Yesterday

• 26 hours ago

  Manual (v8cgi User manual) Wiki page commented on by preston.bannister   -   Note that when compiled for FastCGI (with the fcgi=1 option to scons), there is no command line help for options.

  Note that when compiled for FastCGI (with the fcgi=1 option to scons), there is no command line help for options.
• 37 hours ago

  issue 37 (Can't build on MacOSX) commented on by mr.veged   -   at r696 i success compiled with this diff of SConstruct Index: SConstruct =================================================================== --- SConstruct (revision 696) +++ SConstruct (working copy) @@ -136,6 +136,8 @@ if env["os"] == "darwin": env.Append( + CCFLAGS = ["-m32"], + LINKFLAGS = ["-m32"], SHLINKFLAGS = "-undefined dynamic_lookup", CPPDEFINES = ["DSO_EXT=so"] )

  at r696 i success compiled with this diff of SConstruct Index: SConstruct =================================================================== --- SConstruct (revision 696) +++ SConstruct (working copy) @@ -136,6 +136,8 @@ if env["os"] == "darwin": env.Append( + CCFLAGS = ["-m32"], + LINKFLAGS = ["-m32"], SHLINKFLAGS = "-undefined dynamic_lookup", CPPDEFINES = ["DSO_EXT=so"] )
• 42 hours ago

  issue 51 (MySQL:fetchObjects() crashes on XP running select query on m...) commented on by habibur   -   Found the problem. I was using libmysql.dll from my version 5.0. Just downloaded mysql community release 5.1 from mysql site and copied libmysql.dll into v8's folder and it is working. There the fix is to use libmysql.dll from v 5.1. I shall request v8cgi windows build to distribute libmysql.dll with the zip, just like how PHP does. The program doesn't run otherwise.

  Found the problem. I was using libmysql.dll from my version 5.0. Just downloaded mysql community release 5.1 from mysql site and copied libmysql.dll into v8's folder and it is working. There the fix is to use libmysql.dll from v 5.1. I shall request v8cgi windows build to distribute libmysql.dll with the zip, just like how PHP does. The program doesn't run otherwise.
• 43 hours ago

  issue 51 (MySQL:fetchObjects() crashes on XP running select query on m...) commented on by habibur   -   The problem seems to be related with fetchNames() too, which crashes. I guess fetchObjects internally calls fetchNames. If no one else if looking at it I shall try to fix it myself.

  The problem seems to be related with fetchNames() too, which crashes. I guess fetchObjects internally calls fetchNames. If no one else if looking at it I shall try to fix it myself.

Last 7 days

• Dec 28, 2009

  InstallOnUbuntu9 (Installation instructions for Ubuntu 9.10 32-bit server.) Wiki page commented on by preston.bannister   -   To the v8 build if building for 64-bit change the build command to: {{{ scons arch=x64 library=shared }}} (As you note in the v8cgi INSTALL file.)

  To the v8 build if building for 64-bit change the build command to: {{{ scons arch=x64 library=shared }}} (As you note in the v8cgi INSTALL file.)
• Dec 27, 2009

  issue 51 (MySQL:fetchObjects() crashes on XP running select query on m...) commented on by habibur   -   The os shows crash window, "This application has performed an illigal operation." Running the debugger shows the memory location and instruction set, which I have not noted down...but was trying to load a value from invalid memory address.

  The os shows crash window, "This application has performed an illigal operation." Running the debugger shows the memory location and instruction set, which I have not noted down...but was trying to load a value from invalid memory address.
• Dec 27, 2009

  issue 51 (MySQL:fetchObjects() crashes on XP running select query on m...) commented on by ondrej.zara   -   This certainly does not happen with Linux builds. I will try on Windows soon. By the way, what "crashes" actually means?

