On Tuesday, 13 January 2015 at 00:22:33 UTC, Mike wrote:
On Sunday, 11 January 2015 at 20:17:25 UTC, Iain Buclaw via
Digitalmars-d-announce wrote:
On 10 January 2015 at 20:15, Walter Bright via
Digitalmars-d-announce
<digitalmars-d-announce@puremagic.com> wrote:
On 1/10/2015 9:50 AM, Andrei Alexandrescu wrote:
On 1/10/15 9:49 AM, Andrei Alexandrescu wrote:
On 1/10/15 8:15 AM, Iain Buclaw via Digitalmars-d-announce
wrote:
In any event, are you doing flash talks this year? I
don't think I
could find something to spend more than 15 minutes talking
about this
year.
Yes. -- Andrei
I should add that gdc is a topic of much interest so pretty
much anything
you
say would be interesting. I compel you to prepare a full
talk. -- Andrei
I agree. There's no way you don't have interesting things to
talk about! For
example, what is your process for integrating dmd changes
into gdc? What are
the advantages/disadvantages of gdc? What are the biggest
challenges you
face working on gdc? What's the hardest problem you solved
with gdc? How can
others help out? Etc.
Talking about that probably extends a possible talk to 30
minutes,
covering two subjects. :o)
I have a suggestion for any compiler implementers: How about a
talk on how to get started hacking the compiler. Something
that may lower the entry barrier and encourage participation.
Some random thoughts:
* General structure of the compiler
* Walk through the data flow: Lexer -> parser -> AST -> backend
* How to add a new compiler switch (e.g. -fnotypeinfo)
* How to add a new attribute (e.g. @notypeinfo)
* What's your workflow for debugging the compiler?
* Pick a bug, and fix it (Live demo)
* Overview of CTFE and how it's implemented
* (I'm sure you can think of more)
I realize there's documentation on the wiki, and some of this
was discussed briefly at DConf2013, but there's more that can
be done to make it accessible and interesting.
Mike
I would really like this. It would be difficult to know what
level to target it at though.
