home *** CD-ROM | disk | FTP | other *** search
- Path: ac.dal.ca!cook
- Newsgroups: comp.lang.c++
- Subject: Re: Let's save Borland C++ and OWL
- Message-ID: <1996Jan1.123321.44509@ac.dal.ca>
- From: cook@is.dal.ca (Adrian Ross Cook)
- Date: 1 Jan 96 12:33:20 -0400
- References: <4bmnv0$l9i@serv.hinet.net>
- Nntp-Posting-Host: is.dal.ca
- X-Newsreader: TIN [version 1.2 PL2]
-
- I whole-heartedly agree with what you've said. The quality of Borland
- documentation is really awful. I've had more than one similar experience
- using Turbo C++ for Windows. My favourite memory is of trying to figure
- out what STRICT was all about and how it should be used. Borland led me
- on a wild goose chase through all the documentation I had. I still
- haven't got it to work properly.
-
- It's really too bad, because the OWL library does seem really
- well-designed and Borland compilers seem fine in all aspects but their
- documentation.
-
- Adrian Cook
- cook@is.dal.ca
-
- Lyndon Lin (ll7755@c2.hinet.net) wrote:
- : After thorough experiences with both VC++ and BC++, I believe that BC++
- : and it's OWL framework library is worth of living in the market. But
- : Borland's documentation is too much worse in compared with Microsoft's.
- : This hurts a lot for public acceptance of them. I'm now collecting all my
- : witness about the error, incompleteness, confusing of their documents.
-
- : Let me give you one of them.
-
- : Do you know how to trace into OWL's source code?
- : I was driven to mad in searching for the procedure. After reading
- : through all the relevant documentation (Turbo Debugger User Guide, BC++
- : User Guide, Borland Tech. Info in Borland's Web site, ReadMe files), I
- : was still out of luck. I just found that BC++ installation program didn't
- : copy the necessary diagnostic version of DLL and import library(I choosed
- : to have a FULL installation). So I copy it by hand. It help nothing. It
- : seems to be resonable since in C++ termilogy, diagnostic means exception
- : handling and Run Time Type Identification (RTTI), it is nothing to do
- : with symbolic debugging. Then I reviewed the supplied make file for OWL
- : library source. (I's was never a make file lover.) It's evident that
- : you can build a debug version of the OWL by defining a directive. But
- : it is also evident that all of BC++'s runtime library should follow
- : some naming convention. For example, the BC++ 4.50 dictates the 16-bit
- : DLL version of OWL library be named OWL250.DLL for distribution and
- : OWL250D.DLL for diagnostics version. But what is the name of a debug
- : version? Nothing was said about it. It seemed to me Borland doesn't
- : giave a position for debug version library.
-
- : But in the make file, it mentions a directive to stripe off the debug
- : infomation from a debug version EXE or LIB/DLL. In the stripping process,
- : you can opt to create a .TDS file to hold the debug infomation. It
- : doesn't explained further about what is the .TDS file for. Then I recall
- : that in the Turbo Debugger (but not the integrated debugger), when you
- : tried to load another symbol file, it prompt you to supply file with .TDS
- : extention. So I searched again in the CD and found the OWL250.TDS. Sound
- : interesting. I copy it into BC++ BIN directory. Try again. TDW still
- : prompt me to supply a .TDS file (Shouldn't it be able to find the TDS
- : file in the BIN directory?). OK, I pick up the file with the file browser
- : dialog. No luck. TDW sielently accept the TDS, but I still can't trace
- : into OWL source. Maybe I should stick with VC++ anyway.
-
- : Two days later, I accumulated some more momentum for this mess. Let's
- : try again. In a very occation, I check out the "Diagnostic" check button
- : in the Target Expert dialog to compile my project with diagnostic
- : information. You guess what. When I start up TDW, it bring me to the
- : first assembly instruction in the good old WinMain() function. I got it!
- : In a OWL application, WinMain() exists in the OWL library, not in user
- : code. It means TDW has automatically loaded the OWL library symbol
- : infomation and correctly pinpoint the first line of code to be executed.
-
- : But what about integerated debugger? We usually debug with it for being
- : able to access to other programs and most importantly, all the necessay
- : on-line library reference. It didn't behave as good as TDW, but improve a
- : lot. Immediately at starting up, it prompted you for the Directory to
- : locate a "Applicat.cpp" file. This is the source file for the
- : TApplication class. So the IDE debugger knows what file contains the
- : library source code, it just doesn't know where it is put. After choose
- : the correct directory using the file browser dialog, IDE debugger behaved
- : as good as TDW. Then I try to set some directory information in the
- : project option dialog. The IDE is also able to locate all the necessary
- : source module automatically. This directory setting is documented in the
- : IDE User Guide, but it's not likely that you would think this setting
- : also apply to library source.
-
-
- : You may challenge me why not just call Borland tech support. First, I
- : live in Taiwan, and register with Borland Taiwan. I can't call
- : U.S.A directly for tech support. As for Borland Taiwan, I doubt they have
- : the necessary task force for helping with serious tech. problem regarding
- : C++.
- :
- : Second, this is only one of several hundred of potential problems we
- : are doomed to encounted with, all just because Borland's incapable of
- : suppling the useful documentation we deserve. If Borland doesn't see this
- : as a serious problem, you and me will be forced to switch to Microsoft,
- : and finally the death of an execellent compiler and framework library.
- : It's really a shame to waste you American's telent and effort this way.
-
- : I really wish this posting will arouse all the BC++ user giving more
- : and more of their complaint and witness. Only in that way can we force
- : Borland to improve their documentaion for their next version of software.
-
-
-