bobs@moliman.MoliEnergy.BC.CA (Robert J Silliker) (04/13/89)
In article <780@stag.UUCP> to_stdnet@stag.UUCP writes: >From: stag!thelake!steve@bungia.mn.org (Steve Yelvington) > >A problem, and some questions, on resource files: > >I've been using Bob Silliker's RSC2C.TTP program to decompose (is that the >right word?) GEM resource files into C sources. I ran into a small >incompatibility with the OBDEFS.H file I've been using. Stuff deleted. >RSC2C program produces code in which that field is either a long >hexadecimal number, a string, or the address of any of several other types >of array. If it is the former, Sozobon C refuses to compile the code, >since it is expecting a pointer. If it is the address of a non-char array, >it merely complains. >I recently acquired Mark Williams C, so I took a look at the OBDEFS.H file >that came with it. In that file, the declaration is changed to unsigned >long. >Now (finally) to my questions: > >* Is the MWC alteration of ob_spec from char * to unsigned long in > OSBIND.H appropriate? Should I alter my OSBIND.H file to unsigned long? I don't think this is a good idea unless it comes straight from MW. You will lose compatibility with MWC sources otherwise. >* Should I fix RSC2C so that it casts the field to a char pointer? I'm > leaning toward the this alternative. Yes. Please do not post the complete sources or the binary. Posting a patch is acceptable. I will be creating a new version that will allow you to change the type of cast made into ob_spec. This will probably take the form of a command line argument. Are there any casts other than (char *) and (unsigned long) that should be included? Anybody? +--------------------------------------------------------------------------+ | Disclaimer: The above opinions are my own and not those of Moli Energy. | +--------+----------------------------------------------------+------------+ | Bob Silliker | bobs@moliman.molienergy.bc.ca | | | ...uunet!van-bc!resrch!moliman!bobs | +----------------------------------------------------+