Sergeo7

But on the other hand you can't temporary save big file on 'almost full' disk even if physically downloaded just 10% of that 60Gb. So if download isn't fast preallocated space will be a waste of resources needed just to ensure that you'll have required space in the end (as long as target and temporary dirs on the same logical drive).

What kind of problem do you mean? Is it psychological, like when user see file already occupy "full size", but real space still decreases? For test I had added this function to build and see no technical problems at all. File class just throw FileException when disk is full, but when I freed some space download normally resumed.

2 developers: Please, add some more popular temp's to predefined settings in future releases PS: (Util::stricmp(fileExt.c_str(), ".jc!") == 0) || //FlashGet (Util::stricmp(fileExt.c_str(), ".dmf") == 0) || //Download Master (Util::stricmp(fileExt.c_str(), ".!ut") == 0) || //uTorrent (Util::stricmp(fileExt.c_str(), ".bc!") == 0) || //BitComet (name.find("__padding_") == 0) || //BitComet padding

Maybe someone find this solution useful..? If you don't need it, doesn't mean that no one need it. Some of my friends merge folders from several harddrives with JPs and share it as one. So I think it's not so difficult. It's impossible to implement every single idea just because of limited time :P

To merge folders into one you can use NTFS junction points on NT-based Windows or Symbolic links on Unix / Vista. There are no real need in such feauture within client.

As I see an appropriate API would be needed minimum for: a) External control of ApexDC application (start/stop downloads/and so forth) Custom data storage (for example to replace xml containers with some sort of sql database) c) For a script languages (like LUA and so on; might overlap other needs) d) Chat/PM processing plugins (for example local abuse filter or chatbot)

In principle ADC protocol supports nick change without reconnection.

Hi. In current version of apexdc all external search commands performed throught hardcoded blocks like this: if(Util::stricmp(cmd.c_str(), _T("g")) == 0 || Util::stricmp(cmd.c_str(), _T("google")) == 0) { if(param.empty()) { status = TSTRING(SPECIFY_SEARCH_STRING); } else { WinUtil::openLink(_T("http://www.google.com/search?q=") + Text::toT(Util::encodeURI(Text::fromT(param)))); } } Maybe it will be a good idea to make all such commands loaded from external Commands.xml settings file? (for example with tags of command - hyperlink - help string, where hyperlink contains search string substitution, like http://www.site.com/?q=%s) PS: And also same idea may be applied to file type - extensions correspondence list for search window.

When debugging I found that this (a bit modified from original, to know whats going on) code: CComPtr<IEnumFilters> pEF; HRESULT hr = pFG->EnumFilters(&pEF);returns error hr = 0x80070725 AKA Incompatible version of the RPC stub - "Unhandled Exception." or RPC_X_WRONG_STUB_VERSION

Try to recompile ApexDC on machine with slovak locale and arrayutf copied from StrongDC source, maybe it'll help. PS: Personally I think this kinda way would be better when removing diacritic marks: WinUtil.cpp wchar_t arraydblutf[12] = { L'\xe4' ,L'\xf1' ,L'\xf6' ,L'\xfc' ,L'\xff' ,L'\x00df' ,L'\x00c6' ,L'\x00e6' ,L'\x0132' ,L'\x0133' ,L'\x0152' ,L'\x0153' }; vector<tstring> arraydblwin(12); arraydblwin.push_back(L"ae"); arraydblwin.push_back(L"ny"); arraydblwin.push_back(L"oe"); arraydblwin.push_back(L"ue"); arraydblwin.push_back(L"yu"); arraydblwin.push_back(L"ss"); arraydblwin.push_back(L"AE"); arraydblwin.push_back(L"ae"); arraydblwin.push_back(L"IJ"); arraydblwin.push_back(L"ij"); arraydblwin.push_back(L"Oe"); arraydblwin.push_back(L"oe"); wchar_t arrayutf[22] = { L'\x00d0' ,L'\x0110' ,L'\x00f0' ,L'\x0111' ,L'\x0126' ,L'\x0127' ,L'\x0131' ,L'\x0138' ,L'\x013f' ,L'\x0141' ,L'\x0140' ,L'\x0142' ,L'\x014a' ,L'\x0149' ,L'\x014b' ,L'\x00d8' ,L'\x00f8' ,L'\x017f' ,L'\x00de' ,L'\x0166' ,L'\x00fe' ,L'\x0167 }; wchar_t arraywin[22] = { L'D' ,L'D' ,L'd' ,L'd' ,L'H' ,L'h' ,L'i' ,L'k' ,L'L' ,L'L' ,L'l' ,L'l' ,L'N' ,L'n' ,L'n' ,L'O' ,L'o' ,L's' ,L'T' ,L'T' ,L't' ,L't }; const tstring& WinUtil::disableCzChars(tstring& message) { //Here we must check if locale is really contain removed diacritic (maybe switch(locale) {... //removal... default: return message; } would suit) //Then array of removed chars must vary for each locale. //... //Let's remove them for(size_t j = 0; j < message.length(); j++) { for(size_t l = 0; l < (sizeof(arrayutf) / sizeof(arrayutf[0])); l++) { if (message[j] == arrayutf[l]) { message[j] = arraywin[l]; break; } } for(size_t l = 0; l < arraydblwin.size(); l++) { if (message[j] == arraydblutf[l]) { message.erase(j,1); message.insert(j,arraydblwin[l]); j++; break; } } } return message; }

No, there is no such thing. All just as I said earlier. I think something like this could resolve problem (just disabling removal of diacritic marks in options is really lame solution): text.h extern const string utf8; extern string systemCharset; extern bool isCzechLocale; text.cpp const string utf8 = "utf-8"; // optimization string systemCharset; bool isCzechLocale = false; void initialize() { setlocale(LC_ALL, ""); #ifdef _WIN32 char *ctype = setlocale(LC_CTYPE, NULL); if(ctype) { systemCharset = string(ctype); } else { dcdebug("Unable to determine the program's locale"); } #else systemCharset = string(nl_langinfo(CODESET)); #endif if(systemCharset.find("Czech") != string::npos) //todo: all other locales with diacritic isCzechLocale = true; } and WinUtil.cpp const tstring& WinUtil::disableCzChars(tstring& message) { if (!Text::isCzechLocale) return message; for(size_t j = 0; j < message.length(); j++) { for(size_t l = 0; l < (sizeof(arrayutf) / sizeof(arrayutf[0])); l++) { if (message[j] == arrayutf[l]) { message[j] = arraywin[l]; break; } } } return message; }

The ApexDC++ 1.0.0 Beta4 compilation is going well (I am using STLPort 5.2 and WTL 8 fixed used on VS2005 SP1), but after I run it and type text there is a bug when it is showing in chat window: instead of some cyrillic characters it is showing some other irrelevant chars. Can anyone help to solve this problem?

Eh.. no one wish to answer so I had to find the bug by myself: Function disableCzChars really don't care about current locale so if it is not czech it'll "randomly" destroy non-english chars. And somehow the option to use it was enabled when I run new build with old settings.

Actually, the most new builds are at other page. They don't really bother to update it in original project's (don't know why). PS: And there is one more problem with the MPC spam, I forget to mention before - it isn't showing local non-english codepage letters, just an english ones (the codepage has been set up properly and everywhere else they displays correctly).