproduce

Compile sheldon on Windows.

Current behavior

C:/ProgramData/Chocolatey/lib/mingw/tools/install/mingw64/bin/mingw32-make -C d:/a/rebar3_sheldon/rebar3_sheldon/_build/default/lib/hoedown libhoedown.a
mingw32-make[1]: Entering directory 'd:/a/rebar3_sheldon/rebar3_sheldon/_build/default/lib/hoedown'
cc -g -O3 -ansi -pedantic -Wall -Wextra -Wno-unused-parameter -Isrc -c -o src/autolink.o src/autolink.c
process_begin: CreateProcess(NULL, cc -g -O3 -ansi -pedantic -Wall -Wextra -Wno-unused-parameter -Isrc -c -o src/autolink.o src/autolink.c, ...) failed.
make (e=2): The system cannot find the file specified.
mingw32-make[1]: *** [Makefile:92: src/autolink.o] Error 2
mingw32-make[1]: Leaving directory 'd:/a/rebar3_sheldon/rebar3_sheldon/_build/default/lib/hoedown'
mingw32-make: *** [Makefile:33: ../priv/emarkdown.so] Error 2
mingw32-make: Leaving directory 'd:/a/rebar3_sheldon/rebar3_sheldon/_build/default/lib/emarkdown/c_src'
===> Hook for compile failed!

Expected behavior

The sheldon can be able compile on Windows.

the current version (2.2.3) of worker_pool is not compatible with erlang 20

sheldon version

0.3.0

OS version

Ubuntu

Description

• Motivation
  Not all users want to see bazinga in response from sheldon. It would be great have configuration for this behavior.
• Proposal
  Add new config option show_bazinga true/false to skip bazinga message from sheldon in response.

Current behavior

The bazinga field in response always provided.

Expected behavior

The bazinga field in response should be configurable.

sheldon version

0.4.1

OS version

MacOS 11.6

Description

• Motivation
  I want to be able to write hyphenated words.
• Proposal
  Analyze stuff like non-breaking as two words: non and breaking.

Current behavior

For instance…

src/spillway_srv.erl:51: The word "non-incrementing" in comment is unknown.

Expected behavior

No warnings.

Currently it complains about deprecated functions

Candidates is returning the misspelled word.

The line https://github.com/inaka/sheldon/blob/master/src/sheldon_dictionary.erl#L189 should be replaced by

[]

sheldon version

0.2.0

OS version

Ubuntu

Description

• Motivation
  Speeding up the library.
• Proposal
  • Create typo generator for generate typo_dictionary.txt based on dictionary.txt and put it into priv/lang/eng/* or move into separate repo, because the size of file can be large for eg hex.pm. As result for each word from dictionary.txt will be put into eg ets all possible typos. Which in turn will lead to the fact that all possible options will already be generated and available, the shadlon will not need to spend additional time on additional generation, words are either there or not.
  • Also don't forget about association of similar words which must be linked to each other thus one misspelled word can have several possible variants but no more 1 +- 5 or 10.

Current behavior

The sheldon try generate and provide a lot of candidates for each word in real time.

Expected behavior

All possible candidates should be insert before sheldon will be loaded/started.

Config

It might be worth adding an option to control whether the new behavior is turned on or off.

sheldon version

0.2.1

OS version

Ubuntu

Steps to reproduce

1> sheldon:check("this comment: ’").
** exception error: bad argument
     in function  re:split/2
        called as re:split([116,104,105,115,32,99,111,109,109,101,110,116,58,
                            32,8217],
                           "\n")
        *** argument 1: not an iodata term
     in call from sheldon:do_check/2 (/.../sheldon/src/sheldon.erl, line 58)

Current behavior

The sheldon crashes when put unicode chars.

Expected behavior

The sheldon work with unicode chars.

Try to go up to 100%

sheldon version

0.2.0

OS version

Ubuntu

Description

• Motivation
  An incorrect 4-character word can create about +- 250 possible suggestions. In the current implementation, the generating of suggestions happens each time for any duplicates words which ultimately increases the load to system.
• Proposal
  Add separate table for already generated suggested suggestions as for success as for empty suggestions for each word what was already checked.

Current behavior

Cache of suggested suggestions is missed.

Expected behavior

Cache of suggested suggestions is missed implemented.

Config

Optional: it might be worth considering an additional option to enable and disable the cache of suggested suggestions.

sheldon is spending almost 40 seconds when starts. The problem is converting from the keys list to a set when starting the dictionary

Adding the integration Travis + Codecov