Fix handling of ident-less declarations
PATCH 1/7 Fix handling of ident-less declarations
The rule for ident-less declaration is
declaration -> declaration-specifiers ; not declaration -> declaration-specifiers abstract-declarator;
IOW, struct foo; is OK and so's struct foo {int x; int y;} (and even simply int; is allowed by syntax - it's rejected by constraints, but that's a separate story), but not struct foo (void); and its ilk.
See C99 6.7p1 for syntax; C90 is the same in that area and gcc also behaves the same way. Unlike gcc I've made it a warning (gcc produces a hard error for those).
--- parse.c | 10 +++++++++- validation/missing-ident.c | 18 ++++++++++++++++++ 2 files changed, 27 insertions(+), 1 deletions(-) create mode 100644 validation/missing-ident.c
parse.c +9 -1
@@ -2265,14 +2265,22 @@ struct token *external_declaration(struct token *token, struct symbol_list **lis token =declaration_specifiers (token, &ctype,0 ); decl =alloc_symbol (token->pos,SYM_NODE ); decl->ctype = ctype; +/* Just a type declaration? */ +if (match_op (token,';' )) { +apply_modifiers (token->pos, &decl->ctype); +return token->next; + } + token =declarator (token, decl, &ident,0 );apply_modifiers (token->pos, &decl->ctype); decl->endpos = token->pos;/* Just a type declaration? */ -if (!ident) +if (!ident) { +warning (token->pos,"missing identifier in declaration" );return expect (token,';' ,"end of type declaration" ); + }/* type define declaration? */ is_typedef = (ctype.modifiers &MOD_TYPEDEF ) !=0 ;
validation/missing-ident.c new file +18 -0
@@ -0,0 +1,18 @@ +int [2]; +int *; +int (*); +int (); +int ; +struct foo ; +union bar {int x;int y;}; +struct baz {int x, :3, y:2;}; +/* +* check-name: handling of identifier-less declarations +* +* check-error-start +missing-ident.c:1:8: warning: missing identifier in declaration +missing-ident.c:2:6: warning: missing identifier in declaration +missing-ident.c:3:8: warning: missing identifier in declaration +missing-ident.c:4:7: warning: missing identifier in declaration +* check-error-end +*/
Signature · 1 line
1.5.6.6
Christopher Li
maintainerQuoted 2 lines · Al Viro
The rule for ident-less declaration is
wow, great.
The series applies to the tree fine. First impression of the series looks fine. I am still working on it.
Thanks for the patches. BTW, this is not the "lazy type evaluation" you are talking about right?
Chris
Quoted 80 lines
declaration -> declaration-specifiers ; not declaration -> declaration-specifiers abstract-declarator;IOW, struct foo; is OK and so's struct foo {int x; int y;} (and even simply int; is allowed by syntax - it's rejected by constraints, but that's a separate story), but not struct foo (void); and its ilk.
See C99 6.7p1 for syntax; C90 is the same in that area and gcc also behaves the same way. Unlike gcc I've made it a warning (gcc produces a hard error for those).
Signed-off-by Al Viro <[email protected]>--- parse.c | 10 +++++++++- validation/missing-ident.c | 18 ++++++++++++++++++ 2 files changed, 27 insertions(+), 1 deletions(-) create mode 100644 validation/missing-ident.cparse.c +9
-1@@ -2265,14 +2265,22 @@ struct token *external_declaration(struct token *token, struct symbol_list **lis token =declaration_specifiers (token, &ctype,0 ); decl =alloc_symbol (token->pos,SYM_NODE ); decl->ctype = ctype; +/* Just a type declaration? */ +if (match_op (token,';' )) { +apply_modifiers (token->pos, &decl->ctype); +return token->next; + } + token =declarator (token, decl, &ident,0 );apply_modifiers (token->pos, &decl->ctype); decl->endpos = token->pos;/* Just a type declaration? */ -if (!ident) +if (!ident) { +warning (token->pos,"missing identifier in declaration" );return expect (token,';' ,"end of type declaration" ); + }/* type define declaration? */ is_typedef = (ctype.modifiers &MOD_TYPEDEF ) !=0 ;validation/missing-ident.c new file +18
-1@@ -0,0 +1,18 @@ +int [2]; +int *; +int (*); +int (); +int ; +struct foo ; +union bar {int x;int y;}; +struct baz {int x, :3, y:2;}; +/* +* check-name: handling of identifier-less declarations +* +* check-error-start +missing-ident.c:1:8: warning: missing identifier in declaration +missing-ident.c:2:6: warning: missing identifier in declaration +missing-ident.c:3:8: warning: missing identifier in declaration +missing-ident.c:4:7: warning: missing identifier in declaration +* check-error-end +*/ --1.5.6.6
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 11 lines · Christopher Li
Quoted 2 lines · Al Viro
The rule for ident-less declaration is
wow, great.
