Path: utzoo!utgpu!watmath!clyde!att!rutgers!mailrus!cornell!uw-beaver!rice!titan!phil From: phil@titan.rice.edu (William LeFebvre) Newsgroups: comp.sys.amiga.tech Subject: Re: Unlock? AarRgh! Message-ID: <2262@kalliope.rice.edu> Date: 1 Dec 88 22:55:10 GMT References: <2174@kalliope.rice.edu> <1067@amnesix.liu.se> <1413@aucs.UUCP> Sender: usenet@rice.edu Reply-To: phil@Rice.edu (William LeFebvre) Organization: Rice University, Houston Lines: 40 In article <1413@aucs.UUCP> 850347s@aucs.UUCP (Hume Smith) writes: >In article <2174@kalliope.rice.edu> phil@rice.edu (William LeFebvre) writes: >>Well, imagine my surprise when the linker told me last night that the >>reference "Unlock" was not resolved. Turns out I wanted "UnLock".... > >I can't resist... > >Maybe you should (gasp) edit one of the dos include files (dos.h? >libraries/dos.h? /dosextens.h? /doseverthinelse.h?) and put in a line > >#define Unlock UnLock > >This should be safe under normal use (hand me that fire extinguisher). That's not the point. The point is that the supplied programmer interface is inconsistent. I'm sure that this is not the only example. It just took me by surprise....that's all. The libraries also use the spelling "WorkBench"! Besides, if I did that to my copy of a standard include file, my software would not be very distributable, would it? Besides besides, I'm using assembly. I'd have to do "_LVOUnlock equ _LVOUnLock" and " xref _LVOUnLock". >Why fret about UnLock? Who can remember those BltBitMaskBitMapRastPort >names? Well, but at least the capitalization is right. :-) What really burned me about this is: I double check all my library calls with the appendix in Commander's book (I don't remember the title, but about the only thing the book is good for is that appendix) that has a very compact listing of function and register parameters, and when I checked that call it listed it as "Unlock". That's one reason I was surprised. William LeFebvre Department of Computer Science Rice University