Modern Error Handling with C++23

2025-04-10T00:00:00+00:00

With the introduction of std::expected</code> in C++23 and the use of std::error_code</code> form C++11, it is finally possible to use error codes in C++ without the usual hassle.

Using the Modern Error Handling</h2>
Let's first see an example of how this could look like. A function that might return an error could be defined like this:</p>

[</span>[</span>nodiscard]</span></span>]</span></span> std::</span>expected<</span>int</span>,</span> std::</span>error_code></span>
</span>divide</span></span></span>(</span></span></span>int</span> numerator</span>,</span> int</span> denominator</span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    if</span> (</span>denominator ==</span> 0</span>)</span></span> {</span>
</span></span></span></span>        return</span> std::</span>unexpected</span>{</span></span></span>math_error::</span>division_by_zero</span></span>}</span></span></span>;</span>
</span></span></span></span>    }</span></span>
</span></span></span>
</span></span></span>    return</span> numerator /</span> denominator;</span>
</span></span></span></span></span>}</span></span></span>
</span></code></pre>
The nodiscard</code> attribute (introduced with C++17) is not required, but I strongly recommend using it, so that the compiler will issue a warning, if the result of a call to the function is not used.</p>
The function can be called then just as shown here:</p>
//</span> no need for output parameters, the potential error and the
</span></span>//</span> result can both just be retrieved as a return parameter
</span></span>const</span> auto</span> result =</span> divide</span>(</span></span></span>100</span>,</span> 10</span></span></span>)</span></span></span>;</span>
</span>
</span>//</span> check for a valid result to do error handling (expected)
</span></span>if</span> (</span>!</span>result)</span></span> {</span>
</span></span>    //</span> in case of an error, we can directly print the static
</span></span></span>    //</span> message, that is connected to the error code, for example
</span></span></span>    printf</span>(</span></span></span>"</span>Failed to calculate: %s</span>"</span></span>,</span> result.</span>error</span>(</span></span></span></span>)</span></span></span>.</span>message</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span>}</span></span> else</span> {</span>
</span></span>    //</span> access the calculated value after checking
</span></span></span>    const</span> auto</span> value =</span> result.</span>value</span>(</span></span></span></span>)</span></span></span>;</span>
</span></span>}</span></span>
</span></code></pre>
But what if we do not actually can handle the error, but want to forward it to our caller? Just return the error in another std::unexpected</code>:</p>
const</span> auto</span> result =</span> divide</span>(</span></span></span>100</span>,</span> 10</span></span></span>)</span></span></span>;</span>
</span>if</span> (</span>!</span>result)</span></span> return</span> std::</span>unexpected</span>{</span></span></span>result.</span>error</span>(</span></span></span></span>)</span></span></span></span></span>}</span></span></span>;</span>
</span>
</span>printf</span>(</span></span></span>"</span>The result was: %i</span>"</span></span>,</span> result.</span>value</span>(</span></span></span></span>)</span></span></span></span>)</span></span></span>;</span>
</span></code></pre>
Compared to the indeed very nice error handling in Rust, we are missing a bit of syntactical sugar, but the usage pattern is quite close:</p>
let</span> result =</span> divide</span>(</span>100</span>,</span> 10</span></span>)</span></span>?</span>;</span>
</span>print!</span>(</span></span>"</span>The result was: {}</span>"</span></span></span>,</span> result)</span></span>;</span>
</span></code></pre>
Implementing the Modern Error Handling</h2>
Let's see then, how the modern error handling can actually be implemented in C++.
Using std::expected</code> is straightforward and could directly be used with your existing error codes if you enabled the C++23 standard on your compiler (see also at cppreference</a>):</p>
#include</span> <</span>cstdio></span></span>
</span></span>#include</span> <</span>expected></span></span>
</span></span>
</span>//</span> does not need to be an enum class, could also be plain integers
</span></span>enum class</span> </span>math_error</span></span>
</span></span>{</span></span></span>
</span></span></span>    //</span> the 0 value should always be reserved for 'success'
</span></span></span></span>    division_by_zero =</span> 1</span>,</span>
</span></span></span></span></span>}</span></span></span>;</span>
</span>
</span>[</span>[</span>nodiscard]</span></span>]</span></span> std::</span>expected<</span>int</span>,</span> math_error></span>
</span>divide</span></span></span>(</span></span></span>int</span> numerator</span>,</span> int</span> denominator</span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    if</span> (</span>denominator ==</span> 0</span>)</span></span> {</span>
</span></span></span></span>        return</span> std::</span>unexpected</span>{</span></span></span>math_error::</span>division_by_zero</span></span>}</span></span></span>;</span>
</span></span></span></span>    }</span></span>
</span></span></span>
</span></span></span>    return</span> numerator /</span> denominator;</span>
</span></span></span></span></span>}</span></span></span>
</span>
</span>int</span> main</span></span></span>(</span></span></span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    const</span> auto</span> result =</span> divide</span>(</span></span></span>100</span>,</span> 10</span></span></span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    if</span> (</span>!</span>result)</span></span> {</span>
</span></span></span></span>        printf</span>(</span></span></span>"</span>Failed to calculate: %s</span>"</span></span>,</span> result.</span>error</span>(</span></span></span></span>)</span></span></span>.</span>c_str</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span></span></span>        return</span> -</span>1</span>;</span>
</span></span></span></span>    }</span></span>
</span></span></span>
</span></span></span>    printf</span>(</span></span></span>"</span>The result was: %i</span>"</span></span>,</span> result.</span>value</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span></span></span></span>}</span></span></span>
</span></code></pre>
Combining it with std::error_code</code> instead of plain error codes, however, has some advantages like:</p>

