Mail Archives: djgpp/1996/05/09/07:51:24
On Tue, 7 May 1996, Martin Brundage wrote:
> extender. The app is an ATM adapter diagnostic that has a large number of
> driver-level functions, including some embedded assembler. What are the
> potential risks, benefits, pitfalls, etc?
You will have to learn a different assembly syntax (GCC and GNU Assembler
both use the AT&T syntax). The interrupt handlers and the code that
calls real-mode interrupt services and/or accesses absolute memory
locations or buffers in conventional memory would have to be rewritten.
> How about debugging interrupt routines or displaying/changing virtual
> memory-mapped memory or I/O locations?
What are you used to do about these in BCC?
> How well does gcc work under DOS?
At least as good or better than the real-mode BCC.
> Are the DOCs adequate?
Are they ever? Seriously, I have yet to find a software product that
comes with adequate docs. For me ``adequate'' means that when I look for
an answer to a non-trivial question, I can find it without browsing
through the entire docs set. Most of the docs of any software I saw
until now are voluminous in describing trivial things that don't need to
be described at all, but hide crucial pieces of important info where no
one can find them. And the indices are usually horribly ineffective.
IMHO, DJGPP's docs are as good as any other. At least in DJGPP I can
look into the sources, or ask the people who wrote it, when in doubt,
which is almost never an option when I use Borland products.
> Thanks for all responses.
There is one very important aspect that you didn't ask about. The
support you get with DJGPP blows any other commercial extender off the
water. For any question or problem that you might have, you can just ask
here and you will usually have a solution within a few days at most. You
can change everything in DJGPP because you get the sources for free, and
you can talk to people who wrote it to ask for advice.
- Raw text -