1. RE: fptr / Matt Lewis
- Posted by Matthew Lewis <matthewwalkerlewis at YAHOO.COM> Feb 13, 2003
- 505 views
Yes, just yesterday I considered that, since I came across that problem. For now, I use this piece of code to convert: constant sign_conversion = allocate(4) global function signed( atom x ) poke4(sign_conversion,x) return peek4s( sign_conversion ) end function In following the peek4 example, should I name them call_cdecls() and call_cdeclu()? Or just add call_cdecl_signed() to be really explicit? Matt Lewis > From: Andy Serpa [mailto:ac at onehorseshy.com] > Hi Matt, > > You ought to add duplicate versions of (or add a parameter to) the > call_xxx() functions that return signed values rather than > unsigned. I > find most libraries I use return signed error codes from the > functions > and put the "real" return value in some pre-allocated > location that you > pass as a parameter....
2. RE: fptr / Matt Lewis
- Posted by Andy Serpa <ac at onehorseshy.com> Feb 13, 2003
- 554 views
> In following the peek4 example, should I name them call_cdecls() and > call_cdeclu()? Or just add call_cdecl_signed() to be really explicit? > In your old version I just made a duplicate called call_cdecl_s() -- "signed" is too much typing, "cdecls" is too ugly... -- Andy > Matt Lewis > > > From: Andy Serpa [mailto:ac at onehorseshy.com] > > > Hi Matt, > > > > You ought to add duplicate versions of (or add a parameter to) the > > call_xxx() functions that return signed values rather than > > unsigned. I > > find most libraries I use return signed error codes from the > > functions > > and put the "real" return value in some pre-allocated > > location that you > > pass as a parameter.... > >