|
From: | grischka |
Subject: | Re: [Tinycc-devel] Basic patch for passing W9X short DOS paths to TCC. |
Date: | Wed, 15 Apr 2009 15:50:56 +0200 |
User-agent: | Thunderbird 1.5.0.10 (Windows/20070221) |
lostgallifreyan wrote:
Joshua Phillips <address@hidden> wrote: (15/04/2009 13:46)stricmp or strcasecmp would be provided by the C library. Tcc is only a compiler - it doesn't include a C library. Therefore, the presence of stricmp or strcasecmp will be determined by the C library you're using (GNU libc or MSVCRT or other)Ok, that makes sense, but if it were only the library that determines what's there, TCC might compile C++. As it doesn't I assume it's not just the library that determines this. (That remark is based in my possibly erroneous assumption that C++ is a kind of superset of C with a lot more dedicated functions). In short, doesn't the compiler have to 'understand' what it's compiling?
In short, yes and no. The initial vocabulary (keywords) of a C compiler is rather small: char short int long float double enum void signed unsigned typedef struct union sizeof const extern static if else while do for continue break goto return switch case default And some others. The complete list for TCC is in "tcctok.h". A C++ compiler knows a few additional keywords, like bool class new delete ... However the compiler "learns" more words as it reads the include files and the program text. E.g. with TinyCC it learns "pstrcat" and "dynarray_add" and "put_ext_sym2" and whatnot. It learns enough about these words to speak them, but it doesn't understand their meaning. It understands the meaning of its own keywords though.
[Prev in Thread] | Current Thread | [Next in Thread] |