🚀 go-pugleaf

In article <1990Nov28.125836.10515@watserv1.waterloo.edu> tom@mims-iris.waterloo.edu (Tom Haapanen) writes:
>Alen Shapiro <alen@crash.cts.com> writes:
>	
>> 5) Why do I need 2 compilers 1 for development and the other for
>>    production? Why is the standard comiler so slow (an hour to
>>    compile on the PS2/80, (15 minutes on the macII and 5 minutes
>>    on SPARC))? Why does the development compiler (-qc) not work
>>    in so many cases?
>
>An hour?!?  Our 25,000-line application takes 16 minutes to compile, link
>and rc on a 386/25 (with a 1 MB SmartDrive) with MSC 5.1 and default
>optimization.  Are you using a disk cache?  Is that a 16 MHz PS/2-80?  Or
>is your application really immense?
>
>We don't use -qc --- why bother when even -Ox is fast enough?

	I don't use -qc either, but that's because it doesn't work.
Two comments: MSC 6.0 is visibly (even dramatically) slower than 5.1
was; MSC 5.1 is very slow compared to most other commercial compilers.  

	You seem shocked at his 1 hour compile time (I am too), but are
you really happy that you only wait 16 minutes to compile only 25,000
lines on a 386/25 (using caching and a SmartDrive)?  Either Turbo C
from Borland or Zortech C will compile our 60,000 line (largest build
probably 30,000 lines after preprocessing) in about 7 minutes on a
386/20 with RAM drive.  With MSC 6.0 I usually go out for lunch when
doing a complete rebuild.  The guy in my office porting to DEC 3100 laughs that
he doesn't need to use make because it recompiles all the code while
I'm waiting for one or two modules.

	What seems good is what you are used to--and I'm spoiled and
glad of it.

// *************************************************************** //
John Gossman
SoftSource      (206)676-0999 Phone
WWU Math Dept.
My employer stands behind all my opinions, except in public.
// ************************************************************** //