A failed build now says what actually broke

Almost every failed build is a mistake in the code, and those have always read well: the compiler's own diagnostics, each one pinned to its line.

The rare exception was the confusing case. When a build failed for an infrastructure reason — something going wrong before the compiler ever ran — the build status showed only that fault's bare message. The most common one reads Object reference not set to an instance of an object., which tells you nothing at all, and in particular does not tell you the one thing that matters: that the problem was not in your code and there is nothing in your code to fix.

A build that fails this way now names the fault. Diagnosing one no longer depends on reproducing it — the details are recorded when it happens, which for a fault that appears only occasionally, and only under load, is the difference between a fixable report and a guess.

Nothing changes for an ordinary failed build. Compiler diagnostics are still what you see, still pinned to their lines, and the build status still shows them first.