multiple enums can be used to define error codes in the application and the linker ensures with the help of std::error_category</code> that every error code is unique</li>
different error code types from different parts of the system can be forwarded by a function, without even needing to know which concrete types are defined</li>
it can be casted to bool, hence there is no need to compare to some "no error" value</li>
it contains a human-readable string that can be used for logging, for example</li>
</ul>
How would the same example look with std::error_code</code> then? Let's first define our error codes enum and our custom error category:</p>
//</span> The enum will all error codes of this part of the software (for one
</span></span>//</span> specific error category). As many as needed can be defined for
</span></span>//</span> different subsystems or libraries.
</span></span>enum class</span> </span>math_error</span></span>
</span></span>{</span></span></span>
</span></span></span>    //</span> the 0 value must be reserved for the 'success' case
</span></span></span></span>    division_by_zero =</span> 1</span>,</span>
</span></span></span></span></span>}</span></span></span>;</span>
</span>
</span>//</span> The concrete error category does not need to be part of any interface.
</span></span>//</span> Hence, you would usually define it in an anonymous namespace in an
</span></span>//</span> implementation file.
</span></span>namespace</span>
</span></span></span></span></span>{</span>
</span></span></span>    //</span> The error category needs to implement the name() and message()
</span></span></span></span>    //</span> methods to allow for getting human readable information about any
</span></span></span></span>    //</span> given error code later on.
</span></span></span></span>    struct</span> </span>error_category</span></span> :</span> public</span> std::</span>error_category</span> {</span></span></span>
</span></span></span></span></span>        [</span>[</span>nodiscard]</span></span>]</span></span> const</span> char</span> *</span>name</span>(</span></span></span></span>)</span></span></span> const</span> noexcept</span> override
</span></span></span></span></span>        {</span>
</span></span></span></span></span></span>            return</span> "</span>math_error"</span></span>;</span>
</span></span></span></span></span></span>        }</span></span>
</span></span></span></span></span>
</span></span></span></span></span>        [</span>[</span>nodiscard]</span></span>]</span></span> std::</span>string message</span>(</span></span></span>int</span> ev</span></span>)</span></span></span> const</span> override
</span></span></span></span></span>        {</span>
</span></span></span></span></span></span>            switch</span> (</span>static_cast</span><</span>math_error></span>(</span>ev)</span></span>)</span></span> {</span>
</span></span></span></span></span></span></span>            case</span> math_error::</span>division_by_zero:</span>
</span></span></span></span></span></span></span>                return</span> "</span>Division by zero"</span></span>;</span>
</span></span></span></span></span></span></span>            }</span></span>
</span></span></span></span></span></span>
</span></span></span></span></span></span>            return</span> "</span>Unknown error"</span></span>;</span>
</span></span></span></span></span></span>        }</span></span>
</span></span></span></span></span>    </span></span>}</span></span></span>;</span>
</span></span></span>
</span></span></span>    //</span> This is how the error codes are ensured to be unique. There must
</span></span></span></span>    //</span> be exactly one instance of each error category defined in your
</span></span></span></span>    //</span> application somewhere in any implementation file. Two error codes
</span></span></span></span>    //</span> are assumed to be in the same category, if the addresses of their
</span></span></span></span>    //</span> categories are identical.
</span></span></span></span>    const</span> error_category the_error_category</span>{</span></span></span></span>}</span></span></span>;</span>
</span></span></span>
</span></span></span></span>}</span></span></span> //</span> namespace
</span></span>
</span>//</span> This function needs to be implemented to allow for automatic
</span></span>//</span> conversions of the enum to instances of std::error code. So that you
</span></span>//</span> only need do 'return math_error:division_by_zero', for example, in a
</span></span>//</span> function that returns a std::error_code.
</span></span>std::</span>error_code make_error_code</span></span></span>(</span></span></span>math_error code</span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    return</span> {</span>static_cast</span><</span>int</span>></span>(</span>code)</span></span>,</span> the_error_category}</span></span>;</span>
</span></span></span></span></span>}</span></span></span>
</span>
</span>//</span> The enum must be defined to an error code enum within the std namespace.
</span></span>//</span> This will lead to the templated conversion functions from an enum value
</span></span>//</span> to an error code of the std::error_code class to be instantiated for
</span></span>//</span> this enum class. They will internally call the above defined function
</span></span>//</span> make_error_code() via argument dependent lookup then.
</span></span>namespace</span> </span>std</span></span>
</span></span>{</span>
</span></span></span>    template</span> <</span></span>></span></span>
</span></span></span>    struct</span> </span>is_error_code_enum</span></span><</span>math_error></span> :</span> public</span> std::</span>true_type</span> {</span></span></span>
</span></span></span></span></span>    </span></span>}</span></span></span>;</span>
</span></span></span></span>}</span></span></span> //</span> namespace std
</span></span></code></pre>
There is a bit of boiler-plate code needed, for defining an error category and code. However, this needs to be done only once for every category and you usually won't need many categories. Could be only one for the application code, for example, and one for a 3rd-party dependency. After defining, it can then be used like that:</p>
#include</span> <</span>cstdio></span></span>
</span></span>#include</span> <</span>expected></span></span>
</span></span>#include</span> <</span>system_error></span></span>
</span></span>
</span>//</span> include the error category and code definition from above
</span></span>//</span> or a corresponding header file
</span></span>
</span>[</span>[</span>nodiscard]</span></span>]</span></span> std::</span>expected<</span>int</span>,</span> std::</span>error_code></span>
</span>divide</span></span></span>(</span></span></span>int</span> numerator</span>,</span> int</span> denominator</span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    if</span> (</span>denominator ==</span> 0</span>)</span></span> {</span>
</span></span></span></span>        return</span> std::</span>unexpected</span>{</span></span></span>math_error::</span>division_by_zero</span></span>}</span></span></span>;</span>
</span></span></span></span>    }</span></span>
</span></span></span>
</span></span></span>    return</span> numerator /</span> denominator;</span>
</span></span></span></span></span>}</span></span></span>
</span>
</span>int</span> main</span></span></span>(</span></span></span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    const</span> auto</span> result =</span> divide</span>(</span></span></span>100</span>,</span> 10</span></span></span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    if</span> (</span>!</span>result)</span></span> {</span>
</span></span></span></span>        printf</span>(</span></span></span>"</span>Failed to calculate: %s</span>"</span></span>,</span> result.</span>error</span>(</span></span></span></span>)</span></span></span>.</span>message</span>(</span></span></span></span>)</span></span></span>.</span>c_str</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span></span></span>        return</span> -</span>1</span>;</span>
</span></span></span></span>    }</span></span>
</span></span></span>
</span></span></span>    printf</span>(</span></span></span>"</span>The result was: %i</span>"</span></span>,</span> result.</span>value</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span></span></span></span>}</span></span></span>
</span></code></pre>
Dynamic Memory?</h2>
The major disadvantage of std::error_code</code> that would prevent me from using it on embedded devices is the use of std::string</code> in the interface of std::error_category</code> (and also in the implementations of std::error_code</code> and std::error_condition</code>):</p>
virtual</span> std::</span>string message</span></span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span> =</span> 0</span></span>;</span>
</span></code></pre>
Implementing an error category, would lead to the introduction of dynamic memory via std::string</code>, which is usually unwanted in the firmware development. Luckily, the only things that are needed for implementing system_error</code> are to provide a std::error_category</code> class a std::error_code</code> class and optionally std::error_condition</code> class. The latter could be left out, if the feature or error conditions are not needed.</p>
All of those classes are quite small and simple, as you could check out in the STL implementation from the LLVM project</a>, for example. Let's see what we need to implement for a system_error</code> without dynamic memory.</p>

    The source code of a full implementation can be found in my Zephyr example repository</a>.</span>
</span>
</p>
We need a slightly different interface for the error_category</code> class compared to the version from the STL</abbr> to avoid dynamic memory. Its interface would look like this then:</p>
class</span> </span>error_category</span></span>
</span></span>{</span></span></span>
</span></span></span>public</span>:</span>
</span></span></span>    //</span> class uses virtual inheritance → destructor needs to be virtual
</span></span></span></span>    virtual</span> ~error_category</span></span>(</span></span></span>)</span></span></span> =</span> default</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> class must not be copied
</span></span></span></span>    error_category</span></span>(</span></span></span>)</span></span></span> =</span> default</span></span>;</span>
</span></span></span>    error_category</span></span>(</span></span></span>const</span> error_category &</span>)</span></span></span> =</span> delete</span></span>;</span>
</span></span></span>    error_category &</span>operator=</span></span>(</span></span></span>const</span> error_category &</span>)</span></span></span> =</span> delete</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> a name must be defined by subclasses
</span></span></span></span>    virtual</span> const</span> char</span> *</span>name</span></span>(</span></span></span>)</span></span></span> const</span> =</span> 0</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> define the method for the default error condition, only
</span></span></span></span>    //</span> if you plan to implement and use error conditions
</span></span></span></span>    virtual</span> error_condition default_error_condition</span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> equivalence between error codes and conditions of this category (if error conditions are used)
</span></span></span></span>    virtual</span> bool</span> equivalent</span></span>(</span></span></span>int</span> code</span>,</span> const</span> error_condition &</span>condition</span>)</span></span></span> const</span></span>;</span>
</span></span></span>    virtual</span> bool</span> equivalent</span></span>(</span></span></span>const</span> error_code &</span>code</span>,</span> int</span> condition</span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> here's the difference to the STL implementation → char* instead of std::string
</span></span></span></span>    virtual</span> const</span> char</span> *</span>message</span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span> =</span> 0</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> those operators for equivalence and ordering must be defined
</span></span></span></span>    bool</span> operator==</span></span>(</span></span></span>const</span> error_category &</span>rhs</span>)</span></span></span> const</span></span>;</span>
</span></span></span>    bool</span> operator!=</span></span>(</span></span></span>const</span> error_category &</span>rhs</span>)</span></span></span> const</span></span>;</span>
</span></span></span>    bool</span> operator<</span></span>(</span></span></span>const</span> error_category &</span>rhs</span>)</span></span></span> const</span></span>;</span>
</span></span></span></span></span>}</span></span></span>;</span>
</span></code></pre>
Similarly for the error_code</code> class, where we also need to replace the usage of std::string</code>:</p>
class</span> </span>error_code</span></span>
</span></span>{</span></span></span>
</span></span></span>public</span>:</span>
</span></span></span>    //</span> default constructor must set success value on system_category
</span></span></span></span>    error_code</span></span>(</span></span></span>)</span></span></span>;</span>
</span></span></span>    error_code</span></span>(</span></span></span>int</span> value</span>,</span> const</span> error_category &</span>category</span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    template</span> <</span></span>class</span> Enum,</span> typename</span> =</span> typename</span> std::</span>enable_if_t<</span>is_error_code_enum_v<</span>Enum></span>></span></span>></span></span>
</span></span></span>    error_code</span></span>(</span></span></span>Enum enum_value</span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    void</span> assign</span></span>(</span></span></span>int</span> value</span>,</span> const</span> error_category &</span>category</span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    template</span> <</span></span>class</span> Enum,</span> typename</span> =</span> typename</span> std::</span>enable_if_t<</span>is_error_code_enum_v<</span>Enum></span>></span></span>></span></span>
</span></span></span>    error_code &</span>operator=</span></span>(</span></span></span>Enum enum_value</span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    //</span> back to success value on system_category (like default constructed)
</span></span></span></span>    void</span> clear</span></span>(</span></span></span>)</span></span></span>;</span>
</span></span></span>
</span></span></span>    //</span> just getters for the private attributes
</span></span></span></span>    int</span> value</span></span>(</span></span></span>)</span></span></span> const</span></span>;</span>
</span></span></span>    const</span> error_category &</span>category</span></span>(</span></span></span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> just to forward call to category().default_error_condition()
</span></span></span></span>    error_condition default_error_condition</span></span>(</span></span></span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> as for the error_category, the return value needs to be char* instead of std::string
</span></span></span></span>    const</span> char</span> *</span>message</span></span>(</span></span></span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>    //</span> operator needs to be implemented to return true on any value != 0
</span></span></span></span>    explicit</span> operator</span> bool</span></span>(</span></span></span>)</span></span></span> const</span></span>;</span>
</span></span></span>
</span></span></span>private</span>:</span>
</span></span></span>    int</span> value_;</span>
</span></span></span>    const</span> error_category *</span>category_;</span>
</span></span></span></span></span>}</span></span></span>;</span>
</span></code></pre>
As you might have spotted in the interface definitions already, there are two instances of std::error_category</code> required to be provided by an implementation: a system category and a general category. For compatibility and for being able to default-construct error_code</code> instances, it makes sense to implement those instances as well:</p>
//</span> implementation of the generic error category
</span></span>class</span> </span>generic_error_category</span></span> :</span> public</span> error_category</span>
</span></span>{</span></span></span>
</span></span></span>public</span>:</span>
</span></span></span>    virtual</span> const</span> char</span> *</span>name</span></span>(</span></span></span>)</span></span></span> const</span>
</span></span></span></span>    </span>{</span></span></span>
</span></span></span></span></span>        return</span> "</span>generic"</span></span>;</span>
</span></span></span></span></span>    </span></span>}</span></span></span>
</span></span></span>    virtual</span> const</span> char</span> *</span>message</span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span>
</span></span></span></span>    </span>{</span></span></span>
</span></span></span></span></span>        return</span> strerror</span>(</span></span></span>ev</span></span>)</span></span></span>;</span>
</span></span></span></span></span>    </span></span>}</span></span></span>
</span></span></span></span></span>}</span></span></span>;</span>
</span>
</span>//</span> one instance must be created in an implementation file
</span></span>const</span> generic_error_category the_generic_error_category</span>{</span></span></span></span>}</span></span></span>;</span>
</span>
</span>//</span> this instance should be returned by a function
</span></span>const</span> error_category &</span>generic_category</span></span></span>(</span></span></span>)</span></span></span>
</span></span></span>{</span></span></span>
</span></span></span>    return</span> the_generic_error_category;</span>
</span></span></span></span></span>}</span></span></span>
</span>
</span>
</span>//</span> implementation of the system error category
</span></span>class</span> </span>system_error_category</span></span> :</span> public</span> error_category</span>
</span></span>{</span></span></span>
</span></span></span>public</span>:</span>
</span></span></span>    virtual</span> const</span> char</span> *</span>name</span></span>(</span></span></span>)</span></span></span> const</span>
</span></span></span></span>    </span>{</span></span></span>
</span></span></span></span></span>        return</span> "</span>system"</span></span>;</span>
</span></span></span></span></span>    </span></span>}</span></span></span>
</span></span></span>    virtual</span> const</span> char</span> *</span>message</span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span>
</span></span></span></span>    </span>{</span></span></span>
</span></span></span></span></span>        return</span> strerror</span>(</span></span></span>ev</span></span>)</span></span></span>;</span>
</span></span></span></span></span>    </span></span>}</span></span></span>
</span></span></span>    virtual</span> error_condition default_error_condition</span></span>(</span></span></span>int</span> ev</span>)</span></span></span> const</span>
</span></span></span></span>    </span>{</span></span></span>
</span></span></span></span></span>        return</span> error_condition</span>(</span></span></span>ev,</span> generic_category</span>(</span></span></span></span>)</span></span></span></span></span>)</span></span></span>;</span>
</span></span></span></span></span>    </span></span>}</span></span></span>
</span></span></span></span></span>}</span></span></span>;</span>
</span>
</span>//</span> one instance must be created in an implementation file
</span></span>const</span> system_error_category the_system_error_category</span>{</span></span></span></span>}</span></span></span>;</span>
</span>
</span>//</span> this instance should be returned by a function
</span></span>const</span> error_category &</span>system_category</span></span></span>(</span></span></span>)</span></span></span> noexcept</span>
</span></span></span>{</span></span></span>
</span></span></span>    return</span> the_system_error_category;</span>
</span></span></span></span></span>}</span></span></span>
</span></code></pre>
After doing the necessary implementations for the error_category</code>, error_code</code> and optionally the error_condition</code>, the error handling can be used without dynamic memory.</p>

    Instead of doing your own implementation, feel free to check out what I already implemented in my Zephyr example repository</a>.</span>
</span>
</p>

Comments for Static Website (with Mastodon)

2025-02-28T00:00:00+00:00

When I created this website and blog as a static website with Zola</a>, I wondered how I could add a comment feature.

A quick search revealed the post Adding comments to your static blog with Mastodon</a> from Carl Schwan, in which he explains how he shows responds to his Mastodon posts as comments to his blog. This was already very close to what I wanted to achieve, however, I would prefer a solution that works without the need to dynamically load the comments via JavaScript. Hence, based on his solution I created a completely static integration of Mastodon responses as comments into my website.

Prepare Static Website for Comments</h2>

The content of the static website is provided in Markdown files, which means that there is a Markdown file for each blog post. Comments could be directly added there as well, however, for a cleaner separation of blog content and comments, I decided to use a separate Markdown file for the comments for each blog post. For this post, which is located at content/blog/static-comments-for-static-website.md</code>, the comments would be added to the file content/blog-comments/static-comments-for-static-website.md</code>.

As long as the comments file has the same filename as the Markdown file of the blog post, the comments can be included for the blog post, by adding the following to the blog page template:

{#-</span> Show comments and link to Mastodon post, if the comments variables are set for the page. -#}</span></span>
</span>{%-</span> if</span> page</span>.</span>extra</span>.</span>comments</span>.</span>host</span> and</span> page</span>.</span>extra</span>.</span>comments</span>.</span>username</span> and</span> page</span>.</span>extra</span>.</span>comments</span>.</span>id</span> -%}</span></span>
</span>
</span>  {#-</span> Generate comment filename from current filename (same name different path) and load its content. -#}</span></span>
</span>  {%-</span> set</span> comments_file</span> =</span> page</span>.</span>relative_path</span> |</span> split</span>(</span>pat</span>=</span>"</span>/"</span></span>)</span> |</span> last</span> -%}</span></span>
</span>  {%-</span> set</span> comments</span> =</span> get_page</span>(</span>path</span>=</span>"</span>blog-comments/"</span></span> ~ comments_file</span>)</span> -%}</span></span>
</span>
</span>  {#-</span> Show comments, if there are any in the comments file. -#}</span></span>
</span>  {%-</span> if</span> comments</span>.</span>content</span> -%}</span></span>
</span>    <h2>Comments</h2>
</span>    {{</span> comments</span>.</span>content</span> |</span> safe</span> }}</span></span>
</span>  {%-</span> endif</span> -%}</span></span>
</span>
</span>  {#-</span> Show link to the corresponding Mastodon post. -#}</span></span>
</span>  <a href="https://{{</span> page</span>.</span>extra</span>.</span>comments</span>.</span>host</span> }}</span></span>/@{{</span> page</span>.</span>extra</span>.</span>comments</span>.</span>username</span> }}</span></span>/{{</span> page</span>.</span>extra</span>.</span>comments</span>.</span>id</span> }}</span></span>" target="_blank">Comment on Mastodon</a>
</span>
</span>{%-</span> endif</span> -%}</span></span>
</span></code></pre>
For the comments itself, I added a shortcode</a>, so that the presentation is separated from the content:</p>
<</span>div</span> class</span>=</span>"</span></span>comment</span> flex</span> mt-2</span> mb-4</span> {%</span> if</span> level</span> ></span> 1</span> %}ml-8{%endif</span> %}</span>"</span></span></span>></span></span>
</span>    <</span>a</span> href</span>=</span>"</span>{{ user_url }}"</span></span></span> target</span>=</span>"</span>_blank"</span></span></span> rel</span>=</span>"</span>external nofollow"</span></span></span>></span></span><</span>img</span> src</span>=</span>"</span>{{ avatar }}"</span></span></span> class</span>=</span>"</span></span>h-12</span> w-12</span> object-cover</span> rounded-full</span> mr-2</span>"</span></span></span>></span></span></</span>a</span>></span></span>
</span>    <</span>div</span>></span></span>
</span>        <</span>div</span>></span></span>
</span>            <</span>a</span> href</span>=</span>"</span>{{ user_url }}"</span></span></span> target</span>=</span>"</span>_blank"</span></span></span> rel</span>=</span>"</span>external nofollow"</span></span></span> class</span>=</span>"</span></span>mr-12</span>"</span></span></span>></span></span>{{ fullname }}</</span>a</span>></span></span>
</span>        </</span>div</span>></span></span>
</span>        <</span>a</span> href</span>=</span>"</span>{{ comment_url }}"</span></span></span> target</span>=</span>"</span>_blank"</span></span></span> rel</span>=</span>"</span>external nofollow"</span></span></span> class</span>=</span>"</span></span>text-sm</span>"</span></span></span>></span></span>{{ datetime }}</</span>a</span>></span></span>
</span>        <</span>div</span> class</span>=</span>"</span></span>mt-2</span>"</span></span></span>></span></span>
</span>            {{ body | safe }}
</span>        </</span>div</span>></span></span>
</span>        {% if favorites > 0 %}
</span>        <</span>a</span> href</span>=</span>"</span>{{ comment_url }}/favourites"</span></span></span> target</span>=</span>"</span>_blank"</span></span></span>></span></span><</span>div</span> class</span>=</span>"</span></span>text-sm</span>"</span></span></span>></span></span>{{ content::icon(id="heart") }}{{ favorites }}</</span>div</span>></span></span>
</span>        {% endif %}
</span>    </</span>div</span>></span></span>
</span></</span>div</span>></span></span>
</span></code></pre>
The shortcode can be instantiated like this in the comments Markdown file:</p>
{%</span> comment</span>(</span>user_url</span>=</span>"</span>https://mastodon.social/@awinterstein"</span></span>,</span>
</span></span>    comment_url</span>=</span>"</span>https://mastodon.social/@awinterstein/114081720861078028"</span></span>,</span>
</span></span>    fullname</span>=</span>"</span>Adrian Winterstein"</span></span>,</span>
</span></span>    datetime</span>=</span>"</span>February 28, 2025 at 13:21"</span></span>,</span>
</span></span>    avatar</span>=</span>"</span>https://files.mastodon.social/accounts/avatars/113/945/630/736/175/025/original/08b18b1267a1fe54.jpg"</span></span>,</span>
</span></span>    favorites</span>=</span>1,</span> level</span>=</span>1)</span> %}</span></span>
</span>This comment is just a test to show the Mastodon responses feature in my blog.
</span>{%</span> end</span> %}</span></span>
</span></code></pre>
Which would lead to the following rendering:</p>

