chip@tct.uucp (Chip Salzenberg) (05/30/90)
I am currrently a user of SCO Xenix/386. However, I expect that I
will soon be moving to SCO Unix or perhaps another brand of Unix/386.
The C safety-check program "lint" that is part of the Xenix/386
development system is, well, poor. It gets confused about structure
definitions, it doesn't understand function prototypes (!), etc.
I've gotten info on Gimpel's portable lint ("FlexeLint"). It has
everything including the kitchen sink, but it will set me back $800,
and I'd hate to spend that much money if Unix/386 has a decent lint.
So, my question: How good is the lint provided with modern versions of
Unix/386? Specifically, can it deal with prototypes? Please send
E-Mail or post a followup article, as you feel appropriate.
--
Chip Salzenberg at ComDev/TCT <chip%tct@ateng.com>, <uunet!ateng!tct!chip>walter@mecky.UUCP (Walter Mecky) (06/01/90)
In article <2662C59D.3C5A@tct.uucp> chip@tct.uucp (Chip Salzenberg) writes:
+ I am currrently a user of SCO Xenix/386. However, I expect that I
+ will soon be moving to SCO Unix or perhaps another brand of Unix/386.
+
+ The C safety-check program "lint" that is part of the Xenix/386
+ development system is, well, poor. It gets confused about structure
+ definitions, it doesn't understand function prototypes (!), etc.
+ []
+ So, my question: How good is the lint provided with modern versions of
+ Unix/386? Specifically, can it deal with prototypes? Please send
+ E-Mail or post a followup article, as you feel appropriate.
If you mean SCO UNIX when you say Unix/386, you can stay to XENIX:
lint seems to be the same as in XENIX: no prototype (==> syntax error),
confused about structure adresses too.
--
Walter Meckyhb@vpnet.chi.il.us (hank barta) (06/02/90)
Sigh. The SCO Unix Development Kit I bought late last year included a (draft) ansi C compiler (Microsoft's) and a K&R compliant lint. White Oak Software Inc. Henry Barta 0 South 258 Park (708) 510-0578 (voice) Winfield, IL 60190 ...!vpnet.chi.il.us!pswin!barta Predictable systems by design. hb@vpnet.chi.il.us
wengland@stephsf.UUCP (Bill England) (06/14/90)
In article <2662C59D.3C5A@tct.uucp> chip@tct.uucp (Chip Salzenberg) writes: >So, my question: How good is the lint provided with modern versions of >Unix/386? Specifically, can it deal with prototypes? Please send >E-Mail or post a followup article, as you feel appropriate. >-- >Chip Salzenberg at ComDev/TCT <chip%tct@ateng.com>, <uunet!ateng!tct!chip> Sadly the lint on SCO Unix does not reconize prototypes. However, if you are going to spend the money on a 'decent lint' you could go for a full blown ansi standard c compiler like Metawear's. Tom Pennello and Frank Deremer have been _very_ responsive to porting their compiler to other Unix Systems in the past ( Convergent for Example ) and I'm sure they would love to sell some to the new Unix PC Environment's. There is a MetaWare Inc. in the mail maps with an address off of uunet or ucscc. Try adam@metaware.uucp. I have used the compiler ( Convergent Systems )and it has excellent type checking, prototyping ansii compatable modes, generates great code, use's lots of 386 instructions, - drop em a line, it beats a simple lint program! - I'll bet that if SCO used the MW compiler that their system preformance would improve. ( Goes also for Interactive and Esix. ) +-------- | Bill England | Stephen Software Systems, Inc., Tacoma Wa. | wengland@stephsf.com +1 206 564 2122 | * * H -> He +24Mev * * * ... Oooo, we're having so much fun making itty bitty suns * * *
cpcahil@virtech.uucp (Conor P. Cahill) (06/14/90)
In article <152@stephsf.UUCP> wengland@stephsf.UUCP (Bill England) writes: > I'll bet that if SCO used the MW compiler that their system preformance > would improve. ( Goes also for Interactive and Esix. ) However, new compilers introduce new bugs and usually take a significant effort to implement (i.e. to solve the compilation problems, add additional compiler directives, etc). I'm not saying that it will, just that it probably will. Someone would have to run a test compile using the new compiler and then evaluate the effort required to do the port (changing compilers is equivalent to a system port). The other side of the issue is that someone would have to guage the expected performance increase and whether the performance gain would be perceived by the user (i.e. changing it so that grep only took 1.0 seconds instead of 1.3 seconds would be almost un-perceivable to the normal user even though the statistics would show an almost 30% gain). -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170