Skip to main content

A Go Gotcha that Got Me

Here's a little Go program that has a surprising output:
package main

import (
 "fmt"
 "net"
 "crypto/tls"
)
 
func main() {
 var c net.Conn

 c, err := tls.Dial("tcp", "www.jgc.org:80", nil)

 fmt.Printf("%v %v\n", c, err)

 if c == nil {
  fmt.Printf("Nil\n")
 } else {
  fmt.Printf("Not nil\n")
 }
}
This program tries to connect to my web site using TLS on the non-TLS port 80. That's done to force there to be a TLS error. The output is a little surprising:
  <nil> local error: record overflow
   Not nil
The Printf gives the value of c as but when the test c == nil is performed it's non-nil. 

So, what's going on? 

The answer is in the Go FAQ: Why is my nil error value not equal to nil?

In short, c is an interface (a net.Conn). The implementation of an interface is a type and a value. The type gives the actual type that implements that interface (in the case above it's a *tls.Conn) and the value is a pointer to the concrete example of that type. 

When the error occurs in the code above tls.Dial returns nil, err. The nil is a nil pointer to a tls.Conn. When the assignment to c happens c becomes non-nil (its type is now *tls.Conn) but its value is nil (since tls.Dial returned a nil pointer). Thus the nil test fails. 

The bottom line is that my code above is the wrong thing to do. Don't do nil tests on interface variables.

PS A lot of people have been asking me why I wasn't checking the value of err. In the real code I was, but a defer statement was acting on c and in it I had the code

if c != nil {
  c.Close()
}

In figuring out why that failed sometimes I discovered the gotcha.

Comments

Yves Junqueira said…
Most importantly, check the error result first before doing anything with the value.
Yves Junqueira said…
Most importantly, check the error result first before doing anything with the value.
rog peppe said…
Don't do nil tests on interface variables.

I'm not sure that's the right message to take home from this experience. Doing nil tests on interface variables is just fine. The thing to be aware of is that when you assign a non-interface value to an interface value, the interface value will never be nil.

This issue is why functions always return an error value of type "error" even when the type is known. The particular case you encountered is particularly easy to trip over because the signature of tls.Dial is almost exactly the same as that of net.Dial.

The particular pattern to watch for is:

x, err = something()

Note the "=" rather than :=.

It might be nice if there was a go vet check for this pattern actually. When go/types is ready, there may well be.

Popular posts from this blog

Your last name contains invalid characters

My last name is "Graham-Cumming". But here's a typical form response when I enter it:


Does the web site have any idea how rude it is to claim that my last name contains invalid characters? Clearly not. What they actually meant is: our web site will not accept that hyphen in your last name. But do they say that? No, of course not. They decide to shove in my face the claim that there's something wrong with my name.

There's nothing wrong with my name, just as there's nothing wrong with someone whose first name is Jean-Marie, or someone whose last name is O'Reilly.

What is wrong is that way this is being handled. If the system can't cope with non-letters and spaces it needs to say that. How about the following error message:

Our system is unable to process last names that contain non-letters, please replace them with spaces.

Don't blame me for having a last name that your system doesn't like, whose fault is that? Saying "Your last name …

All the symmetrical watch faces (and code to generate them)

If you ever look at pictures of clocks and watches in advertising they are set to roughly 10:10 which is meant to be the most attractive (smiling!) position for the hands. They are actually set to 10:09.14 if the hands are truly symmetrical. CC BY 2.0image by Shinji
I wanted to know what all the possible symmetrical watch faces are and so I wrote some code using Processing. Here's the output (there's one watch face missing, 00:00 or 12:00, because it's very boring):



The key to writing this is to figure out the relationship between the hour and minute hands when the watch face is symmetrical. In an hour the minute hand moves through 360° and the hour hand moves through 30° (12 hours are shown on the watch face and 360/12 = 30).
The core loop inside the program is this:   for (int h = 0; h <= 12; h++) {
    float m = (360-30*float(h))*2/13;
    int s = round(60*(m-floor(m)));
    int col = h%6;
    int row = floor(h/6);
    draw_clock((r+f)*(2*col+1), (r+f)*(row*2+1), r, h, floor(m…

Importing an existing SSL key/certificate pair into a Java keystore

I'm writing this blog post in case anyone else has to Google that. In Java 6 keytool has been improved so that it now becomes possible to import an existing key and certificate (say one you generated outside of the Java world) into a keystore.

You need: Java 6 and openssl.

1. Suppose you have a certificate and key in PEM format. The key is named host.key and the certificate host.crt.

2. The first step is to convert them into a single PKCS12 file using the command: openssl pkcs12 -export -in host.crt -inkey host.key > host.p12. You will be asked for various passwords (the password to access the key (if set) and then the password for the PKCS12 file being created).

3. Then import the PKCS12 file into a keystore using the command: keytool -importkeystore -srckeystore host.p12 -destkeystore host.jks -srcstoretype pkcs12. You now have a keystore named host.jks containing the certificate/key you need.

For the sake of completeness here's the output of a full session I performe…