View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0002801 | unreal | ircd | public | 2006-02-06 00:52 | 2007-04-27 03:46 |
| Reporter | Xuefer | Assigned To | |||
| Priority | normal | Severity | feature | Reproducibility | always |
| Status | closed | Resolution | wont fix | ||
| Product Version | 3.2.4 | ||||
| Summary | 0002801: feature request for clean valgrind output on shutdown(s_die) | ||||
| Description | it looks like most module is shutdown cleanly, but from the output of valgrind, there's many blocks not free'ed in ircd itself(not module) after SIGTERM/die (s_die) imho, this makes valgrind output confusing. an IsupportDel even malloc blocks and make valgrind think it's the module not shuting down correctly. | ||||
| 3rd party modules | |||||
|
|
I wasn't planning to make it free all kinds of stuff, actually. Too much work. IMO the 'definately lost' / 'possibly lost' stuff is working ok and enough. I did see your other Isupport thing bugreport though :P |
|
|
the Isupport problem here isn't the one i report in another bug the "malloc blocks" is inside make_isupportstrings: if (tokcnt == 13 || bufsize < len+toklen+1) { tokcnt = 0; len = 0; IsupportStrings[++i] = MyMallocEx(bufsize); } there's also same problem if 3rd'party/homemade module invoke m_topic m_nick directly. well, the pain may be easy while more and more code is made into a clean module :) thanks for your notice. |
