Re: Why does Pro*C suck so bad?????

Bob Petrocelli (rrp_at_cfm.brown.edu ) writes:
>
> Hello Fellow Oracle Victims and Co-dependents!
 How apropos.
>
>
> I just wanted to share a little PRO*C nightmare with you all.
> I have recently purchased Oracle 7.1.4 and Pro*c 2.0.4.0 for
> My Hp 9000/725-100 running Hp-ux 9.05.
>
> To my dismay after installing Pro*C and compiling the sample
> programs I discovered the following:
>
> sample1
>
> bus error
> core dump
>
>
> I inserted some printf's and discovered that the sample
> program crashes and burns on the CONNECT statement!
>
> After several calls on the resulting TAR I am told that Pro*C has
> only been tested using CC version 9.34 (and on only an HP 9000/800).
>
> The current version of the CC compiler is 9.7!
>
> I am now desperate to get hold of HP C Compiler 9.34. User
> support had the audacity to tell me that I should get HP to
> FIX THE PROBLEM (which is clearly with the Oracle product) HA!
>
> Am I the only one who is discovering that Oracle products are
> SHIT! Should I have specified SYBASE for this project?
>

Actually, a number of their products are actually shinola. I have worked with Oracle products since the late stages of version 4, and have installed a number of versions on more different kinds of platforms than I like to think about. Once installed, a properly configured Oracle instance is practically bulletproof short of concurrent disk crashes on the data and log spindles. The tools are in general quite reliable, although not to the degree of the DBMS. Not so the installation software or the Pro*C interface, in my experience. Although I may have done an Oracle install where there were no explosions or nasty surprises, I can't recall it. On version after version and platform after platform, I've seen syntax errors in install scripts, illegal defaults, inadvertent tight recursions, and incomplete media contents (especially link libraries). To add insult to injury, installation support is inconsistent, frequently inaccurate, and about as speedy as a glacier. I also have running reports from programmers concerning Pro*C stuff that works in one environment (e.g. SCO Unix) and not in another (e.g. hpux). Last one I followed up on personally appeared to work on SCO only because of a bug in the Oracle Pro*C pre-compiler; the failure under hpux was actually the correct behavior.

If you're trying to do this via client-server architecture, you have more interesting times in store. Hope you can find a way out of your nightmare-- with a site license and a bunch of platforms in place, I'm going to have to endure mine.

W.C. Epperson     "These are my own opinions, but you can help yourself:
Senior S.E.        I got a plenty where these came from."