Tag Archives: Tech

Sending mail from a script on a Raspberry Pi

I’m working on a project where I need to send email from my Raspberry Pi. Installing a full-blown SMTP server would be overkill, I just need something where I can send messages from a bash script.

A brief search led me to a forum post from 2013 which talked about configuring the ssmtp package. That post in turn referenced a step-by-step guide from 2009. Unfortunately, both seem to be out of date, and the latter is for installing it on CentOS?RHEL/RedHat/Fedora. So here’s my attempt at an updated version for the Pi (which should apply to any Debian-based Linux distribution).


  • These instructions send via Gmail. If you’re using two-factor authentication (and you really should), you’ll need to set up an application -specific password. Otherwise, you’ll get authentication errors.
  • The password is stored in plain text. This solution is not suitable for use on a shared system.

The Steps

sudo apt update -y && sudo apt upgrade -y
sudo apt install -y ssmtp
sudo vi /etc/ssmtp/ssmtp.conf

Make these changes to the ssmtp.conf file


I also set the root= setting to my email address. I don’t believe this is necessary, but it does allow me to get notified when something goes wrong with one of my messages. (The way I first found out my configuration was working was a message from a cron job which had some unexpected output.)


Part of the installation is to set up a symlink so that sendmail becomes an alias for ssmtp. You can use either command.

The ssmtp command doesn’t seem to include command-line options for specifying the subject line or the name of the recipient..

So, here’s a command line you can use. Edit the email address as suits your needs. (The sender name and email address will be embedded by GMail.)

Ignore the word-wrap, this is all one line.

echo -e "Subject: Test Message\nTo: Your Name Here <you@example.com>\nThis message was sent via ssmtp." | ssmtp -t

Alternatively, you can put the recipient’s email address on the command line (the message will then be received as a BCC).

echo -e "Subject: Test Message\nThis message was sent via ssmtp." | ssmtp you@example.com


Four files are written to /var/logs

  • mail.err – contains an entry for each time there’s a problem sending a message.
  • mail.info – contains an entry for each attempt (successful or failed) at sending a message
  • mail.log – duplicates mail.info.
  • mail.warn – duplicates mail.err.

(Image via Pixabay user Deans_icons used under Pixabay License.)

Typing Emoji on WIndows 10 😲

This has the potential to be dangerous. 😲 I was writing an email to a friend and wanted it to be perfectly clear that what I was writing was joke. I find the purely textual emoticons such as the sideways smiley 🙂 are often mistaken for punctuation and their intent lost, so I wanted to use a graphical emoji. We both use GMail, but I don’t care for the “melted lump” characters Google put in there.

On a whim, I did a search for how to type emoji on windows 10 and found a PC World article explaining how to type emoji if you have the Fall Creators Update. My immediate thought was to wonder if it was safe to assume everyone had it yet, and then I realized the article was from 2017! So the feature’s been there for a while and I just didn’t know about it.

So, if you on Windows 10, you can type emoji by pressing the Windows Key, followed by either the period or the semi-colon, and an emoji keyboard will appear. This is much more convenient – and universal – than any per-website or per-application emoji button.

Don’t get me wrong, I’m not a huge fan of emoji, but sometimes you just really want to type 🚡 or ✈ without first visiting emojipedia. 😃

(Only bummer on this is GMail replaces the emoji characters with the “melted lump” equivalents.)

By the way, Mac users can do the same by pressing and holding down both the [Control] and [Command] keys and then hitting the space bar.

Image by Pixaline from Pixabay. Used under the Pixabay license.

Turn off Chrome notification pop-ups.

Talking to my Dad over the weekend, I learned he’d recently started seeing ads appear on his computer. These days, everyone’s used to seeing ads pop up in the web browser, but these were in the lower right corner of the screen, and would stick around if you moved the browser window.

My first thought was some sort of malware infection, but there didn’t seem to be any unusual processes running. The vital clue was when Dad mentioned that the ads only showed up when Chrome was running, and would disappear if you closed all the Chrome windows. An online search revealed a new suspect – Chrome’s browser notifications.

If you’ve used Chrome at all in the past year or two, you’ve very likely seen the browser display a message asking for permission from one site or another (or dozens) to display notifications. The idea behind notifications is that even if you aren’t actively browsing a particular site, you can still let it put up a message about an important update (e.g. event tickets going on sale). In this case, someone had decided to use notifications to display ads and Mom or Dad had accidentally given the site permission to display them.

