On Friday, 20 September 2013 at 07:16:09 UTC, Manu wrote:
[...]
It needs a reasonable amount of support from the compiler and
presumably cooperation from the debugger too. If people have
never heard of it, chances are, it doesn't exist :(
Google has been hitting close to this:
http://www.youtube.com/watch?v=U98rhV6wONo (slides:
http://llvm.org/devmtg/2012-04-12/Slides/Manuel_Klimek.pdf)
Lot of potential here, even for a die-hard Vim user like myself.
Mmmm, a concept that I've always found completely amazing
actually. How is it that Linux - truly an OS for developers
(certainly not for end-users) - can consistently be plagued by
the worst dev tools out there? Surely someone in the past 30-40
years get's frustrated at some stage, looks at what MS have
been doing for over a decade, and think "shit, that's awesome,
I'd like that too!". I'm actually amazed that MS managed to
invent it in the first place. You'd think that Linux should have
gotten to it first...
Part of this would seem to be the simple fact that the GNU
toolchain is the de facto standard for working in Linux and other
Unixen. If you've ever encountered a GNU project's general
attitude toward patches and ideas from "outsiders", the concerns
presented in this SO answer might offer some clarity as to why
it's taken so long:
http://stackoverflow.com/a/4440794/432364
In short: getting all the people involved to agree on answers to
all these questions is sort of a hard problem.
(Though it looks like "Fix-and-continue" was added to the GDB
roadmap about a year ago:
http://gcc.gnu.org/wiki/cauldron2012?action=AttachFile&do=get&target=jkratoch.pdf)
I agree, I'm really looking forward to what they bring to the
table. I expect it's a lot of work though... they have over a
decade of catching up to do.
Per the links above, they might be closer than it initially
appears. Of course, it's also a matter of integration and
coordination across multiple projects. There are, occasionally,
advantages to monolithic vertically-integrated dictatorships.
-Wyatt
