www.pudn.com > 文件恢复及修补 C 语言源程序.zip > ARPA-C.DMP
Retransmitted from ARPANET BY J Morrell 10/03/83: From: Martin SchoffstallSubject: Lattice C V1.04 and C-food I just got a copy of the "new" Lattice C compiler (v1.04), (was the article in Byte reviewing v1.24 a typo?). Over all it looks much better than v1.01 with a real debugging switch and a disassembler to work with it. C-food makes me want to cry! All those .asm programs that I massaged to interface with Lattice ALL IN A LIBRARY. (IBM why don't you supply us with a LIBRARIAN!!!) C-food contains code to handle graphics, the asynch port(s), directories getting the time in a usable form etc. The only think still lacking is floating point, and a larger model for the compiler (>64K text would be great, but >64k of text and >64k of data would be worth sacrificing a bullock about, you can even make structures larger than 32K illegal) Those two I can wait for and I certainly am not flaming, however, I do have three flames so far, one for IBM, and the others for Lattice. 1) the IBM linker (2.0) declares it can't find "bdos" in library lcx when given in this order Libraries: [.LIB] lc lcx but it can for Libraries: [.LIB] lcx lc How can the linker be that stupid, (don't answer that)? 2) the IBM linker declares that cscanf and something else are defined twice when in your program you declare #include #include but #include #include is fine. I understand why but it still is pretty foolish. 3) C-food has very few examples, take handling the async port: do you do ainit(); pcasm(); . . aterm(); or pcasm(); ainit(); . . aterm(); As it turns out the latter seems to work. The cost is very high, $500 list for the C compiler, and $150 list for C-food, of course you can get it for less. If you are going to get the compiler I strongly suggest you get C-food, it will be very useful. It doesn't irk me too badly since what I hate is paying a premium price for junk!