[9fans] Binit and Bterm

[9fans] Binit and Bterm

[Joel “chesky” Salomon]

I’ve got a function in the Brain-Dead SHell™ that I’m writing for
homework that looks like this:
	Biobuf in;

	void bdsh(int f) {
		Binit(&in, f, OREAD); yyparse(); Bterm(&in);
	}
This function gets called for each file named in argv[] after main()
open(2)s it.

My question is: is the call Bterm necessary, or will Binit
reinitialize the global Biobuf correctly on its own—or am I courting
trouble some other way?  The man page for bio(2) is a bit vauge on
reusing Biobufs.

--Joel

Re: [9fans] Binit and Bterm

[Tim Wiess]

    since you are just doing reads then no i don't believe
    the Bterm is necessary.


> I’ve got a function in the Brain-Dead SHell™ that I’m writing for
> homework that looks like this:
> 	Biobuf in;
> 
> 	void bdsh(int f) {
> 		Binit(&in, f, OREAD); yyparse(); Bterm(&in);
> 	}
> This function gets called for each file named in argv[] after main()
> open(2)s it.
> 
> My question is: is the call Bterm necessary, or will Binit
> reinitialize the global Biobuf correctly on its own—or am I courting
> trouble some other way?  The man page for bio(2) is a bit vauge on
> reusing Biobufs.
> 
> --Joel

Re: [9fans] Binit and Bterm

[Joel “chesky” Salomon]

> since you are just doing reads then no i don't believe
> the Bterm is necessary.

So will Binit properly reinitialize its argument, reusing the buffers
rather than allocating new ones and leaking the old?

--Joel

Re: [9fans] Binit and Bterm

[Russ Cox]

The only buffer is on the stack inside the Biobuf.
There is no dynamic allocation in this case.

Russ


On 10/5/06, Joel "chesky" Salomon <JoelCSalomon@gmail.com> wrote:
> > since you are just doing reads then no i don't believe
> > the Bterm is necessary.
>
> So will Binit properly reinitialize its argument, reusing the buffers
> rather than allocating new ones and leaking the old?
>
> --Joel
>
>

Re: [9fans] Binit and Bterm

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 240 bytes --]

according to /sys/src/libbio/binit.c, Binit doesn't
allocate buffers (that's done by Bopen), and Bterm
for a Biobuf initialised to OREAD by Binit doesn't do much.

i'd still call Bterm though, especially in these days of cut and paste

[-- Attachment #2: Type: message/rfc822, Size: 2472 bytes --]

From: Joel “chesky” Salomon <JoelCSalomon@Gmail.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] Binit and Bterm
Date: Thu, 5 Oct 2006 16:33:32 -0400
Message-ID: <8eeb3016afcfba5a3719301de1824a5f@plan9.jp>

> since you are just doing reads then no i don't believe
> the Bterm is necessary.

So will Binit properly reinitialize its argument, reusing the buffers
rather than allocating new ones and leaking the old?

--Joel

Re: [9fans] Binit and Bterm

[Joel “chesky” Salomon]

> i'd still call Bterm though, especially in these days of cut and paste

I think I’d rather have things break on cut-and-paste-ers. ☺

--Joel

Re: [9fans] Binit and Bterm

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 74 bytes --]

i suppose i don't see that problem in calling the function.
what is it?

[-- Attachment #2: Type: message/rfc822, Size: 2427 bytes --]

From: Joel “chesky” Salomon <JoelCSalomon@Gmail.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] Binit and Bterm
Date: Thu, 5 Oct 2006 17:00:32 -0400
Message-ID: <96d9086fc9539edad73f1242fb629a9a@plan9.jp>

> i'd still call Bterm though, especially in these days of cut and paste

I think I’d rather have things break on cut-and-paste-ers. ☺

--Joel

Re: [9fans] Binit and Bterm

[Joel “chesky” Salomon]

> > I think I’d rather have things break on cut-and-paste-ers. ☺
> 
> i suppose i don't see that problem in calling the function.
> what is it?

I’m actually leaving the Bterm in the code for clarity; I was just making a joke about leaving brittle code about so it can’t be copied without understanding it.

--Joel

Re: [9fans] Binit and Bterm

[Martin Neubauer]

* Joel âcheskyâ Salomon (JoelCSalomon@Gmail.com) wrote:
> I'm actually leaving the Bterm in the code for clarity; I was just making
> a joke about leaving brittle code about so it can't be copied without
> understanding it.

It will anyway.

Martin

Re: [9fans] Binit and Bterm

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 295 bytes --]

yes, i was going to suggest that perhaps there should be a rich text
format  in plan 9 for source code storage allwoing subtle code to be tagged so
could not be edited without understanding it, or at least reading it,
but then i realised that i wouldn't then be able to edit code of my own.

[-- Attachment #2: Type: message/rfc822, Size: 3113 bytes --]

From: Martin Neubauer <m.ne@gmx.net>
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Subject: Re: [9fans] Binit and Bterm
Date: Thu, 5 Oct 2006 23:38:09 +0200
Message-ID: <20061005213809.GA11079@shodan.homeunix.net>

* Joel âcheskyâ Salomon (JoelCSalomon@Gmail.com) wrote:
> I'm actually leaving the Bterm in the code for clarity; I was just making
> a joke about leaving brittle code about so it can't be copied without
> understanding it.

It will anyway.

Martin

Re: [9fans] Binit and Bterm

[Tim Wiess]

> So will Binit properly reinitialize its argument, reusing the buffers
> rather than allocating new ones and leaking the old?

    yes the buffer will be reinitialized, but as Russ said, it is
    not dynamically allocated.

Re: [9fans] Binit and Bterm

["Nils O. Selåsdal"]

Joel “chesky” Salomon wrote:
> I’ve got a function in the Brain-Dead SHell™ that I’m writing for
> homework that looks like this:
> 	Biobuf in;
> 
> 	void bdsh(int f) {
> 		Binit(&in, f, OREAD); yyparse(); Bterm(&in);
> 	}
> This function gets called for each file named in argv[] after main()
> open(2)s it.
> 
> My question is: is the call Bterm necessary, or will Binit
> reinitialize the global Biobuf correctly on its own—or am I courting
> trouble some other way?  The man page for bio(2) is a bit vauge on
> reusing Biobufs.

What's one extra, correct, Bterm call among friends ?