Debugging is not a problem per se. As an example you can see how Nim generates debugging info while transpiling to C with the --debuginfo --linedir:on options. The real problem with C transpilation is proper garbage collection support. A C compiler doesn't track roots and that means that for a language transpiled to C it is not practically feasible as of today to support garbage collection algorithms that are not conservative. There are published tecniques that should allow implementing full non conservative GC on top of C but they have remained in research settings so far.As a matter of fact the de facto standard GC for such langauages is Bohem's which has known limitations.
> The V compiler hasn't crashed once for me, and I haven't received any crash reports from other users.<p><a href="https://github.com/vlang/v/issues/2838" rel="nofollow">https://github.com/vlang/v/issues/2838</a> "Compiler crash during filtering's assignment operation on a mutable array"<p><a href="https://github.com/vlang/v/issues/2565" rel="nofollow">https://github.com/vlang/v/issues/2565</a> "Compiler crash when using inexistent struct type in const block"<p>(One of these is from two days after the author wrote the above statement, but the other isn't.) There are other reports where the V compiler seems to generate C code that does not compile, or only with very nasty type warnings, for example: <a href="https://github.com/vlang/v/issues/2960" rel="nofollow">https://github.com/vlang/v/issues/2960</a> <a href="https://github.com/vlang/v/issues/2981" rel="nofollow">https://github.com/vlang/v/issues/2981</a>
This looks like an improved version of Go. The website claims that it will guarantee no data races at compile time (work in progress). I'd like to know how he plans to accomplish this as it's not a trivial problem to solve completely. Also, a lack of algebraic (sum) data types makes it much less attractive to me. Otherwise, neat little language.