The series applies to the tree fine. First impression of the series looks fine. I am still working on it.
Thanks for the patches. BTW, this is not the "lazy type evaluation" you are talking about right?
No, just the first batch of declaration parser fixes... The next group is about separating ctype-as-part-of-symbol from ctype-as-parser-state uses. After that - getting declaration_specifiers() to sane shape, which, BTW, will relieve the situation with mode bits. Then - cleaning up the type handling...
Quoted 10 lines · Al Viro
Quoted 9 lines · Christopher Li
Quoted 2 lines · Al Viro
The rule for ident-less declaration is
wow, great.
The series applies to the tree fine. First impression of the series looks fine. I am still working on it.
Quoted 8 lines
Quoted 2 lines
Thanks for the patches. BTW, this is not the "lazy type evaluation" you are talking about right?
No, just the first batch of declaration parser fixes... The next group is about separating ctype-as-part-of-symbol from ctype-as-parser-state uses. After that - getting
declaration_specifiers()to sane shape, which, BTW, will relieve the situation with mode bits. Then - cleaning up the type handling...
BTW, I wonder whether if it would be better to just scan for the end of attributes after ( when we are deciding whether it's a function or nested declarator, leaving the actual handling of these guys to after the decision. The thing is, unlike gcc we have the token list anyway, and it's easier to not bother with passing that crap to parameter_type_list(), etc.
I'll try to do that and see what falls out; potentially that replaces 7/7 in this series.
Quoted 27 lines · Al Viro
Quoted 10 lines · Al Viro
Quoted 9 lines · Christopher Li
Quoted 2 lines · Al Viro
The rule for ident-less declaration is
wow, great.
The series applies to the tree fine. First impression of the series looks fine. I am still working on it.
Quoted 8 lines
Quoted 2 lines
Thanks for the patches. BTW, this is not the "lazy type evaluation" you are talking about right?
No, just the first batch of declaration parser fixes... The next group is about separating ctype-as-part-of-symbol from ctype-as-parser-state uses. After that - getting
declaration_specifiers()to sane shape, which, BTW, will relieve the situation with mode bits. Then - cleaning up the type handling...BTW, I wonder whether if it would be better to just scan for the end of attributes after ( when we are deciding whether it's a function or nested declarator, leaving the actual handling of these guys to after the decision. The thing is, unlike gcc we have the token list anyway, and it's easier to not bother with passing that crap to
parameter_type_list(), etc.
Quoted 2 lines
I'll try to do that and see what falls out; potentially that replaces 7/7 in this series.
... no, it doesn't ;-/ Too much PITA with error recovery, so the original version still stands. Alas.
A message that is not in the archive (sent off-list or not yet received): [email protected]
A message that is not in the archive (sent off-list or not yet received): [email protected]
Christopher Li
maintainerQuoted 9 lines · Al Viro
Quoted 2 lines · Christopher Li
All applied and pushed. I update some of the validations error because of the tab width change.
BTW, I'm not sure that this messing with tabstops is a good idea. Note that tokenizer is _still_ the hottest part of the entire thing, so we need to be damn careful around it. You are getting the slow path of
nextchar()on \t now, which can add up to a lot of extra overhead...
How about the attached patch?
I move white space handling out side of the nextchar_slow(). Let nextchar_slow() only handle two case: possible EOF and '\\'.
Chris