Recently, I had to customize NSIS installers for work.

It was my first time, and I had to get to know this tool first. While this system already offers many functionalities, if you want to change anything in its built-in pages, you must rely on Win32 APIs.

The welcome and finish pages are an exception, up to a certain sense, because they are made with nsDialogs, NSIS’s plugin for creating completely custom pages.

As a result, it is possible to pass to add new widgets using the same API in a function passed MUI_PAGE_CUSTOMFUNCTION_SHOW. However, I encountered a major problem: the default widgets fill the available space. Therefore, any new widgets will be hidden, and it will not be possible to interact with them with the mouse. I looked for a solution online - NSIS is more than 20 years old, so you can find a lot of documentation - but I did not find an answer that worked for me.

Eventually, I found that you can resize widgets on Win32 with MoveWindow. … [Leggi il resto]

A while ago, I started working on a dataset I captured a few years ago with a Microsoft Kinect One.

I immediately realized the data looked much cleaner than the newer datasets I created with my Intel RealSense D435.

I had already noticed that, above a certain distance, the depth data was full of craters. I already knew the error is proportional to the squared distance, but for me, it was much bigger than expected. Therefore, I calibrated the sensors and now I stay closer to my targets during the acquisitions.

But for the last dataset I captured, I tried another strategy: I decided to save also the raw IR footage to process it offline.

Stereo vision

RealSense cameras are RGBD sensors: they provide simultaneously a color (RGB) and depth (D) stream.

There are several types of techniques to measure depth. For example, the original Kinect for the Xbox 360 uses “structured light”, and the Kinect One included a time-of-flight camera.

The RealSense D400 series is based on stereo vision, which works by matching the same point in frames captured by two different cameras. There is a relation between the displacement of this point (disparity), the relative position of the two cameras, and the depth. … [Leggi il resto]

In the last few months, in my free time, I have been developing a small application to elaborate RGBD datasets capturing people. In particular, my goal is to create 3D scans of heads. I do not expect these models to be well-made or even usable without some processing, but I wonder if I can transform this data at least to starter models.

I first started to work with RGBD cameras during my internship at Altair. At the time, I built a small pipeline to reconstruct models based on Kinect Fusion.

Sadly, Kinect Fusion is an online method. The advantage is that it will give immediate feedback if it loses the camera tracking. But the disadvantage is that it needs a pretty powerful GPU. I have one on my desktop, but I took all my datasets using laptops.

Also, in my experience, getting a usable dataset with Kinect Fusion requires several attempts and time (the acquisition must proceed very slowly), which, generally speaking, is not always compatible with… people 😄️. They might move, or lose patience, etc etc… … [Leggi il resto]

Two weeks ago, I needed to mass-download a few IMAP mailboxes before migrating them to another provider.

Gist contains many scripts to do so, and I tried one of them, but it did not work as I wanted. I wanted to download all the IMAP folders, not only Inbox.

Therefore, I wrote my version of that script 😄️.

I used imapclient instead of the built-in imaplib because it is more Pythonic and handles all the tedious conversions between bytes, strings, and other types.

The center of the script is the process function. It connects to an IMAP server, queries the list of folders, and downloads all the messages from each one but trash and spam (but notice that the comparison is case-sensitive!).

client.fetch downloads all the messages you provide to it, which caused an OOM in my case. Therefore, I split the list with the message IDs into chunks. I had to write the batches function because I was running on Python 3.11. From Python 3.12, you can use itertools.batched instead. … [Leggi il resto]

Once upon a time, the Tor Browser Bundle was an actual bundle. It included Firefox, the Tor daemon, and Torbutton, the extension to turn on and off the Tor mode in the browser.

This toggle model was not great and extremely confusing to some users. This and other problems led to the creation of Tor Browser: this article contains more details about this story.

From a technical point of view, Torbutton did not really go away. The visible button disappeared, but much of the related code remained.

Part of the state isolation code was not necessary anymore because Tor Browser always runs in private browsing mode or was dropped over the years thanks to Firefox improvements and the Tor Uplift initiative. However, the circuit display, the first-party domain circuit isolation, and other parts of the existing code were still needed. As a result, Torbutton continued to live for many years as a Tor Browser-only built-in extension on its separate repository and included in the browser with git submodules (even though the browser was non-functional without it). New patches and functionalities were written in the Firefox code that constitutes Tor Browser, and the Torbutton code was changed only to fix existing bugs or to keep it working in new versions of Firefox. … [Leggi il resto]