Fortunately, once you know what’s going on, it’s pretty easy to turn notifications back off.

The steps below turn notifications off altogether; you should’t even be prompted anymore. I generally try to avoid distractions, but if you don’t want to turn them completely off, you can stop at step 6 to review and edit the list of sites where they’re allowed and/or blocked.

  1. In the upper right corner of the Chrome browser, click on the three vertical dots.
  2. Near the bottom of the menu, click settings.
  3. Scroll to the bottom of the page and click “Advanced.”
  4. Go to the “Privacy and Security” section and click “Site Settings”.
  5. Find the line for “Notifications” (there will be a small bell to the left).
  6. Click Notifications.
  7. At the top of the page, where it says “Ask before sending (recommended)”, click on the blue “slider” button.  It will turn gray.
  8. At the top, click on  the “Notifications” arrow.
  9. Click on the “Site Settings” arrow.


Dynamic Type Selection in PHP

So, check out this block of PHP code:

class MyClass {
  public function doSomething() {
    echo "Hello there.";
$class = "MyClass";
$method = "doSomething";
$instance = new $class();

The “new” statement is using a string variable to specify the class being instantiated! What’s more, on the very next line, another string variable is being used, this time to specify the method being invoked.

My background is C-like languages (C, C++, C# and Java), so I was somewhat surprised to discover that this not only executes, but does so without errors. Turns out, this is just how PHP’s new statement works.

In Java, you’d use Class<T> to get the class object, retrieve a specific constructor, and then call the newInstance method, passing any required parameters. So the above example would end up looking something like this:

package myprogram;

import java.lang.Class;
import java.lang.reflect.Method;

class MyClass {
    public MyClass() {}
    public void doSomething() {
        System.out.println("Hello there.");

    public static void main(String[] args) {
        String className = "myprogram.MyClass";
        Class<?> class = Class.forName(className);
        MyClass instance = class.getConstructor().newInstance();

        Method method = class.getMethod("doSomething");

In a more complex code sample (e.g. passing arbitrary class objects into a method to be instantiated and used in a callback), the Java version has definite advantages in terms of compile time type checking.

But I can’t deny that the simplicity of the PHP version also has some appeal.

Password rules

Some very basic rules for managing your passwords:

  1. Don’t even think about using “password” as your password. That’s the number one most used password in the world.
  2. Consider using a password manager. No one will ever guess that your password is qwb5Qauz36H9Kleqyotx and with a password manager, you won’t have to remember it.
  3. If you must use a password you can remember, at least use a passphrase. “SixSillySwansSangSonnets” is much more secure than “Tr0ubad0r” (and a darn sight easier to remember the correct spelling).
  4. Never, ever, ever use the same password on two different sites. In short: if one site has a breach and the bad guys get hold of your username and password, they’re going to try using them on other sites as well.
  5. Faithfully following those rules doesn’t guarantee that none of your accounts will ever get hacked, too much of that’s out of your hands. But they’re a solid start and they’ll definitely help limit the damage.

    A non-technical relative admits to not understanding why people would use a password manager. Couldn’t someone just hack your password manager?

    Yes. That could potentially happen. The aforementioned password rules also apply when setting the password for your password manager.

    And you have to ask yourself, which system is more secure? A well-vetted, “battle tested” password manager (and I’m referring to the likes of LastPass, 1Password, or KeePass), storing passwords which are composed of 20 random letters and numbers? Or just using the site’s name with a couple letters and maybe a number?

    And which is easier? Keeping track of a single strong password for the password manager? Or trying to remember what password you used for 30, 40, or more different web sites? (Hint: you’re gonna remember the Six Silly Swans example for a long time.) The main reason people re-use passwords is that they need to keep track of so doggone many of them!

    The idea behind a password manager is that you only have to remember one really good password, and then the password manager remembers the rest of them.

    And the good password managers (I personally use LastPass and KeePass) use heavy-duty encryption. If you use a good password, it’s extraordinarily unlikely that anyone’s going to break into your password manager by brute-force guessing.

