Iconv

iconv is an essential utility commonly found on Unix-like systems. It stands for "international conversion," and its primary function is to convert text from one character encoding to another. This capability is particularly useful when dealing with files created in different regions or systems, where encoding formats may vary (for example, converting between UTF-8 and ISO-8859-1). By enabling seamless conversion of file encodings via the command line or within scripts, iconv helps ensure compatibility and proper display of text data across various platforms and applications.

On modern versions of macOS, iconv is bundled as a standard part of the operating system—thoroughly tested and maintained by Apple. For Iguana, this means it can confidently rely on the system-supplied implementation of iconv on macOS, which is better integrated and more likely to receive timely updates and security fixes. As a result, it is preferable for Iguana to use the macOS-provided version of iconv rather than including its own copy, which would only add unnecessary bloat and maintenance overhead.

The situation on Linux, however, requires more careful consideration. The availability and versions of iconv can differ widely between distributions, so shipping a bundled version could help ensure consistent behavior across all environments.

I first noticed a problem with iconv when the initial build of Iguana 6.2.1 was crashing on my system. The crash was traced to file logic that invoked iconv to handle issues arising from an old Apple file system called HFS.

To some, cross-platform development might seem like an unnecessary complication, but it actually encourages a heightened awareness of simplicity. This focus on simplification is what has made Iguana so stable—and why our customers have trusted it for so many years. But it’s important to remember: simplicity doesn't come easily.

I hope to eventually remove the iconv library from Iguana 6, but first, I’ll need to carefully determine if any customer code still depends on it. Perhaps we can detect calls to iconv and ask customers to report their usage to us?

Today, most systems have standardized on UTF-8, so perhaps soon this friction will no longer be needed.

However, it turns out that MD Anderson has one interface where users, when copying and pasting from an older application that uses Unicode, need iconv—they use iconv calls to filter out certain Unicode characters. So I will at least need to look at how they use it.

Funnily enough, the National University Hospital in Singapore does not have this requirement. It’s interesting what drives these needs—ultimately, it’s the applications and their workflows.