On Thu, Jun 15, 2017 at 1:20 AM, Luc Van Oostenryck
<luc.vanoostenryck@xxxxxxxxx> wrote:
>
> Please, drop what you have already pulled and take the
> following instead.
Here is the feed back for "finer control over error vs. warnings"
+#define ERROR_CURR_PHASE (1 << 0)
+#define ERROR_PREV_PHASE (1 << 1)
+extern int has_error;
The current phase vs previous phase define is a bit messy to
maintain when you advance to next phase. Currently your
patch does:
+
+ if (has_error & ERROR_CURR_PHASE)
+ has_error = ERROR_PREV_PHASE;
I purpose each different phase/stage we have a dedicate bits.
#define PHASE_PARSING (1 << 0)
#define PHASE_EVALUATE (1<<1)
Have one variable to track which phase we are currently in,
initialized to parsing.
int current_phase = PHASE_PARSING;
When move to evaluation:
current_phase = PHASE_EVALUATE;
The following code:
- max_warnings = 0;
+ has_error |= ERROR_CURR_PHASE;
will change to:
has_error |= current_phase;
Then you don't need to move bits between different phase:
+
+ if (has_error & ERROR_CURR_PHASE)
+ has_error = ERROR_PREV_PHASE;
That will be gone. The rest of the patch should be very similar.
This has the advantage of:
- The code clearly show which stage it is in.
- No need to move error bits from current phase to previous phase.
- more friendly if we have more than two phase.
Chris
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