    </a>
    
        
            Adrian Winterstein</a>
        </div>
        February 28, 2025 at 13:21</a>
        
            This comment is just a test to show the Mastodon responses feature in my blog.</p>
        </div>
        
        
With the templates and comment files being prepared, it was then possible to create a script to retrieve the comments from Mastodon and fill them into the comments files.</p>
Retrieving Comments from Mastodon</h2>
As described by Carl Schwan, there is a simple API available on Mastodon, to retrieve all replies to a given post. I created a small Python script that iterates through all blog entries, retrieves the Mastodon post IDs from the entries and checks the Mastodon API for the responses to this post.</p>
#</span>!/usr/bin/env python3
</span></span>
</span>from</span></span> glob</span></span> import</span></span></span></span></span> glob</span></span>
</span>from</span></span> urllib</span>.</span>request</span></span> import</span></span></span></span></span> urlopen</span></span>
</span>import</span> argparse</span></span></span>
</span>import</span> json</span></span></span>
</span>import</span> re</span></span></span>
</span>import</span> tomllib</span></span></span>
</span>import</span> os</span>.</span>path</span></span> as</span> path</span></span></span>
</span>from</span></span> dateutil</span></span> import</span></span></span></span></span> parser</span> as</span> datetime_parser</span></span>
</span>from</span></span> babel</span>.</span>dates</span></span> import</span></span></span></span></span> format_date</span>,</span> format_time</span></span>
</span>
</span>
</span>#</span> the TOML header of the Markdown files (as defined for Zola)
</span></span>HEADER</span></span> =</span> "</span></span></span>+++\n</span>render=false\n</span>+++\n</span>\n</span>"</span></span></span>
</span>
</span>#</span> path where the markdown files of the blog entries are stored
</span></span>BLOG_PATH</span></span> =</span> '</span></span></span>content/blog'</span></span></span>
</span>
</span>#</span> path where the markdown files for the blog comments should be written
</span></span>BLOG_COMMENTS_PATH</span></span> =</span> '</span></span></span>content/blog-comments'</span></span></span>
</span>
</span>#</span> the default language of the website as defined in the config.toml
</span></span>DEFAULT_LANGUAGE</span></span> =</span> "</span></span></span>en"</span></span></span>
</span>
</span>#</span> additional languages of the website as defined in the config.toml
</span></span>ADDITIONAL_LANGUAGES</span></span> =</span> [</span>"</span></span></span>de"</span></span></span>]</span></span>
</span>
</span>
</span>def</span></span> get_blog_entry_files</span></span></span>(</span></span>path</span>,</span> exclude_translations</span>)</span></span>:</span></span>
</span>    """</span>
</span></span>    Returns a list of all markdown blog entry filenames from the given directory.
</span></span>
</span></span>    The index files are always excluded from this list and it can optionally be
</span></span>    requested with the exclude_translations parameter, to only return the files
</span></span>    of the default language, but exclude all translation files.
</span></span>    """</span></span>
</span>
</span>    #</span> pattern for markdown file names in additional languages that contain a language code (e.g, blog-entry.de.md)
</span></span>    translation_pattern</span></span> =</span> re</span></span>.</span></span>compile</span></span></span></span>(</span>"</span></span></span>.*\\</span>.[a-z]{</span>2}</span></span>\\</span>.md$"</span></span></span>)</span></span>
</span>
</span>    files</span></span> =</span> [</span>]</span></span>
</span>
</span>    for</span> file</span> in</span></span> glob</span></span></span></span>(</span>f</span>'</span></span></span>{</span>path</span></span></span></span>}</span></span>/*.md'</span></span></span>)</span></span></span>:</span></span>
</span>        file</span></span> =</span> file</span></span>.</span></span>replace</span></span></span></span>(</span>f</span>'</span></span></span>{</span>path</span></span></span></span>}</span></span>/'</span></span></span>,</span> '</span></span></span>'</span></span></span>)</span></span>
</span>
</span>        if</span> not</span> file</span></span>.</span></span>startswith</span></span></span></span>(</span>'</span></span></span>_index'</span></span></span>)</span></span> and</span> (</span>not</span> exclude_translations</span></span> or</span> not</span> translation_pattern</span></span>.</span></span>match</span></span></span></span>(</span>file</span></span>)</span></span>)</span></span>:</span></span>
</span>            files</span></span>.</span></span>append</span></span></span></span>(</span>file</span></span>)</span></span>
</span>
</span>    return</span> files</span></span>
</span>
</span>
</span>def</span></span> get_blog_entry_comments_config</span></span></span>(</span></span>path</span>,</span> exclude_translations</span>)</span></span>:</span></span>
</span>    """</span>
</span></span>    Returns a dictionary with the Mastodon comments configuration from each
</span></span>    blog entry file from the given path.
</span></span>
</span></span>    It can optionally be requested with the exclude_translations parameter, to
</span></span>    only return the configuration from the files of the default language.
</span></span>    """</span></span>
</span>    files_config</span></span> =</span> {</span>}</span></span>
</span>
</span>    for</span> file</span> in</span></span> get_blog_entry_files</span></span></span></span>(</span>BLOG_PATH</span></span>,</span> exclude_translations</span></span>)</span></span></span>:</span></span>
</span>        with</span> open</span></span></span></span>(</span>f</span>'</span></span></span>{</span>BLOG_PATH</span></span></span></span>}</span></span>/</span>{</span>file</span></span></span></span>}</span></span>'</span></span></span>)</span></span> as</span></span></span> f</span>:</span></span>
</span>            config_part</span></span> =</span> f</span></span>.</span></span>read</span></span></span></span>(</span>)</span></span>.</span></span>split</span></span></span></span>(</span>'</span></span></span>+++\n</span>'</span></span></span>)</span></span>[</span></span>1</span></span>]</span></span>
</span>            data</span></span> =</span> tomllib</span></span>.</span></span>loads</span></span></span></span>(</span>config_part</span></span>)</span></span>
</span>
</span>            if</span> '</span></span></span>extra'</span></span></span> in</span> data</span></span> and</span> '</span></span></span>comments'</span></span></span> in</span> data</span></span></span>[</span></span>'</span></span></span>extra'</span></span></span></span>]</span></span>:</span></span>
</span>                files_config</span></span></span>[</span></span>file</span></span></span>]</span></span> =</span> data</span></span>
</span>
</span>    return</span> files_config</span></span>
</span>
</span>
</span>def</span></span> create_empty_comment_files</span></span></span>(</span></span>BLOG_COMMENTS_PATH</span>,</span> files_comments_config</span>)</span></span>:</span></span>
</span>    """</span>
</span></span>    Creates the markdown files for the comments of each blog entry.
</span></span>
</span></span>    There are no comments in the files, only the header is written. Existing
</span></span>    files are not overwritten.
</span></span>    """</span></span>
</span>    for</span> file</span> in</span></span> files_comments_config</span></span></span>:</span></span>
</span>        comments_file</span></span> =</span> f</span>'</span></span></span>{</span>BLOG_COMMENTS_PATH</span></span></span></span>}</span></span>/</span>{</span>file</span></span></span></span>}</span></span>'</span></span></span>
</span>        if</span> not</span> path</span></span>.</span></span>isfile</span></span></span></span>(</span>comments_file</span></span>)</span></span>:</span></span>
