log.Flags（）是否应该与write共享相同的sync.Mutex？

Question

I have the following code:

package main

import(
    "log"
    "os"
)

type LogFilter struct {}

func (t *LogFilter) Write(p []byte) (int, error) {
    _ = log.Flags()
    return os.Stderr.Write(p)
}

func main() {
    log.SetOutput(&LogFilter{})
    log.Println("Hello, playground")
}

Which Deadlocks because of http://golang.org/src/pkg/log/log.go line 135 defers the lock until after the write. Which in the write I'm calling Flags which tried to get the lock.

Is there any reason that they (Write & Flags) should share the same mutex?

Answer 1

It shares the mutex, because the Logger also reads its internal flags field when writing output.

Granted it could be more granularly locked, but you'd have to make a good case to justify the added complication. If you need the flags in your filter, I would copy them in when you initialize your structure.