This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
| Other format: | [Raw text] | |
One of the problems with slimming down the common bits of the tree data structure is the multi-purposing of fields in that common part. While abstraction macros have been used for many other multi-purposed fields, they have (generally) not been used for the fields in tree_common. The attached patch addresses (part of) that by introducing DECL_CHAIN to make it obvious when DECLs are being chained together as opposed to generic TREE_LISTs or similar. Such a change has already been done for BLOCK_CHAIN. Making this change paves the way for an eventual pushing down of tree_common.chain into only structures that use it. (There are a number of changes remaining before that pushdown can happen: cp/cp-tree.h needs extensive housekeeping, there's some GC issues to address, etc.) This patch does not add specialized functions like decl_nreverse to be used on chains of decls. That would be an appropriate followup patch. I don't claim to have caught all places where this substitution needs to occur. My strategy for producing this patch was, roughly: - make TREE_CHAIN error when given a DECL; - replace erroring occurrences with DECL_CHAIN. (I'm not submitting the erroring bits at this time, though they would be a useful followup patch once, say, the C++ FE is cleaned up.) So it's certainly possible that there are some codepaths that still use TREE_CHAIN when they could be using DECL_CHAIN. This patch is a good start, though. The tree this patch is against is a couple days old. If the patch is approved, I'd like to rebase it against trunk with the above error checking enabled, re-bootstrap, fix errors, and commit. If the reviewer(s) don't like that idea, I'm happy to submit followup patches for smaller-scale DECL_CHAINification. Bootstrapped and tested on x86_64-unknown-linux-gnu with all languages and checking enabled, no regressions. Tested with cross to arm-eabi, powerpc-eabispe, and mips-elf with --enable-languages=c, no regressions. (Testing with powerpc and mips still ongoing, but no surprises expected.) OK to commit? -Nathan gcc/ * tree.h (DECL_CHAIN): Define. * alias.c: Carefully replace TREE_CHAIN with DECL_CHAIN. * c-decl.c: Likewise. * c-parser.c: Likewise. * c-typeck.c: Likewise. * cfgexpand.c: Likewise. * cgraph.c: Likewise. * cgraphunit.c: Likewise. * combine.c: Likewise. * config/alpha/alpha.c: Likewise. * config/arm/arm.c: Likewise. * config/frv/frv.c: Likewise. * config/i386/i386.c: Likewise. * config/i386/winnt-cxx.c: Likewise. * config/ia64/ia64.c: Likewise. * config/iq2000/iq2000.c: Likewise. * config/mep/mep.c: Likewise. * config/mips/mips.c: Likewise. * config/pa/som.h: Likewise. * config/rs6000/rs6000.c: Likewise. * config/s390/s390.c: Likewise. * config/sh/sh.c: Likewise. * config/sh/symbian-cxx.c: Likewise. * config/sparc/sparc.c: Likewise. * config/spu/spu.c: Likewise. * config/stormy16/stormy16.c: Likewise. * config/vxworks.c: Likewise. * config/xtensa/xtensa.c: Likewise. * coverage.c: Likewise. * dbxout.c: Likewise. * dwarf2out.c: Likewise. * emit-rtl.c: Likewise. * expr.c: Likewise. * function.c: Likewise. * gimple-low.c: Likewise. * gimple-pretty-print.c: Likewise. * gimplify.c: Likewise. * integrate.c: Likewise. * ipa-inline.c: Likewise. * ipa-prop.c: Likewise. * ipa-split.c: Likewise. * ipa-struct-reorg.c: Likewise. * ipa-type-escape.c: Likewise. * langhooks.c: Likewise. * lto-cgraph.c: Likewise. * omp-low.c: Likewise. * stor-layout.c: Likewise. * tree-cfg.c: Likewise. * tree-complex.c: Likewise. * tree-dfa.c: Likewise. * tree-dump.c: Likewise. * tree-inline.c: Likewise. * tree-mudflap.c: Likewise. * tree-nested.c: Likewise. * tree-object-size.c: Likewise. * tree-pretty-print.c: Likewise. * tree-sra.c: Likewise. * tree-ssa-live.c: Likewise. * tree-ssa-loop-niter.c: Likewise. * tree-ssa-math-opts.c: Likewise. * tree-ssa-reassoc.c: Likewise. * tree-ssa-sccvn.c: Likewise. * tree-ssa-structalias.c: Likewise. * tree-tailcall.c: Likewise. * tree-vrp.c: Likewise. * tree.c: Likewise. * var-tracking.c: Likewise. * varasm.c: Likewise. gcc/ada/ * gcc-interface/decl.c: Carefully replace TREE_CHAIN with DECL_CHAIN. * gcc-interface/trans.c: Likewise. * gcc-interface/utils.c: Likewise. * gcc-interface/utils2.c: Likewise. gcc/c-family/ * c-common.c: Carefully replace TREE_CHAIN with DECL_CHAIN. * c-format.c: Likewise. gcc/cp/ * cp-tree.h: Carefully replace TREE_CHAIN with DECL_CHAIN. * call.c: Likewise. * class.c: Likewise. * cp-gimplify.c: Likewise. * decl.c: Likewise. * decl2.c: Likewise. * init.c: Likewise. * mangle.c: Likewise. * name-lookup.c: Likewise. * optimize.c: Likewise. * parser.c: Likewise. * pt.c: Likewise. * rtti.c: Likewise. * search.c: Likewise. * semantics.c: Likewise. * typeck.c: Likewise. * typeck2.c: Likewise. gcc/fortran/ * f95-lang.c: Carefully replace TREE_CHAIN with DECL_CHAIN. * trans-common.c: Likewise. * trans-decl.c: Likewise. * trans-types.c: Likewise. * trans.c: Likewise. gcc/java/ * java-tree.h: Carefully replace TREE_CHAIN with DECL_CHAIN. * boehm.c: Likewise. * class.c: Likewise. * decl.c: Likewise. * expr.c: Likewise. * jcf-parse.c: Likewise. * typeck.c: Likewise. * verify-glue.c: Likewise. gcc/objc/ * objc-act.c: Carefully replace TREE_CHAIN with DECL_CHAIN. gcc/testsuite/ * g++.dg/plugin/attribute_plugin.c: Carefully replace TREE_CHAIN with DECL_CHAIN.
Attachment:
decl-chain.patch.gz
Description: Binary data
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |