Re:Go’s runtime C to Go rewrite, by the numbers
by sanderjd on Monday September 01, 2014 @10:04AM
(#8252844)
To expand on "real binaries" a bit. It's somewhat ironic, but distributing statically compiled binaries for common systems is actually easier than distributing source (or bytecode) for VM-based languages like Python, Ruby, and Java.
Re:Go’s runtime C to Go rewrite, by the numbers
by pjc50 on Monday September 01, 2014 @11:07AM
(#8253184)
True statically compiled Linux binaries for complex applications are surprisingly hard; libnss insists on dynamically loading, and libstdc++ doesn't like it either. I think we ended up with unpleasant RPATH hacks to load the copy of the standard libraries that shipped alongside the app rather than the system ones.
Re:Go’s runtime C to Go rewrite, by the numbers
by sanderjd on Monday September 01, 2014 @06:52PM
(#8255004)
Interesting! Thanks for the info. I've never run into trouble myself, but I've never tried to link a go application to libnss or libstdc++.
Re:Go’s runtime C to Go rewrite, by the numbers
by pjmlp on Monday September 01, 2014 @05:58PM
(#8254845)
- there are lots of JVMs to choose from.
- also possible in Java via Unsafe package, which is being promoted to official package in Java 9
- Excelsior JET, Atego JVM, RoboVM and many others AOT compilers do exist for Java.
- C-style systems programming without C-style catastrophic security bugs
- "real binaries"