| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 456.1 | Command line results same same | SMURF::GAF | Jerry Feldman, Unix Dev. Environment, DTN:381-2970 | Tue Feb 04 1997 14:04 | 4 |
| Just to add a bit more. Went to the command line, using just the /Za
switch on version 10.10.6079 and 11.10.6332.
Again, no warnings. I would expect a syntax error in ANSI mode.
|
| 456.2 | | DECCXL::OUELLETTE | | Wed Feb 05 1997 08:59 | 9 |
| I would expect a diagnostic and would much prefer a warning to an error.
With the C 1989 standard oddities like:
a = b //* a commment */ c;
could appear in C programs. (Although this is a strange thing to do.)
R.
|
| 456.3 | Internesting behavior in roland's example | SMURF::GAF | Jerry Feldman, Unix Dev. Environment, DTN:381-2970 | Thu Feb 13 1997 06:27 | 23 |
| Re .2
Roland retorts:
>I would expect a diagnostic and would much prefer a warning to an
>error.
In your example,
a = b //* a commment */ c;
A non-compliant program interpreting the // as a comment will receive
an error that the next line is missing a ";".
An ANSI compliant program should treat the /* comment as a real comment
and evaluate the expression as a = b / c;
VC++5.0 with the /Za switch evaluates "a = b / c;".
VC++5.0 without the /Za switch complains about the semicolon.
DECC with -std1 behaves as /Za, with -std0 and -std, behaves as VC++
without the /Za
In the expression, a = b // c;
VC++5.0 treats the // as a comment. I would have expected that an ansi
compliant compiler would complain that the // is an invalid expression
or something like that.
|
| 456.4 | | DECC::OUELLETTE | | Thu Feb 13 1997 11:51 | 1 |
| Credit for the interesting example goes to Rich Peterson.
|
| 456.5 | The culprit is exposed/ | SMURF::GAF | Jerry Feldman, Unix Dev. Environment, DTN:381-2970 | Mon Feb 17 1997 10:08 | 2 |
| RE:Credit for the interesting example goes to Rich Peterson.
Thanks for exposing the culprit.
|