info-mac@uw-beaver (01/25/85)
From: Scott Comer <wert@rice.ARPA> I am less than pleased. 1) Toolbox routines don't follow Apple case conventions. Porting software from Sumacc is therefore not a "snap", because some poor soul must change all the toolbox calls. Bletch. 2) The C compiler doesn't use the nice file opening interface (like MacWrite, or even their editor). "C-boot:edit.c" Yum, Yum. Moreover, if you get the file wrong (or right, after it does its bit) the compiler exits. 3) Same for the linker. Also, it bombed out inexplicably while loading external symbols and ran perfectly the second time. Yow! Are we having fun yet? All in all, this is a good idea, if only someone had done it right (maybe used the standard user interface guidelines). Just to let people know that there are problems, too. We've heard too much good news about this compiler to believe it could be true, and it isn't. Scott Comer and David Chase of Rice University, checking this product out
info-mac@uw-beaver (01/28/85)
From: Mike Caplinger <mike@rice.ARPA> In response to the recent posting by David Chase and Scott Comer of Rice, I thought I would express the opinion of the other one-third of our office. True, the file interface in the compiler and linker are poor. I surmise that the system was compiled by itself, and they just didn't bother to call the toolkit routines. On the up side, though, this is the first Mac program that isn't so SLOW as to be a pain to use! The compiler is quite fast; the linker is rather doggy, but presumably you link less often. The compiler error messages are dumped in an errors file; just as well, since the compiler exits almost immediately after displaying them, clearing the screen. And the resultant applications start up very quickly. [At least the Consolair editor is pretty nice. These guys didn't make the P-system mistake!] So the interface is lacking. I would rather have some speed than a slick user interface, at least in something as hacky as a C compiler. [Once the documentation for our evaluation copy comes in, I will probably have more to say.] - Mike
info-mac@uw-beaver (02/04/85)
From: (Mike O'Dell[x-csam]) mo@lbl-csam THere is one serious Gotcha in the Megamax documentation I looked at for the much rumored V2. A "short" is a 1-byte integer!! This is at lease strange, because char's are naturally signed on the 68K, so why they would screw short's is beyond me. While this interpretation is within the province of the White Book, one would think intellegent people would try to make their compiler work in ways similar to all the other C compilers which have come before it. But after having seen several Mac C compilers (no to mention 8086 compilers), I am conviced most of the people doing them don't have the slightest idea as to how their software should work. -Mike O'Dell