Re: Spawn support in pkgsrc tools - Google Summer of Code 2016

[Andreas Theodosiou <andreas%atheod.net@localhost>]

On Sun, Mar 13, 2016 at 03:29:33AM +0300, Aleksej Saushev wrote:
>   Hello,
>
> Andreas Theodosiou <andreas%atheod.net@localhost> writes:
>
> > I am a second year Engineering student, planning to specialise in
> > Electrical and Information Sciences, at the University of Cambridge. I
> > am writing to let the community know of my desire to contribute to the
> > project via Google Summer of Code 2016. I would like to take on the
> > "Spawn support in pkgsrc tools" project as listed on the ideas page at
> > "https://wiki.netbsd.org/projects/project/pkgsrc_spawn_support/";.
> >
> > The aim of this project will be to implement posix_spawn support in
> > pkgsrc tools such as:
> > * devel/bmake
> > * shells/pdksh
> > * NetBSD base make
> > * NetBSD sh
> > * NetBSD base ksh
> >
> > Upon implementation of posix_spawn in the aforementioned tools, the
> > performance of posix_spawn in these tools will be evaluated and compared
> > to the traditional way of launching processes (fork and then clone with
> > an exec-family function).
>
> ...
>
> > I would really like to hear some feedback on this project so that I can
> > work further on developping a full project proposal. I would greatly
> > appreciate an opportunity to contribute to the community.
>
> To reiterate and to add to what we discussed already.
>
> Ideally I'd like to see project document reminiscent of some classical one:
> Russian "Tekhnicheskoe zadanie" GOST 25123-82 or analogous
> ("product/technical requirements document" or similar).
> Given that you study engineering, you should understand what I mean.
>
Yes I understand what you mean, I am familiar with the format of a
technical design proposal document.

> Please, note that I don't require it in all its bureaucratic glory.
> What I'd like to see is the most essential parts: project's purpose and scope,
> current state, wanted state, list of deliverables, acceptance criteria,
> testing procedures, high-level plan with milestones.
>
As I have also asked Martin could you elaborate on the preferred
method of testing procedures in NetBSD? Is there a preferred way of
evaluating performance of a program that NetBSD developers use?

> Ideally, I'd like to see calendar plan too since the time frame is quite tight.
> Probably, you want to devise your plan so that fallback is possible in the case
> of time shortage. Again, I don't want it in all its bureaucratic glory.
>

I aggree having something to fall back to is reasonable.

> Note that some of these may require you to spend some time on reading the source code
> and running some experiments, if you're not familiar.
>
>
> --
> HE CE3OH...
>

I have already cloned the source code so as to identify the
architectures that implement a kernel level syscall for
posix_spawn.

Best regards,

Andreas