home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.databases.oracle
- Path: sparky!uunet!mcsun!chsun!osnbe!
- From: rheiger@renext.eiger.olivetti.ch (Richard H. E. Eiger)
- Subject: Re: Pro*C
- Message-ID: <1992Dec20.105229.955@osnbe.Olivetti.ch>
- Sender: @osnbe.Olivetti.ch
- Nntp-Posting-Host: renext
- Reply-To: rheiger@renext.eiger.olivetti.ch
- Organization: Olivetti (Schweiz) AG, Branch Office Berne
- References: <1992Dec17.220804.12585@tigger.jvnc.net>
- Date: Sun, 20 Dec 1992 10:52:29 GMT
- Lines: 50
-
- In article <1992Dec17.220804.12585@tigger.jvnc.net> fogelinc@nmr1.Cyanamid.COM
- (Carl Fogelin) writes:
- > In article 13161@den.mmc.com, richard@crowded-house.den.mmc.com (Richard
- Armstrong) writes:
- > >
- > >I have plenty of good things to say about Oracle Pro*C. However, let me
- > >focus on one area where I find Pro*C completely brain-dead.
- > >
- > [stuff deleted]
- > >
- > >Why doesn't the Pro*C precompiler do #define substitutions?
- >
- > Because the scope of the Pro*C preprocessor would probably get out of hand.
- > Sure, it would be nice if the Pro*C preprocessor would at least recognize
- > #defines which are at the top of the current module, but the c preprocessor
- > can be much more complicated. What about #defines found in an #include file?
- > What about #defines which use conditional assignments (i.e., #ifdef)? These
- > are just a couple of complications, but if you think about it, I'm sure you
- > can come up with more.
- >
- > Yes, I agree that I wish Pro*C did the #define substitutions, but I
- understand
- > why they didn't do it. What I don't understand is why "struct"ures are not
- > valid for host variables, but VARCHARs (which are structures) are... 8-(
- >
- > -------------------------------------
- > Carl Fogelin (fogelinc@cyanamid.com)
- >
-
- The problems described here are a nuisance, I agree. But there are things I
- find much much worse in PRO*C. For instance, you can specify ireclen and
- oreclen. But, PRO*C still has an upper limit. When that is reached your output
- will be erratic without warning!
-
- The other real stupidity is that PRO*C is not clever enough to put "#line nnn
- file.pc" statements into the output. It would be so much less of a pain to find
- errors in the .pc source when the C-compiler complains. And also runtime output
- from the assert macro would finally be useful.
-
- The solution to this is to preprocess the .pc file before feeding it to pcc.
-
- Just my 2 cents.
- Richard
- --
-
- ___ _____2
- / ) / / / Richard H. E. Eiger
- /_ _/ /__/ /__ Ing. Informatik HTL
- /\ / / / Unregistered NeXT fan
- / \ / / /_____ rheiger@renext.eiger.olivetti.ch (NeXT mail welcome)
-