  This certainly does not happen with Linux builds. I will try on Windows soon. By the way, what "crashes" actually means?
• Dec 27, 2009

  issue 51 (MySQL:fetchObjects() crashes on XP running select query on m...) reported by habibur   -   What steps will reproduce the problem? mysql> drop table if exists t1; create table t1 (f1 int,f2 int,f3 int,f4 int,f5 int,f6 int, f7 int, f8 int); insert into t1 values (1,2,3,4,5,6,7,8); test.js var MySQL = require("mysql").MySQL var db = new MySQL().connect("localhost", "root", "", "test") var result = db.query("select * from t1") var rows=result.fetchObjects() I am running from v8cgi-0.7.5-win32 -habib.

  What steps will reproduce the problem? mysql> drop table if exists t1; create table t1 (f1 int,f2 int,f3 int,f4 int,f5 int,f6 int, f7 int, f8 int); insert into t1 values (1,2,3,4,5,6,7,8); test.js var MySQL = require("mysql").MySQL var db = new MySQL().connect("localhost", "root", "", "test") var result = db.query("select * from t1") var rows=result.fetchObjects() I am running from v8cgi-0.7.5-win32 -habib.
• Dec 27, 2009

  API_MySQL (Module "mysql") Wiki page commented on by habibur   -   > "Error opening shared library 'd:\Server\usr\local\apache\v8cgi\mysql.dll'" You need to keep libmysql.dll on your path. Note that %path% variable is different from within a webserver call and cmd prompt on the same machine. First try copyting libmysql.dll to the same folder where v8cgi.exe resides. You can find libmysql.dll in mysql\bin folder.

  > "Error opening shared library 'd:\Server\usr\local\apache\v8cgi\mysql.dll'" You need to keep libmysql.dll on your path. Note that %path% variable is different from within a webserver call and cmd prompt on the same machine. First try copyting libmysql.dll to the same folder where v8cgi.exe resides. You can find libmysql.dll in mysql\bin folder.

Last 30 days

• Dec 21, 2009

  issue 49 (require('process') dont work on MacOSX because module file i...) commented on by mr.veged   -   i don't know, some very old, because i can't build http://code.google.com/p/v8cgi/issues/detail?id=37

  i don't know, some very old, because i can't build http://code.google.com/p/v8cgi/issues/detail?id=37
• Dec 21, 2009

  Manual (v8cgi User manual) Wiki page edited by ondrej.zara   -   Revision r696 Edited wiki page through web user interface.

  Revision r696 Edited wiki page through web user interface.
• Dec 21, 2009

  issue 50 (One require() per module) Status changed by ondrej.zara   -   Commited in r695.

  Status: Fixed

  Commited in r695.

  Status: Fixed
• Dec 21, 2009

  r695 (#50 - individual require for each module) committed by ondrej.zara   -   #50 - individual require for each module

  #50 - individual require for each module
• Dec 20, 2009

  issue 50 (One require() per module) reported by ondrej.zara   -   Every module needs its own require.

  Every module needs its own require.
• Dec 20, 2009

  issue 49 (require('process') dont work on MacOSX because module file i...) commented on by ondrej.zara   -   Current SConstruct file defines the process library to build as .so ... what SVN revision are you experimenting with?

  Current SConstruct file defines the process library to build as .so ... what SVN revision are you experimenting with?
• Dec 20, 2009

  issue 49 (require('process') dont work on MacOSX because module file i...) reported by mr.veged   -   That seems *.dylib need be added to modules lookup.

  That seems *.dylib need be added to modules lookup.
• Dec 18, 2009

  issue 48 (Building fcgi=1 on centos) Status changed by ondrej.zara   -   Fixed in r694.

  Status: Fixed

  Fixed in r694.