    (Image via Life of Pix on Pexels.com under Creative Commons 1.0 Universal)

Two Git bookmarks

I follow Mark Hamill on Twitter because I find him entertaining. For example, this exchange:

On the other hand, I primarily follow Scott Hanselman because he drops interesting tech nuggets, such as when he retweeted this:

Don’t get me wrong, Scott can be entertaining too, but the git config linked from that tweet is full of things I didn’t even know you could configure! (Six months in, I’m still finding entirely new realms within git that I didn’t realize I didn’t know about!)

Seeing the config sections for the merge and diff tools, I also verified that yes, the bc merge/diff tool referenced in the git documentation really is (or at least, can be) Beyond Compare. I’ve been using TortiseGit for my Git GUI needs, but I also like Beyond Compare. (The default vi-based diff tool is just painful.)

And then I find an article about how to configure Beyond Compare to work with Git.

String Types

Not quite a year ago, I received a .Net Rocks! mug from Richard Campbell and Carl Franklin after a comment I’d left for a previous episode was read on the show. History repeated itself on Thursday when they used another of my comments, this time one about C++, as the lead-in for the show’s main topic.

Thursday’s show was about a scripting language, chaiscript, that allows you to write scripts in C++ and use them from other C++ projects. (C++ as a scripting language is a neat trick since it’s normally compiled ahead of time and shipped to the user as a binary executable.) It’s an interesting show and you should absolutely give it a listen. There’s also an interesting bit around the 20 mark, talking about the Commodore 64 (I had no idea those disk drives had CPUs).

The gist of my comment was that some of the features added to C++ since I’d last used it sounded rather compelling (particularly “stack semantics” which sound like there’s a sharply reduced need for new and delete, and that even pointers are largely hidden). I still have reservations though because of “scars from working with a half-dozen different, not-quite compatible string types.”

The first web application I ever worked on was a bit of a brownfield product, sharing code for the business logic with a desktop product that used the Microsoft Foundation Classes library (MFC). The resulting web application started off with char * along with the MFC CString class. (That’s two string types right there.)

Because this application ran on Active Server Pages (so-called “Classic ASP”), we soon added the BSTR and CComBSTR types in order to work with COM. And then, every so often, a new “sheriff” would attempt to unify things under a single “standard” class, which meant the introduction of TCHAR, wchar_t *, std::string and std::wstring. (Of course, as we all know, unifying under a new standard just makes things worse.)

So that’s really eight not-quite-compatible string types.

It was definitely a learning experience (if for no other reason than the anti-patterns), but I very much enjoy the fact that the C#, Java, and JavaScript languages only have one string type apiece.

Docker Tips

I’ve been working with Docker the past few months and all-in-all, I’ve been very pleased with the quality of the documentation. But, as with any other tool, there are always a few tricks to pick up, particularly when trying to script things out for an automated build or deployment. I’ve listed some of the more useful ones below and will update this post as I learn new ones.

Note: These are mainly oriented around running Docker in a Linux environment, as that’s where I’m currently using it.

How do I stop typing sudo all the time?

Docker runs as root, so when you’re working with an out-of-the-box installation, the docker command must be preceded by sudo. Since it may not be desirable for all Docker users to be able to execute commands as root, the installation creates a docker group. Members of the group may execute docker commands without elevated privileges.

To add users to the group, execute the command:
sudo usermod -aG docker <username>

How do I remove all stopped containers?

When a container is stopped, it remains loaded. You can remove it by issuing the command docker rm container_name, but that can be a hassle if you have a large number of containers loaded and they all have random names (a frequent occurrence when you’re first learning Docker).

You can remove all stopped containers by executing the command:
docker rm $(docker ps --quiet -a --filter status=exited)

(The –filter option prevents errors from attempting to remove containers which are currently running.)

You can also cause your containers to remove themselves automatically by including the –rm option on the docker run command line.

How do I know if a container is running?

To determine if a named container (e.g. “clever_leakey”) is currently running

containerID=$(docker ps --quiet --filter status=running --filter name=clever_leakey)

if $containerID is non-null, the named container is running. If it’s null, then the container is no longer running.

Do note however that there are other non-running states, e.g. paused, which will also return a null containerID for this test. As an alternative, to find only the containers which are stopped, use status=exited.

If the docker run command includes the –rm option, the container will be removed from memory.