</span>            with</span> open</span></span></span></span>(</span>comments_file</span></span>,</span> "</span></span></span>x"</span></span></span>)</span></span> as</span></span></span> f</span>:</span></span>
</span>                f</span></span>.</span></span>write</span></span></span></span>(</span>HEADER</span></span>)</span></span>
</span>
</span>
</span>def</span></span> get_display_name_without_emojis</span></span></span>(</span></span>comment</span>)</span></span>:</span></span>
</span>    """</span>
</span></span>    Returns the display name from the given comment dict, but with all emojis
</span></span>    removed from the name.
</span></span>    """</span></span>
</span>    name</span></span> =</span> comment</span></span></span>[</span></span>'</span></span></span>account'</span></span></span></span>]</span></span>[</span></span>'</span></span></span>display_name'</span></span></span></span>]</span></span>
</span>    for</span> emoji</span> in</span></span> comment</span></span></span>[</span></span>'</span></span></span>account'</span></span></span></span>]</span></span>[</span></span>'</span></span></span>emojis'</span></span></span></span>]</span></span></span>:</span></span>
</span>        name</span></span> =</span> name</span></span>.</span></span>replace</span></span></span></span>(</span>f</span>"</span></span></span>:</span>{</span>emoji</span></span></span>[</span></span>'</span></span></span>shortcode'</span></span></span></span>]</span></span></span></span>}</span></span>:"</span></span></span>,</span> "</span></span></span>"</span></span></span>)</span></span>
</span>    return</span> name</span></span>
</span>
</span>
</span>def</span></span> create_markdown_entry</span></span></span>(</span></span>comment</span>,</span> locale</span>)</span></span>:</span></span>
</span>    """</span>
</span></span>    Returns the string for the macro instance of the comment.
</span></span>
</span></span>    The macro is defined in templates/shortcodes/comment.html. For now, only
</span></span>    the date and time are translated according to the given locale.
</span></span>    """</span></span>
</span>    connecting_word</span></span> =</span> {</span></span>'</span></span></span>en'</span></span></span></span>:</span></span> </span>'</span></span></span>at'</span></span></span></span>,</span> </span>'</span></span></span>de'</span></span></span></span>:</span></span> </span>'</span></span></span>um'</span></span></span></span>}</span></span>
</span>
</span>    level</span></span> =</span> 1</span> if</span> comment</span></span></span>[</span></span>'</span></span></span>in_reply_to_id'</span></span></span></span>]</span></span> ==</span> str</span></span></span></span>(</span>config</span></span></span>[</span></span>'</span></span></span>id'</span></span></span></span>]</span></span>)</span></span> else</span> 2</span></span>
</span>    datetime</span></span> =</span> datetime_parser</span></span>.</span></span>parse</span></span></span></span>(</span>comment</span></span></span>[</span></span>'</span></span></span>created_at'</span></span></span></span>]</span></span>)</span></span>
</span>    date</span></span> =</span> format_date</span></span></span></span>(</span>datetime</span></span>,</span> format</span>=</span>'</span></span></span>long'</span></span></span>,</span> locale</span>=</span>locale</span></span>)</span></span>
</span>    time</span></span> =</span> format_time</span></span></span></span>(</span>datetime</span></span>,</span> '</span></span></span>H:mm'</span></span></span>,</span> locale</span>=</span>locale</span></span>)</span></span>
</span>    return</span> f</span>'</span></span></span>{{</span>% comment(user_url="</span>{</span>comment</span></span></span>[</span></span>'</span></span></span>account'</span></span></span></span>]</span></span>[</span></span>'</span></span></span>url'</span></span></span></span>]</span></span></span></span>}</span></span>", comment_url="</span>{</span>comment</span></span></span>[</span></span>'</span></span></span>url'</span></span></span></span>]</span></span></span></span>}</span></span>", fullname="</span>{</span>get_display_name_without_emojis</span></span></span></span>(</span>comment</span></span>)</span></span></span></span>}</span></span>", '</span></span></span> \</span>
</span>        f</span>'</span></span></span>datetime="</span>{</span>date</span></span></span></span>}</span></span> </span>{</span>connecting_word</span></span></span>[</span></span>locale</span></span></span>]</span></span></span></span>}</span></span> </span>{</span>time</span></span></span></span>}</span></span>", avatar="</span>{</span>comment</span></span></span>[</span></span>'</span></span></span>account'</span></span></span></span>]</span></span>[</span></span>'</span></span></span>avatar_static'</span></span></span></span>]</span></span></span></span>}</span></span>", favorites=</span>{</span>int</span></span></span></span>(</span>comment</span></span></span>[</span></span>'</span></span></span>favourites_count'</span></span></span></span>]</span></span>)</span></span></span></span>}</span></span>, level=</span>{</span>level</span></span></span></span>}</span></span>) %}}</span>'</span></span></span> \</span>
</span>        f</span>'</span></span></span>{</span>comment</span></span></span>[</span></span>'</span></span></span>content'</span></span></span></span>]</span></span></span></span>}</span></span>{{</span>% end %}}</span>'</span></span></span>
</span>
</span>
</span>if</span> __name__</span></span> ==</span> "</span></span></span>__main__"</span></span></span>:</span></span>
</span>    #</span> argument handling to allow for the parameter for creating the missing files
</span></span>    parser</span></span> =</span> argparse</span></span>.</span></span>ArgumentParser</span></span></span></span>(</span>
</span></span>        prog</span>=</span>'</span></span></span>Retrieve Comments'</span></span></span>,</span>
</span></span>        description</span>=</span>'</span></span></span>Retrieves comments from Mastodon and creates corresponding markdown files'</span></span></span>)</span></span>
</span>    parser</span></span>.</span></span>add_argument</span></span></span></span>(</span>'</span></span></span>--create-empty'</span></span></span>,</span> action</span>=</span>'</span></span></span>store_true'</span></span></span>,</span>
</span></span>                        help</span>=</span>'</span></span></span>Only create the missing empty comment files'</span></span></span>)</span></span>
</span>    args</span></span> =</span> parser</span></span>.</span></span>parse_args</span></span></span></span>(</span>)</span></span>
</span>
</span>    #</span> create the markdown files for the comments of each blog entry (without comments)
</span></span>    if</span> args</span>.</span>create_empty</span></span>:</span></span>
</span>        files_comments_config</span></span> =</span> get_blog_entry_comments_config</span></span></span></span>(</span>
</span></span>            BLOG_PATH</span></span>,</span> exclude_translations</span>=</span>False</span>)</span></span>
</span>        create_empty_comment_files</span></span></span></span>(</span>
</span></span>            BLOG_COMMENTS_PATH</span></span>,</span> files_comments_config</span></span>)</span></span>
</span>
</span>    #</span> retrieve the comments from Mastodon and write them to the markdown files
</span></span>    else</span>:</span></span>
</span>        files_comments_config</span></span> =</span> get_blog_entry_comments_config</span></span></span></span>(</span>
</span></span>            BLOG_PATH</span></span>,</span> exclude_translations</span>=</span>True</span>)</span></span>
</span>
</span>        for</span> file</span>,</span> data</span> in</span></span> files_comments_config</span></span>.</span></span>items</span></span></span></span>(</span>)</span></span></span>:</span></span>
</span>            config</span></span> =</span> data</span></span></span>[</span></span>'</span></span></span>extra'</span></span></span></span>]</span></span>[</span></span>'</span></span></span>comments'</span></span></span></span>]</span></span>
