View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000570 | LDMud 3.3 | Implementation | public | 2008-09-10 13:54 | 2018-01-29 21:57 |
| Reporter | zesstra | Assigned To | zesstra | ||
| Priority | low | Severity | tweak | Reproducibility | always |
| Status | resolved | Resolution | fixed | ||
| Fixed in Version | 3.3.718 | ||||
| Summary | 0000570: fatal() fails to dump a core if the master is not loaded | ||||
| Description | fatal() calls shutdown() in the master and in that process assert_master_ob_loaded() is called. If there is no master object yet, exit(1) is called. In that way, a core dump is not available and to debug it, you have to reproduce with breakpoints at the right places in the debugger. I think, it would make more sense to call fatal() (again). fatal() will anyway detect that it was called recursively and directly dump a core by calling dump_core() without doing anything else which is the desired behaviour I guess. Or do you see a special reason why the driver should exit normally in this case? After all some part of the driver called fatal()... (In my case it was called from slaballoc.c because of a double free.) | ||||
| Tags | No tags attached. | ||||
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2008-09-10 13:54 | zesstra | New Issue | |
| 2008-12-14 16:35 | zesstra | Note Added: 0000819 | |
| 2008-12-14 16:35 | zesstra | Assigned To | => zesstra |
| 2008-12-14 16:35 | zesstra | Status | new => resolved |
| 2008-12-14 16:35 | zesstra | Resolution | open => fixed |
| 2008-12-14 16:35 | zesstra | Fixed in Version | => 3.3.718 |
| 2010-11-16 09:42 | zesstra | Source_changeset_attached | => ldmud.git master fefd0966 |
| 2018-01-29 18:59 | zesstra | Source_changeset_attached | => ldmud.git master fefd0966 |
| 2018-01-29 21:57 | zesstra | Source_changeset_attached | => ldmud.git master fefd0966 |
