gemnasium / logrus-postgresql-hook Goto Github PK

In the readme log.Hooks.Add(hook) should be log.AddHook(hook)

Related to issue #4: Some months ago logrus package has added the TraceLevel logging level. I believe the Levels() method should be updated with the new level logrus.TraceLevel:

Hi,

While reading the code I noticed that the sync.RWMutex is only used with RLock, I could not find a corresponding Lock call.

RWMutex is doing states:
A RWMutex is a reader/writer mutual exclusion lock. The lock can be held by an arbitrary number of readers or a single writer.

The newEntry method is also not having a critical section, correct?

Am I missing something?

When method name is being logged, i.e. log.SetReportCaller(true) , the Caller field is lost in hook. This is because this field (Caller) is not copied when new entry is created.

The newEntry should be:

newEntry := &logrus.Entry{
		Logger:  entry.Logger,
		Data:    data,
		Time:    entry.Time,
		Level:   entry.Level,
		Message: entry.Message,
                Caller: entry.Caller,
	}

Currently the database will log for every event level except for panic, as seen here:
func (hook *Hook) Levels() []logrus.Level { return []logrus.Level{ logrus.FatalLevel, logrus.ErrorLevel, logrus.WarnLevel, logrus.InfoLevel, logrus.DebugLevel, } }

It would be ideal if this method can be provided externally with something like the following:

func (h *Hook) init() { h.Levels = func (hook *Hook) Levels() []logrus.Level { return []logrus.Level{ logrus.FatalLevel, logrus.ErrorLevel, logrus.WarnLevel, logrus.InfoLevel, logrus.DebugLevel, } } }

Then in my caller I can update Levels with something like:
hook := pglogrus.NewAsyncHook(db.DB, map[string]interface{}{}) hook.Levels = func() []logrus.Level { topLevel := int(settings.Log.DbLogLevel) + 1 return logrus.AllLevels[0:topLevel] } l.AddHook(hook)

Alternatively it would be good if we can assign levels at declaration to a slice, and that could be used during execution. Something like the following:

type Hook struct { ... LogLevels []Level }

Hello, I'm just opening this issue to let you know, the next release of logrus (v1.2.0) will add a new trace level named Trace below Debug.
Here is the PR for reference sirupsen/logrus#844
You may want to take into account this new level.

Hello,

I have 2 questions/concerns, which seem to be related:

1. I would like to ask about the rationale behind setting the *time.Ticker to 1 second when the new async hook is created
   

   logrus-postgresql-hook/postgresql_hook.go

   Line 77 in 102da18

   Ticker: time.NewTicker(time.Second),

and then overwriting this time ticker with a 'quicker' one in the Flush() method

2. When running the test with race condition on i.e. with -race flag, there is data racing problem. While one goroutine tries to write to hook.Ticker in Flush() method, another one want to read in fire() method, line 202.

Here is the output that I get when running test with race condition:

Thank you!

So you don't have to call "defer hook.Flush()" everywhere.