  Status: Fixed
• Dec 18, 2009

  r694 (#48 fcgi bugfix) committed by ondrej.zara   -   #48 fcgi bugfix

  #48 fcgi bugfix
• Dec 17, 2009

  issue 48 (Building fcgi=1 on centos) commented on by kraklog   -   I get it to compile when applying the single thing I know about fastcgi and that is that fcgi_stio.h overwrites the definititons in stdio.h and thus should be included last. I moved: #ifdef FASTCGI # include <fcgi_stdio.h> # include <signal.h> #endif

  I get it to compile when applying the single thing I know about fastcgi and that is that fcgi_stio.h overwrites the definititons in stdio.h and thus should be included last. I moved: #ifdef FASTCGI # include <fcgi_stdio.h> # include <signal.h> #endif
• Dec 16, 2009

  issue 48 (Building fcgi=1 on centos) commented on by ondrej.zara   -   Unfortunately, I have no access to CentOS to debug this. Are you able to provide a patch here?

  Unfortunately, I have no access to CentOS to debug this. Are you able to provide a patch here?
• Dec 16, 2009

  issue 48 (Building fcgi=1 on centos) reported by kraklog   -   Building with fcgi=1 on centos gives these errors: src/v8cgi.cc: In member function 'virtual size_t v8cgi_CGI::reader(char*, size_t)': src/v8cgi.cc:61: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fread(void*, size_t, size_t, FILE*)' src/v8cgi.cc: In member function 'virtual size_t v8cgi_CGI::writer(const char*, size_t)': src/v8cgi.cc:68: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' src/v8cgi.cc: In member function 'virtual void v8cgi_CGI::error(const char*, const char*, int)': src/v8cgi.cc:75: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' src/v8cgi.cc:76: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' scons: *** [src/v8cgi.o] Error 1

  Building with fcgi=1 on centos gives these errors: src/v8cgi.cc: In member function 'virtual size_t v8cgi_CGI::reader(char*, size_t)': src/v8cgi.cc:61: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fread(void*, size_t, size_t, FILE*)' src/v8cgi.cc: In member function 'virtual size_t v8cgi_CGI::writer(const char*, size_t)': src/v8cgi.cc:68: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' src/v8cgi.cc: In member function 'virtual void v8cgi_CGI::error(const char*, const char*, int)': src/v8cgi.cc:75: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' src/v8cgi.cc:76: error: cannot convert 'FCGI_FILE*' to 'FILE*' for argument '4' to 'size_t fwrite(const void*, size_t, size_t, FILE*)' scons: *** [src/v8cgi.o] Error 1
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by ondrej.zara   -   Okay, your fix added as r693. I will try to investigate more leaks.

  Okay, your fix added as r693. I will try to investigate more leaks.
• Dec 10, 2009

  r693 (#47 memory leak caused by improper memory releasing, thanks ...) committed by ondrej.zara   -   #47 memory leak caused by improper memory releasing, thanks to neocoder

  #47 memory leak caused by improper memory releasing, thanks to neocoder
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by neocoder   -   Sure

  Sure
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by neocoder   -   Sure

  Sure
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by ondrej.zara   -   Okay, you are right. Can you please send me the patch you have so far? As for the V8, please see http://code.google.com/p/v8/wiki/Source?tm=4 .

  Okay, you are right. Can you please send me the patch you have so far? As for the V8, please see http://code.google.com/p/v8/wiki/Source?tm=4 .
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by neocoder   -   Yes they freed with free() function but should be freed with delete[] as they are created with new[]. I don't really think that this could cause the leak bu as far as I saw on the net everyone recommend to use standard malloc()/free() or new[]/delete[] pairs. I ran Valgrind memory leak detector and it showed only these problems. Everything else is a possible leaks because of using apr pool.(two logs are attached. before "delete[]" patch and after). Ondrej, where can I download the latest libv8 source to see if it has this problem? Now I'm using source provided with v8cgi 0.7.5. Oh, here is the link to Valgrind if it will be helpful - http://valgrind.org/

  Yes they freed with free() function but should be freed with delete[] as they are created with new[]. I don't really think that this could cause the leak bu as far as I saw on the net everyone recommend to use standard malloc()/free() or new[]/delete[] pairs. I ran Valgrind memory leak detector and it showed only these problems. Everything else is a possible leaks because of using apr pool.(two logs are attached. before "delete[]" patch and after). Ondrej, where can I download the latest libv8 source to see if it has this problem? Now I'm using source provided with v8cgi 0.7.5. Oh, here is the link to Valgrind if it will be helpful - http://valgrind.org/
• Dec 10, 2009

  issue 47 (massive memory leak) commented on by ondrej.zara   -   Where exactly in mod_v8cgi_handler do you see the leak? The only new'ed structure is the char ** envp and this is freed in the end of functions...

