View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000617 | LDMud 3.3 | Compilation, Installation | public | 2009-03-31 05:53 | 2009-05-05 15:50 |
| Reporter | _xtian_ | Assigned To | |||
| Priority | normal | Severity | feature | Reproducibility | N/A |
| Status | new | Resolution | open | ||
| Summary | 0000617: relax syntax and warn about obvious coding errors? | ||||
| Description | Since most of the MUD coders are not necessary experienced programmers, why not modify the compiler to catch some obvious mistakes. These could be issued as warnings or even compile-errors. I just stumbled over: if (...); There is no useful case for this statement, even if it is syntactically correct. There surely are a lot of other items like this one. What is the general opinion here? Should we adhere to syntactic rules or enhance useability? | ||||
| Tags | No tags attached. | ||||
|
|
Just a few hours ago I stumbled upon the same. But in a if-elseif-else chain it's only useless if it is the last statement that is empty. Also a foreach loop with an empty statement is useless. |
|
|
Mhmm, isn't that actually making the syntax stricter? Ok, maybe not the syntax itself but I mean, we would make syntactically correct code to raise errors/warning upon compilation. That is not necessarily a bad thing, I appreciate if errors can be caught at compile-time, but I think the title of this issue is confusing in this context. |
|
|
Yes, the title is weird. I was thinking around the lines of "relaxing the necessity to adhere to syntactic correcteness" ... well, anyhow ... Having given it some thought I would suggest to issue warnings for these cases and if #pragma pedantic is set - errors. Also pragmas #warn/no_warn_common_mistakes to switch them on/off. If the user can switch them off/on on a per-file basis, we can also think about issueing warnings for common mistake cases that could even have real uses, like: probably: for(...;...;...); and *maybe* while(...); |
|
|
Hmm, I'd be more likely to have an empty body in a for loop than in a while loop. Anyway, as long as these warnings can be switched off, I'm happy. We can also think about adding a keyword or pseudo efun for intentionally empty statements. I.e. for(;;) ; would give a warning while for(;;) nop; (or whatever) would not. |
|
|
There is a difference between statements that are guaranteed to be useless ('if(expr);' is always the same as 'expr;' as long as no 'else' is following) and statements that might be a mistake. I wouldn't lump them together. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2009-03-31 05:53 | _xtian_ | New Issue | |
| 2009-03-31 06:01 | Gnomi | Note Added: 0001009 | |
| 2009-04-08 04:36 | zesstra | Note Added: 0001012 | |
| 2009-04-15 22:51 | _xtian_ | Note Added: 0001041 | |
| 2009-04-16 11:39 | fufu | Note Added: 0001043 | |
| 2009-04-16 12:00 | Gnomi | Note Added: 0001044 | |
| 2009-05-05 15:50 | zesstra | Project | LDMud => LDMud 3.3 |
