Re: [TUHS] Looking for final C compiler by Dennis Ritchie

[Dibyendu Majumdar <mobile@majumdar.org.uk>]

[-- Attachment #1: Type: text/plain, Size: 2406 bytes --]

On 19 July 2018 at 15:50, Clem Cole <clemc@ccc.com> wrote:

>
> On Wed, Jul 18, 2018 at 5:37 PM, Dibyendu Majumdar <mobile@majumdar.org.uk
> > wrote:
>
>> I am interested in finding out if the last C compiler code (not the
>> earliest versions which I know
>> are available) written by Dennis Ritchie is available somewhere. I
>> assume that the C compiler in V7 code was written by him?
>>
>> ​I'm not sure if this is the last.   This is a pointer to the V7 Ritchie
> Compiler:  https://minnie.tuhs.org//cgi-bin/utree.pl?file=V7/usr/src/cmd/c
> - The sources and the makefile build the three passes /lib/c[012]
> This should be a good starting point/base line.
>
>
Many thanks - I will compare this with
https://github.com/eunuchs/unix-archive/tree/master/PDP-11/Trees/V7/usr/src/cmd/c
which is what I have been looking at.

I guess that by this time the work had transitioned to pcc so probably
there isn't a later version available?


> Be careful because the Johnson Compiler (pcc) was also included with V7
> and is a different technology.
>

Yes understood.


>
> This is important because their are modifications to both the Ritchie and
> Johnson compilers 'in-the-wild' for other back-ends and new optimizations.
>   I for instance, re-targeted the Ritchie compiler to what would become the
> 68000 (it was not yet numbered, it was an experimental chip when we had
> access to it in the late 1970s in Tek Labs - mine was a 16 bit 'int' as I
> was coming primarily from the PDP-11 at the time and the chip was a 16 bit
> chip internally - so the code was tight and clean and I basically
> substituted PDP-11 instruction sequences for 68000 sequences).   IIRC, Jack
> Test's 68000 compiler from MIT which was about 18 mons later was based on
> the Johnson compiler but he used a 32 bit 'int' which proved easier for
> porting programs from the Vax, as the chip supported 32 bit words even
> though it took 2 ticks to do anything [so Jack's compiler generated slower
> code for many simple ops].
>
> I recommend, that google for the old USENIX tapes and see what you turn up
> and compare.
>
> ᐧ
>

Thank you for the info - I will certainly look at the USENIX tapes.

I will try to port the C compiler to amd64 - while preserving as much of
the original code as I can. But not sure if this is even feasible.

Thanks and Regards
Dibyendu

[-- Attachment #2: Type: text/html, Size: 5055 bytes --]