[talk] 2015-10-07 NYCBUG- meeting TBA?
Isaac (.ike) Levy
ike at blackskyresearch.net
Thu Oct 8 12:22:35 EDT 2015
> On Oct 7, 2015, at 11:40 PM, Raúl Cuza <raulcuza at gmail.com> wrote:
>
> Tonight was a lot of fun. Thank you everyone who took part. If nothing else it is a reminder that you can learn a lot of what to expect from your operating system by looking at the code, irrespective of whether you know what `#ifndef` means or not.
Absolutely.
I'm really excited that George Brocklehurst will be continuing the Meetup in NYC!
>
> It was very interesting to see how build practices imposed a lot of bloat into the two simple programs we looked at.
Inspired by Wietse Venema's comments, I'll offer this up:
https://github.com/dotike/honest
Ha! Everyone, submit patches!!!
Best,
.ike
>
> I offer the Darwin version, lifted from NetBSD, as a balanced way to write `true` in c:
>
> http://www.opensource.apple.com/source/shell_cmds/shell_cmds-118/true/
I dig how they have a true.sh in there too, perhaps is that there for build tools to use, before the binary is compiled?
>
> The code itself is simple and thus easy to audit, but the folder it resides in still retains the standard structure the project as a whole needs to simplify the build, test, and documentation-creation processes. It is written in C, so it doesn't have the shell overhead a nameless OS choose to do.
>
> Speaking of which, anyone feel like running the numbers to compare the differences in speed of the various versions of `true`? If you have a super-cluster handy, you could distribute the job to 10,000 cores to get a good statistical measurement.
>
> Raúl
>
I'd like to see that too... The truth is out there... (haha)
Rocket-
.ike
More information about the talk
mailing list