[9fans] 3/27/01 update of C compiler

[9fans] 3/27/01 update of C compiler

[Ish Rattan]

Probably I should not be complaining.

It comes out that 3/27/01 update of C compiler does something
strange to compilation of Alef (8al) from 2ed.

The source is compiled correctly (no errors) if 'typestr' is taken
care of with a redefine.

The generated image (8al) compiles syntactically correct, alef programs
but faults (sys/trap) on programs with syntax errors.

If one generates the same image using pre 3/27/01 C compiler (thanks
Lucio) the image produces correct behavior.

So what was changed? Is there a way to undo it?

-ishwar

Re: [9fans] 3/27/01 update of C compiler

[rob pike]

> I remember mention of a pointer to the "typestr" extension (but
> not the pointer itself).  Are there other such?  Is it time for me
> to read the code? :-)  Is there a list of changes?

I'm pretty sure 'typestr' is the only significant change.  There are
a couple of new pragmas, but they're documented.

> PS: Rob, is there any sentiment within Bell Labs about releasing the
> Alef sources from 2ed?

Not much, no.

-rob

Re: [9fans] 3/27/01 update of C compiler

[rob pike]

'Typestr' is a new undocumented keyword in Plan 9 C.

-rob

Re: [9fans] 3/27/01 update of C compiler

[Lucio De Re]

On Fri, Sep 21, 2001 at 12:38:35AM -0400, rob pike wrote:
>
> 'Typestr' is a new undocumented keyword in Plan 9 C.
>
Well, I guess there are two possibilities:  (a) Ish didn't correct
all intances of "typestr" in the Alef sources or (b) the new "C"
compiler has changed at least enough to damage the Alef binaries.

I remember mention of a pointer to the "typestr" extension (but
not the pointer itself).  Are there other such?  Is it time for me
to read the code? :-)  Is there a list of changes?

++L

PS: Rob, is there any sentiment within Bell Labs about releasing the
Alef sources from 2ed?  I assume there are no non-disclosure
restrictions, and I'll take on the burden of supporting what limited
demand there may be.  Of course, it may be better to encourage use of
Limbo.