Lexing of ACSL characters
ID0000714: This issue was created automatically from Mantis Issue 714. Further discussion may take place here.
| Id | Project | Category | View | Due Date | Updated |
|---|---|---|---|---|---|
| ID0000714 | Frama-C | Kernel > ACSL implementation | public | 2011-02-11 | 2014-02-12 |
| Reporter | patrick | Assigned To | patrick | Resolution | fixed |
| Priority | normal | Severity | minor | Reproducibility | have not tried |
| Platform | - | OS | - | OS Version | - |
| Product Version | - | Target Version | - | Fixed in Version | Frama-C Nitrogen-20111001 |
Description :
frama-c cannot parse the ACSL expression: '\' + 'a'
The lexer considers, as a unique token, the expression: '\' + '
Activity
Original note by patrick
I started to correct the problem. The actual tests are ok
The new following test is ok. //@ behavior esc: assumes c!= ''' && c!='a'; void f(char c) { }
But a problem is still there with the following expression when using -pp-annot option: //@behavior other: assumes c!='\' && c!='a'; void f(char c) { }
Note: I modified the lexing rule as follow: | ('L'? "'" as prelude) (([^ '\\' '\'' '\n']|("\"[^ '\n']))+ as content) "'"
-
Original note by patrick
Description of the problem with -pp-annot and '\' character :
cat t.c //@behavior b: assumes c!='\'; void f(char c) { }
frama-c t.c [kernel] preprocessing with "gcc -C -E -I. t.c"
frama-c -pp-annot t.c [kernel] preprocessing with "gcc -C -E -I. -dD t.c" "t.c":3:[kernel] user error: Can't preprocess annotation: eof while parsing a char literal Annotation will be kept as is t.c:1:[kernel] user error: syntax error [kernel] user error: skipping file "t.c" that has errors. [kernel] error occurring when exiting Frama-C: stopping exit procedure. Frama-C aborted because of invalid user input.