</span>
</span>            #</span> retrieve JSON from the Mastodon API
</span></span>            url</span></span> =</span> f</span>'</span></span></span>https://</span>{</span>config</span></span></span>[</span></span>'</span></span></span>host'</span></span></span></span>]</span></span></span></span>}</span></span>/api/v1/statuses/</span>{</span>config</span></span></span>[</span></span>'</span></span></span>id'</span></span></span></span>]</span></span></span></span>}</span></span>/context'</span></span></span>
</span>            response</span></span> =</span> json</span></span>.</span></span>loads</span></span></span></span>(</span>urlopen</span></span></span></span>(</span>url</span></span>)</span></span>.</span></span>read</span></span></span></span>(</span>)</span></span>)</span></span>
</span>
</span>            #</span> write the default language comment files
</span></span>            with</span> open</span></span></span></span>(</span>f</span>'</span></span></span>{</span>BLOG_COMMENTS_PATH</span></span></span></span>}</span></span>/</span>{</span>file</span></span></span></span>}</span></span>'</span></span></span>,</span> "</span></span></span>w"</span></span></span>)</span></span> as</span></span></span> f</span>:</span></span>
</span>                f</span></span>.</span></span>write</span></span></span></span>(</span>HEADER</span></span>)</span></span>
</span>                for</span> comment</span> in</span></span> (</span>response</span></span></span>[</span></span>'</span></span></span>descendants'</span></span></span></span>]</span></span>)</span></span></span>:</span></span>
</span>                    f</span></span>.</span></span>write</span></span></span></span>(</span>create_markdown_entry</span></span></span></span>(</span>comment</span></span>,</span> locale</span>=</span>DEFAULT_LANGUAGE</span></span>)</span></span>)</span></span>
</span>
</span>            #</span> write the additional language comment files
</span></span>            for</span> language</span> in</span></span> ADDITIONAL_LANGUAGES</span></span></span>:</span></span>
</span>                file_with_language</span></span> =</span> file</span></span>.</span></span>replace</span></span></span></span>(</span>'</span></span></span>.md'</span></span></span>,</span> f</span>'</span></span></span>.</span>{</span>language</span></span></span></span>}</span></span>.md'</span></span></span>)</span></span>
</span>                with</span> open</span></span></span></span>(</span>f</span>'</span></span></span>{</span>BLOG_COMMENTS_PATH</span></span></span></span>}</span></span>/</span>{</span>file_with_language</span></span></span></span>}</span></span>'</span></span></span>,</span> "</span></span></span>w"</span></span></span>)</span></span> as</span></span></span> f</span>:</span></span>
</span>                    f</span></span>.</span></span>write</span></span></span></span>(</span>HEADER</span></span>)</span></span>
</span>                    for</span> comment</span> in</span></span> (</span>response</span></span></span>[</span></span>'</span></span></span>descendants'</span></span></span></span>]</span></span>)</span></span></span>:</span></span>
</span>                        f</span></span>.</span></span>write</span></span></span></span>(</span>create_markdown_entry</span></span></span></span>(</span>comment</span></span>,</span> locale</span>=</span>language</span></span>)</span></span>)</span></span>
</span></code></pre>
In addition to retrieving the responses, this script also provides the functionality to create the empty comment files for each blog entry. This is needed, because the used static site generator Zola does not support checking file existence before including a file. Hence, the comment files that are included as described before, all need to exist when building the site.</p>
Integration with CI Pipelines</h2>
I didn't want to run the script manually, hence I added a Github workflow to regularly retrieve the responses from Mastodon and notify me by creating a pull-request as soon as there are any new, changed or deleted comments. The workflow calls the retrieve-comments.py</code> script and, if there are any changes in the files afterwards, it creates a new branch and a pull-request with those changes.</p>
name</span></span>:</span> Add Comments</span>
</span>
</span>on</span>:</span>
</span>  schedule</span></span>:</span>
</span>  -</span> cron</span></span>:</span> '</span>30 20 * * *'</span></span>
</span>  workflow_dispatch</span></span>:</span>
</span>
</span>permissions</span></span>:</span>
</span>  contents</span></span>:</span> write</span>
</span>  pull-requests</span></span>:</span> write</span>
</span>
</span>jobs</span></span>:</span>
</span>  comments</span></span>:</span>
</span>    runs-on</span></span>:</span> ubuntu-latest</span>
</span>    steps</span></span>:</span>
</span>      -</span> name</span></span>:</span> Checkout</span>
</span>        uses</span></span>:</span> actions/checkout@v4</span>
</span>        with</span></span>:</span>
</span>          submodules</span></span>:</span> recursive</span>
</span>      -</span> name</span></span>:</span> Retrieve Comments</span>
</span>        run</span></span>:</span> ./retrieve-comments.py</span>
</span>      -</span> name</span></span>:</span> Create Pull-Request</span>
</span>        run</span></span>:</span> |</span>
</span>          if [ ! -z "$(git status --untracked-files=no --porcelain)" ]; then
</span></span>            git config --local user.email "adrian@winterstein.biz"
</span></span>            git config --local user.name  "Adrian Winterstein"
</span></span>
</span></span>            # checkout a branch, commit the comment changes and push the branch
</span></span>            git checkout -b "feature/comments-adaption"
</span></span>            git commit -a -m "Changed blog comments according to changes on Mastodon."
</span></span>            git push origin $(git branch --show-current) --force-with-lease -f
</span></span>
</span></span>            # create pull-request, if it dows not exist yet
</span></span>            if [ -z $(gh pr list --search "head:feature/comments-adaption" ) ]; then
</span></span>              gh pr create -B main -l comments -a awinterstein \
</span></span>                -t "Changed blog comments according to changes on Mastodon" \
</span></span>                -b "This is an automatically created pull-requests due to changes of the blog comments on Mastodon."
</span></span>            fi
</span></span>          fi
</span></span></span>        env</span></span>:</span>
</span>          GH_TOKEN</span></span>:</span> ${{ secrets.ACTIONS_ACCESS_TOKEN }}</span>
</span></code></pre>
Running the workflow once a day should keep the load on the Mastodon server low, while still being frequent enough for a small blog, where I do not expect a lot of comments.</p>
Creating a Blog Post</h2>
After publishing a blog post, a corresponding post on Mastodon needs to be created then. So that readers have a place to respond on. When the Mastodon post was created, the following variables can be added to the blog post, to link the blog to the Mastodon post:</p>
[</span>extra</span></span>]</span>
</span>comments</span>.</span>host</span></span> =</span> "</span>mastodon.social"</span></span> #</span> my Mastodon instance</span>
</span>comments</span>.</span>username</span></span> =</span> "</span>awinterstein"</span></span> #</span> my Mastodon username</span>
</span>comments</span>.</span>id</span></span> =</span> 114081713837118160</span> #</span> the id of my Mastodon post</span>
</span></code></pre>
The retrieve-comments.py</code> script uses those and will retrieve all replies of the Mastodon post from then on. Hence, the next pipeline execution after someone responded, will create a pull-request to integrate the comment into the static blog. If someone updates or deletes their response, this will be updated automatically as well.</p>

Next Steps</h2>
The local update could be extended to automatically install updates from an USB stick, as soon as it gets connected. An alternative would be the downloading of update images from an HTTP server. I'm showing this in the follow-up post Setting up Remote SWUpdate with Yocto</a>.</p>

Dev Ops</h3>
Designing, implementing and maintaining the environment from building the software, via analysis and testing to release creation and deployment.</span> View Projects

Adrian Winterstein

Architecture for Liquid Handling Platform

Optical Beam Smoke Detector

Modern Error Handling with C++23

Setting up Remote SWUpdate with Yocto

Comments for Static Website (with Mastodon)

Setting up SWUpdate with Yocto

Creating a Yocto Project

My Focus

Sensor Connection Module for a Patient Monitor

Digital Beacon for Traffic Diversions

Next Generation of Coffee Machines

Development of the Test Stand Software for a Motor Control Unit

IoT Edge Device for Campervans

Security Platform

IoT Device for the Acquisition of Sensor Data

Household Appliance for Hot Drinks Preparation

Architecture Review of a Control Software for Industrial 3D Printers

Responsibilities</h2>
Implementation of the architecture review, analysis of the results and development of concrete suggestions for improvement.</p>

Software Platform for Laboratory Analysis Devices

Prototype “Pump & Heater” for a Coffee Machine

Test Stand for an Oven

Control System for Innovative Chillers

Proof of Concept for a System for Tracking Transported Goods

Diagnostic Tool for High-end Converters for Railroad Applications

Navigated Implantation of Radiation Therapy Markers

Detection of Ears on 3D Profile Images

Adrian Winterstein

Architecture for Liquid Handling Platform

Optical Beam Smoke Detector

Modern Error Handling with C++23

Setting up Remote SWUpdate with Yocto

Comments for Static Website (with Mastodon)

Setting up SWUpdate with Yocto

Configure SWUpdate</h3> In case, that you want to change more configuration values of the SWUpdate build, you can open the menuconfig as follows:</p>

Bootloader Command</h3> When using meta-raspberrypi</code> on a Raspberry Pi, make sure to have the following set in the layer.conf</code>, so that the bootloader gets actually enabled in the image:</p>

Next Steps</h2> The local update could be extended to automatically install updates from an USB stick, as soon as it gets connected. An alternative would be the downloading of update images from an HTTP server. I'm showing this in the follow-up post Setting up Remote SWUpdate with Yocto</a>.</p>

Creating a Yocto Project

Git Submodules</h3> Alternatively, a git repository could be created for the Yocto project and the layers - including Poky - could be added as git submodules in this repository. With the structure then looking like this:</p>

My Focus

Software Development (Embedded)</h3> Developing (embedded) software in C++ with or without real-time operating systems like FreeRTOS - or on Linux. Use of modern C++ standards and best practices.</span> View Projects</svg></a> </p> </div> </div> </svg> </div>

Software Architecture</h3> Definition and documentation of software architectures. Preferably in interdisciplinary teams together with e.g., electronics and mechanics architects.</span> View Projects</svg></a> </p> </div> </div> </svg> </div>

Technical Lead</h3> Technical leading of (distributed) international software development teams. Clarification of requirements and agile implementation with the development team.</span> View Projects</svg></a> </p> </div> </div> </svg> </div>

Sensor Connection Module for a Patient Monitor

Digital Beacon for Traffic Diversions

Next Generation of Coffee Machines

Development of the Test Stand Software for a Motor Control Unit

IoT Edge Device for Campervans

Security Platform

IoT Device for the Acquisition of Sensor Data

Household Appliance for Hot Drinks Preparation

Architecture Review of a Control Software for Industrial 3D Printers

Responsibilities</h2> Implementation of the architecture review, analysis of the results and development of concrete suggestions for improvement.</p>

Software Platform for Laboratory Analysis Devices

Prototype “Pump & Heater” for a Coffee Machine

Test Stand for an Oven

Control System for Innovative Chillers

Proof of Concept for a System for Tracking Transported Goods

Diagnostic Tool for High-end Converters for Railroad Applications

Navigated Implantation of Radiation Therapy Markers

Detection of Ears on 3D Profile Images

Next Steps</h2>
The local update could be extended to automatically install updates from an USB stick, as soon as it gets connected. An alternative would be the downloading of update images from an HTTP server. I'm showing this in the follow-up post Setting up Remote SWUpdate with Yocto</a>.</p>

Responsibilities</h2>
Implementation of the architecture review, analysis of the results and development of concrete suggestions for improvement.</p>