(Image via openclipart under Creative Commans CC0 1.0 Universal)

Fixing Evernote’s “Could not add tray icon, error: An attempt was made to reference a token that does not exist.” message

I reinstalled Evernote a week or so back and every time I fired it up, a background window would also open containing the message “Could not add tray icon, error: An attempt was made to reference a token that does not exist.” Every time this happened, I’d dismiss the message and move on with what I was working on.

This routine got old pretty quickly so I did what any other geek would do and Googled for the message. Apparently the message has been around for a while, with the suggested fix being to reinstall Evernote. So I uninstalled Evernote, waited a few minutes, and then reinstalled it. Then I went back into the application and a background window opened with the same message.

This time, after closing both the pop-up and the main application window, I took a look in the system tray and discovered that Evernote’s “running in the background” icon was also missing. I also realized I’d never been prompted to run the installer as an administrator.

I run my computer differently than most people – the user account where I do my day-to-day work has reduced privileges. There’s a separate login for anything requiring elevated privileges, such as installing software. Most installers will either prompt you to either login as an administrator, or else they’ll install to an alternate location (generally somewhere in the %APPDATA% folder). I didn’t dig too deeply, but my best guess is that Evernote was doing the latter, but the system tray icon requires something to be installed with higher privileges.

In the end, I uninstalled Evernote again and this time made sure to re-install with admin privileges.

I haven’t seen the error message since.

(Public domain image, via pixabay)

Problem: chmod is ignored in the Git Bash prompt

So here’s a strange one that had me baffled for a bit – the chmod command is pretty much a null operation from the Git Bash prompt (MingW64). This initially showed up on a script for launching a Docker container, but as nearly as I can tell, it happens for any shell script.

So, we have a simple script that prints out “Hello World!”.

blair@Squawk MINGW64 ~/test
$ cat foo
echo Hello World!

Simple enough. Now the thing is, I want to make this script executable. Now this particular Bash implementation will let me run ./foo and it’ll execute, but my real use case (running a Docker container) is going to have a somewhat longer name. Just as a matter of convenience, I’d like to to type just the first few characters, press tab, and have the filename expanded. And besides, your executable files should always be marked as executable.

blair@Squawk MINGW64 ~/test
$ ls -l
total 2
-rwxr-xr-x 1 blair 197121 28 Oct 18 00:20 bar*
-rw-r--r-- 1 blair 197121 18 Oct 18 00:10 foo

blair@Squawk MINGW64 ~/test

OK, this is an easy fix, I just need to run chmod and set the execute bit to on, right?

blair@Squawk MINGW64 ~/test
$ ls -l
total 2
-rwxr-xr-x 1 blair 197121 28 Oct 18 00:20 bar*
-rw-r--r-- 1 blair 197121 18 Oct 18 00:10 foo

blair@Squawk MINGW64 ~/test
$ chmod 744 foo
blair@Squawk MINGW64 ~/test
$ ls -l
total 2
-rwxr-xr-x 1 blair 197121 28 Oct 18 00:20 bar*
-rw-r--r-- 1 blair 197121 18 Oct 18 00:10 foo

The execute bit didn’t change. Maybe I need to use the u+x syntax instead?

$ chmod u+x foo
blair@Squawk MINGW64 ~/test
$ ls -l
total 2
-rwxr-xr-x 1 blair 197121 28 Oct 18 00:20 bar*
-rw-r--r-- 1 blair 197121 18 Oct 18 00:10 foo

Still no luck. So why is bar marked as executable? What’s the difference between these two scripts? The answer turns out to be one line of code:

blair@Squawk MINGW64 ~/test
$ chmod u+x foo
blair@Squawk MINGW64 ~/test
$ cat bar
echo Hello World!

Do you see that first line, where it says “#!/bin/sh”. That’s how Bash knows what interpreter to pass the script to. It also turns out, in this particular implementation, that’s how Bash knows the file contains an executable script instead of just text.

So we modify foo, and get this result:

blair@Squawk MINGW64 ~/test$ cat foo
echo Hello World!
blair@Squawk MINGW64 ~/test
$ ls -l
total 2
-rwxr-xr-x 1 blair 197121 28 Oct 18 00:20 bar*
-rwxr--r-- 1 blair 197121 18 Oct 18 00:10 foo*

(Image credit: Screenshot by ThatBlairGuy)