0001026: aMuleD crash on a specific EC_OP_SEARCH_START packet

(0002190)
Kry (manager)
2006-12-14 15:53

That's not an unhandled problem, it's a handled problem, actually :P That's why the assert is there. It's not a crash.

(0002191)
quinox (reporter)
2006-12-14 23:02

OK, but aMule shouldn't go down because of it should it? Just ignoring the packet should be fine, perhaps with sending back an error

(0002192)
Kry (manager)
2006-12-16 16:54

don't compile on debug mode if you don't want debug behaviour?

(0002198)
quinox (reporter)
2006-12-17 19:31

Unless CVS code has some debug turned on by default, I'm not:

  **** aMule Core ****
  Prefix where aMule should be installed? /home/quinox/chroot/
  Should aMule be compiled with i18n support? yes
  Should aMule be compiled in debug mode? no
  Should aMule be compiled with profiling? no
  Should aMule be compiled with optimizations? no
  Should aMule monolithic application be built? yes
  Should aMule daemon version be built? yes
  Should aMule remote gui be built? (EXPERIMENTAL) yes

(0002200)
Kry (manager)
2006-12-17 21:05

You have wx compiled with debug.

(0002201)
quinox (reporter)
2006-12-18 16:54

Right, I'll look into that - I guess the problem isn't with aMule then, feel free to close this bug :)

(0002202)
Kry (manager)
2006-12-18 21:49

Nah... I'm leaving this around. Maybe some logging is betetr than an assert, but then again some hard debug is always welcome... I'll think about it.

(0002648)
Xaignar (manager)
2008-02-22 03:21

This situation has been improved in current SVN, where wxASSERT is now used instead of assert in EC code, allowing for aMule to continue running after an asserted has been triggered.

Notes
(0002190) Kry (manager) 2006-12-14 15:53	That's not an unhandled problem, it's a handled problem, actually :P That's why the assert is there. It's not a crash.

(0002191) quinox (reporter) 2006-12-14 23:02	OK, but aMule shouldn't go down because of it should it? Just ignoring the packet should be fine, perhaps with sending back an error

(0002192) Kry (manager) 2006-12-16 16:54	don't compile on debug mode if you don't want debug behaviour?

(0002198) quinox (reporter) 2006-12-17 19:31	Unless CVS code has some debug turned on by default, I'm not: ** aMule Core ** Prefix where aMule should be installed? /home/quinox/chroot/ Should aMule be compiled with i18n support? yes Should aMule be compiled in debug mode? no Should aMule be compiled with profiling? no Should aMule be compiled with optimizations? no Should aMule monolithic application be built? yes Should aMule daemon version be built? yes Should aMule remote gui be built? (EXPERIMENTAL) yes

(0002200) Kry (manager) 2006-12-17 21:05	You have wx compiled with debug.

(0002201) quinox (reporter) 2006-12-18 16:54	Right, I'll look into that - I guess the problem isn't with aMule then, feel free to close this bug :)

(0002202) Kry (manager) 2006-12-18 21:49	Nah... I'm leaving this around. Maybe some logging is betetr than an assert, but then again some hard debug is always welcome... I'll think about it.

(0002648) Xaignar (manager) 2008-02-22 03:21	This situation has been improved in current SVN, where wxASSERT is now used instead of assert in EC code, allowing for aMule to continue running after an asserted has been triggered.

Issue History
Date Modified	Username	Field	Change
2006-12-13 21:27	quinox	New Issue
2006-12-13 21:27	quinox	Operating System	=> Any
2006-12-14 15:53	Kry	Note Added: 0002190
2006-12-14 23:02	quinox	Note Added: 0002191
2006-12-16 16:54	Kry	Note Added: 0002192
2006-12-17 19:31	quinox	Note Added: 0002198
2006-12-17 21:05	Kry	Note Added: 0002200
2006-12-18 16:54	quinox	Note Added: 0002201
2006-12-18 21:49	Kry	Note Added: 0002202
2008-02-22 03:21	Xaignar	Status	new => resolved
2008-02-22 03:21	Xaignar	Fixed in Version	=> SVN
2008-02-22 03:21	Xaignar	Resolution	open => fixed
2008-02-22 03:21	Xaignar	Assigned To	=> Xaignar
2008-02-22 03:21	Xaignar	Note Added: 0002648

Relationships