Xref: utzoo comp.os.msdos.programmer:3647 comp.windows.ms.programmer:1066 Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!ames!dftsrv!oneel From: oneel@heawk1.rosserv.gsfc.nasa.gov ( Bruce Oneel ) Newsgroups: comp.os.msdos.programmer,comp.windows.ms.programmer Subject: First impressions of Borland C++ 2.0 Message-ID: Date: 28 Feb 91 15:19:40 GMT Sender: news@dftsrv.gsfc.nasa.gov Distribution: na Organization: STX/GSFC Nasa Lines: 64 I got my copy last night and here are my first impressions. This is a long note. First, don't laugh, but here is the machine. PS/2 M25, 8mhz 8086, 640k memory, 20 meg 35ms HD. Not a very capable windows machine. Windows runs OK on it. Not fast but tolerable. The package has about the same size/weight factors as TC++ Pro. It comes on 11 disks (720k 3.5"). I ordered it on 2/14 at 12:30 EST and it was shipped from IL on 2/19 via UPS. I started out with 11 meg free and didn't install the assembler (I don't do assembly), the profiler (It doesn't work in windows to the best of my knowledge), and the debugger ( I was worried about the space). I also only installed the small and large models and and the windows support. While install doesn't tell you how much space you've used/have left it does give you an idea how much you need to install each piece. It complained that I might not have enough space to install before it let me start because they recommend about 15 meg free. It took about 1/2 hr. I went around and deleted all of the executables for protected mode, bccx, bcx, etc... When I was done I had 3 meg free for a net use of 8 meg. Since it was getting late and I wanted to try out windows I used their example book to compile whello.cpp. On a quick look this is quite similiar to the code in the slick mag they sent called something like Borland Language Express. It compiled and ran great!!! This is even in real mode, something that the readme warns against. I next tried the dlldemo program and I couldn't get that to work. Immediate big red switch crash. This had to be compiled within the IDE (no make files, just project files, but prj2mak claimed to exist) and it went kind of slowly. The pre-compiled headers seemed to speed things up quite a bit though. Windows.h compiled slowly the first time but from then on was very fast. The system was definitely working the disk quite hard. Within windows the debugger won't run on my system. It claims it's out of memory and I only have the program manager open. The whitewater resource tool kit runs, but it is painfully slow. Overall summary? For $100 it is great. I'm using a very underpowered machine and it is usable. Overall the documentation lacks a coherent feel from the windows point of view though. Most of the intro material is for DOS, when you hit the windows material there is very little hand holding. The readme is quite important. According to borland on CI$ the limitation on real mode windows apps comes from a problem with the linker which was resolved but not QC'd properly before shipment. It works but isn't supported until they QC it enough so that they are happy. It looks a bit rushed in places but this is really a combination of a 1.0 product and a 2.0 product. The windows part is 1.0 and the c++ is 2.0. I'm very pleased. bruce -- | Bruce O'Neel | internet : oneel@heasfs.gsfc.nasa.gov| | Code 664/STX | span : lheavx::oneel | | NASA/GSFC |compuserve: 72737,1315 | | Greenbelt MD 20771 | AT&Tnet : (301)-286-1119 |