reagentx/logria

A powerful CLI tool that puts log aggregation at your fingertips.

logria's Issues

Terminal not cleared for write to second Parser step, cursor set to end

This issue replicates on MacOS, but not under WSL.

Investigate Alert feature for preconfigured searches

Exiting regex entry does not clear input

1. Start to enter a regex
2. Press ⎋ to cancel
3. Press / to start another regex

After the first keystroke, all of the previously entered text will appear with the new character at the end.

Sometimes `stdout` is slow to receive streams when `stderr` is fine

Example to reproduce:

"""
Simple script to generate test logs
"""

import random
import time
import logging

def first():
    RANDINT = random.randint(1, 100)
    if RANDINT % 3 == 0:
        logging.warning('I am a first log! %s\x00', RANDINT)

def second():
    RANDINT = random.randint(1, 100)
    if RANDINT % 5 == 0:
        logging.error('I am a second log! %s\x00', RANDINT)
        
def third():
    RANDINT = random.randint(1, 100)
    if RANDINT % 7 == 0:
        logging.info('I am a third log! %s\x00', RANDINT)

logging.info('\x00I am the first log in the list! %s\x00', 0)
while True:
    first()
    second()
    third()
    RANDINT = random.randint(1, 100)
    print(f'I am standard output, {RANDINT}')
    time.sleep(0.001)

Confirm destructive actions from remove command

Since these commands remove data from the disk, we should double-check to avoid unwanted data loss.

App is rendering two blank lines instead of one

Refactor `String` `Err()` variants to their own Exceptions that implement `Display`

I wrote these as if they were Python, not Rust.

Split `LogiraConfig` into multiple config structs

Right now it is way too big and annoying to deal with.

`less`-like search instead of just filtering

Sometimes we want to see the context for a log, not just the logs that match a filter.

Page Up and Page Down scrolling is inverted

Subcommand to show current config directory and contents

• using defaults or env vars
• print paths
• print path contents

Spawn a separate process to find all the matches in the list of messages for regex search and parsers

Support piping to Logria

From the Python repo:

• Support optional piping as input stream - SO Link
  • Not possible to implement as Logria requires the user to be in control of stdin
  • stdin gets taken over by whatever we pipe to this program, and we cannot move that pipe away from stdin until the pipe finishes
  • We can overwrite the pipe with sys.stdin = open(0) however this will not work until the original pipe ends, which will never happen when tailing a stream

Entering command mode while a regex is active breaks regex rendering

Open app → RegEx search → type :

Pressing ⎋ returns to normal, but without the proper status:

Deeper instructions for parsers

Use single UserInputHandler for all other _Handler classes

Currently we initialize a new one every time we want to get input.

Sessions and Parsers display their fully qualified path, not just their names

Allow users to delete Parsers with `:r` when on the Parser setup screen

Investigate why app runs in the config directory

Using the below session, running Logria on Linux emits the user's config directory (~/.config) no matter the actual working directory:

{
  "commands": [
    "pwd"
  ],
  "stream_type": "Command"
}

Screen not cleared after regex search with no matches

1. Perform a regex search that has matches
2. Perform a second one that has no matches
3. The results of the first are not cleared

Use `MultipleChoiceHandler` in `StartupHandler`, not `UserInputHandler`

Right now the StartupHandler manually supports selecting items from a list when MultipleChoiceHandler already provides those features.

Screen not cleared when invoking a new Parser while a current Parser is running

Probably needs to call some stuff from return_to_normal() inside of reset().

Load session file from any arbitrary location, not just the default

One option is to invoke like logria --load [path] where path is a valid configuration template.

Scroll state should be an enum and not 3 booleans on a struct

Right now the scroll state is defined by three booleans in a LogiraConfig; it should be a single enum.

Update contribution guidelines with concrete rules about what PRs will be accepted

Subprocesses that want to look at stdin make the app freeze

This is caused by crossterm-rs/crossterm#692

If the first keystroke is `esc` then the cursor is placed in the incorrect spot

Clarify parser/pattern in docs

These terms are used interchangeably, which can be confusing.

No new messages alert shows in command line when it should not

If a buffer is empty and we swap to another buffer that has messages, the empty buffer alert is maintained.

Streams sometimes do not close with the app

Termination signal isn't getting properly sent to all of the threads.
See thread for details.

Cursor flickers when streams are moving very fast

Deeper instructions for sessions

Remove unit test files for parsers and sessions after assertions succeed

Startup should use `auxiliary_messages` instead of its own `startup_messages` vector

GitHub Actions run in infinite loop

Using the following configuration:

name: Test

on:
  pull_request:
    branches:
        - main
        - develop

env:
  CARGO_TERM_COLOR: always

jobs:
  Test:

    runs-on: ubuntu-latest

    steps:
    - uses: actions/checkout@v2
    - uses: actions-rs/toolchain@v1
      with:
        toolchain: stable
    - name: Lint
      uses: actions-rs/clippy-check@v1
      with:
        token: ${{ secrets.GITHUB_TOKEN }}
      continue-on-error: true
    - name: Test
      uses: actions-rs/cargo@v1
      with:
        command: test

GitHub Actions do not work. They run infinitely on tests that access the file system, which is most of them.

https://github.com/ReagentX/Logria/runs/2205703953?check_suite_focus=true