On Monday 07 May 2007 8:33 pm, Harald van Dijk wrote: > Hi, > > Here's a patch to handle spacing during stringizing.
Very nice. :) > It saves an extra integer per token in token strings, containing the token's > flags. This can include TOK_FLAG_BOW (beginning of word), indicating that a > space needs to be inserted in front. Since this is the only flag of > relevance, it wastes space. You might be able to move this flag into a bit > in the token number, if the waste concerns you. It's not the waste of memory that worries me, it's the extra argument being added to all those functions adding complexity to the code. Hmmm... Why did you renumber all the existing TOKEN_FLAG_* entries? (Why not just add TOKEN_FLAG_BOW to the end and make it 8?) Ok, so the rationale for this patch being so intrusive is that the current global "tok_flags" is transient, and isn't saved in tok_str_add. I've just started reading into this area of the code (so far I've read about 20% of the total tcc code thoroughly enough to more or less understand it), I have to make perplexed faces at this for a while longer... Rob _______________________________________________ Tinycc-devel mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/tinycc-devel
