I hear you, being in a similar situation myself. I believe I roughly understand what is going on (though mostly I only care about up until the parsing stage), but not well enough to explain. Unless there is a trove of documentation I am missing, sparse at the moment for better or worse seems to be an experts only (experts in sparse that is!). The excessive amounts of comments I assume is said with a little 2nd degree. For me what is lacking is the type of broad comments (at the top of files say) that give an overall idea about how things fit together - insightful remarks. Probably most importantly the principal types are largely undocumented. What invariants apply to them, when and how they will be used. Ok, tokens are fairly self-explanatory and conventional, but Symbols... Now I would hasard to say a symbol represents anything that will be manifested in the compiled code, but this is probably inaccurate (I am not sure what happens to unused types for instance if this is true). A couple of questions then:- -Preprocessing operates on tokens (i.e. before symbols are created). So what role to NS_PREPROCESSOR and SYM_PREPROCESSOR have (and why are both required)? -What is the meaning of SYM_NODE? (A block or statement in the code??)
From the top of symbol.h:
/* * An identifier with semantic meaning is a "symbol". * * There's a 1:n relationship: each symbol is always * associated with one identifier, while each identifier * can have one or more semantic meanings due to C scope * rules. * * The progression is symbol -> token -> identifier. The * token contains the information on where the symbol was * declared. */ Perhaps this could be elaborated upon. "The progression is symbol -> token -> identifier." Is this correct? Not wanting to embarrass myself, I would have thought it would be something like tokens->(identifier,symbol), i.e. we start with tokens, which groups of which form symbols. One in the group will be the identifier. The identifier structure itself is not namespace specific, so they are shared between namespaces. Sorry could say more but that was already longer than intended! - Mike On 5/9/07, Mish Mash <khalota@xxxxxxxxx> wrote:
dear all, I'm trying to use sparse for writing various source code analysis tools (I was using ctool previously). I'm currently going through the source trying to figure out what I need. Given the excessive amounts of comments I'm having to wade through, this probably isn't the fastest way of figuring the beast out ;) Does anyone have any notes they've taken or can point me at code that uses sparse to help get a rough idea of how to use the beast? (Ideally slightly more detailed than Randy Dunlap's howto) I noticed smatch now uses sparse and i've applied their patch which I'm going over. Is there anything better? (I've already found the no longer very useful presentation at http://developer.osdl.org/dmo/sparse/) Thanks! nessim - To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
