X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-bounces using -f From: "'Raf256' Rafal Maj" Newsgroups: comp.os.msdos.djgpp Subject: let's debug together ;) Date: 27 Jan 2002 11:13:36 GMT Organization: aa Lines: 67 Sender: raf256 AT poczta DOT onet DOT pl@chello062179025129.chello.pl Message-ID: References: NNTP-Posting-Host: chello062179025129.chello.pl Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-2 Content-Transfer-Encoding: 8bit X-Trace: news.onet.pl 1012130016 11421 62.179.25.129 (27 Jan 2002 11:13:36 GMT) X-Complaints-To: abuse AT onet DOT pl NNTP-Posting-Date: 27 Jan 2002 11:13:36 GMT User-Agent: Xnews/4.11.09 To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com Eli Zaretskii wrote in news:Pine DOT SUN DOT 3 DOT 91 DOT 1020127105615 DOT 14549M-100000 AT is: i'm wondering - maybe it's not so bad idea to try debug something together :) ? Imho, this will be interesting... >> char * p = new char[l]; >> eax=0000000a ebx=746f6f72 ecx=000baf58 edx=00000002 esi=00000054 > in the EBX register. The register dump above shows that the value of > EBX looks like ASCII text ("root", written right to left, because x86 > is a little endian machine). Does that "root" string ring a bell? Is it should be easy, program crashes every time in exacly same place : My program crashes in : -------------------------- Call frame traceback: c:/djgpp/lang/cxx-v3/bits/locale_facets.tcc(75) in function $(RHIDE_INCLUDES) $(C_DEBUG_FLA c:/djgpp/lang/cxx-v3/bits/locale_facets.tcc(75) in function $(RHIDE_INCLUDES) $(C_DEBUG_FLA c:/djgpp/lang/cxx-v3/bits/locale_facets.tcc(75) in function $(RHIDE_INCLUDES) $(C_DEBUG_FLA ustr.cpp(66) in function _ZN4cStr4_IncEPKc ustr.h(37) in function _ZN4cStrpLIcEERS_T_ 0acmd.cpp(736) in function _ZN10cCommander5ParseE4cStr 0acmd.cpp(684) in function _ZN10cCommander10RunCommandE4cStrb 0acmd.cpp(781) in function _ZN10cCommander3RunEv 0acmd.cpp(818) in function _ZN11cUptApplCmd3RunEv shell.cpp(74) in function _ZN9cUptShell10StartShellEv main.cpp(30) in function main c:/djgpp/lang/cxx-v3/bits/locale_facets.tcc(75) in function $(RHIDE_INCLUDES) $(C_DEBUG_FLA -------------------------- cStr is my cals simmilar to std::string, but my own :) It's possible that is still contains bugs... let's debug it : Ok, string allocated in t=0x1483c4 cStr::_Put now allocating 12 bytes for string, oryginla string is 'run auto.ub'... Ok, string allocated in t=0x148434 Ok, string allocated in t=0x14844c Ok, string allocated in t=0x148464 Ok, string allocated in t=0x1484ec Ok, string allocated in t=0x14849c Ok, string allocated in t=0x14850c Ok, string allocated in t=0x148404 etc... this is sample of normal working, each string (when i.e. copying cStr, etc) is allocated in diffrent memory region. In last lines before crash, program often is allocating strings twice in same memory area... Hmm I'll add some debugging code and post results here. Thanks for Help ! :) -- Raf256 ,--, _ ,- _ _ _ |,-'`_|_|_ `_||_'|_' raf256 AT wp DOT pl GG-1175498 ICQ-143712878 | `\|_| | |_,,_||_| C++; (_2b||!_2b)?this->is->a:question; {B.Stroustrup, W.Shakespeare}