  Where exactly in mod_v8cgi_handler do you see the leak? The only new'ed structure is the char ** envp and this is freed in the end of functions...
• Dec 10, 2009

  issue 47 (massive memory leak) Status changed by ondrej.zara   -  

  Status: Accepted

  Status: Accepted
• Dec 10, 2009

  issue 47 (massive memory leak) reported by neocoder   -   Hi, Yesterday I tried to check performance of the server with Apache Benchmark. I made 1000 requests and monit restarted apache because of 200MB memory limit it reached. I looked through code and found one place where memory allocated with new char[] but freed with free(). it's in mod_v8cgi.cc (mod_v8cgi_handler function). I've fixed this but module still eats memory. The bug should be somewhere else, but it's hard for me to find it because I don't have normal dev environment yet.

  Hi, Yesterday I tried to check performance of the server with Apache Benchmark. I made 1000 requests and monit restarted apache because of 200MB memory limit it reached. I looked through code and found one place where memory allocated with new char[] but freed with free(). it's in mod_v8cgi.cc (mod_v8cgi_handler function). I've fixed this but module still eats memory. The bug should be somewhere else, but it's hard for me to find it because I don't have normal dev environment yet.
• Dec 04, 2009

  API (API Documentation) Wiki page commented on by ondrej.zara   -   Yes, I mean some location set in %path%, so Windows can find the dll if it is requested to load. If you use v8cgi as a CLI application, it is sufficient to put the in v8cgi's directory. If you use it as Apache module, it must be somewhere in %path%. For further troubleshooting, I suggest using the v8cgi mailing list instead of this comment-based form.

  Yes, I mean some location set in %path%, so Windows can find the dll if it is requested to load. If you use v8cgi as a CLI application, it is sufficient to put the in v8cgi's directory. If you use it as Apache module, it must be somewhere in %path%. For further troubleshooting, I suggest using the v8cgi mailing list instead of this comment-based form.
• Dec 04, 2009

  API (API Documentation) Wiki page commented on by tothanh7604   -   Yes. I copied all modules in "...\dlls\pgsql" to v8cgi directory. System path? You mean Environment variable? I can add a system path that point the neccesary dlls directory ?

  Yes. I copied all modules in "...\dlls\pgsql" to v8cgi directory. System path? You mean Environment variable? I can add a system path that point the neccesary dlls directory ?
• Dec 04, 2009

  API (API Documentation) Wiki page commented on by ondrej.zara   -   You probably forgot to put the necessary additional DLLs to system path. Please see the "dlls" subdirectory and copy those files to some convenient location, where Windows can find them.

  You probably forgot to put the necessary additional DLLs to system path. Please see the "dlls" subdirectory and copy those files to some convenient location, where Windows can find them.
• Dec 03, 2009

  API (API Documentation) Wiki page commented on by tothanh7604   -   I can not include all of module. I get an error: " Error opening shared library 'E:\RocketWebIDE\v8cgi\lib\pgsql.dll' " I think this message mean: v8cgi can find pgsql module, but it can not read it. The same error with other modules.I'm using the lastest version v8cgi

  I can not include all of module. I get an error: " Error opening shared library 'E:\RocketWebIDE\v8cgi\lib\pgsql.dll' " I think this message mean: v8cgi can find pgsql module, but it can not read it. The same error with other modules.I'm using the lastest version v8cgi
• Dec 02, 2009

  issue 46 (system.usleep) Status changed by ondrej.zara   -   Commited in r692.

  Status: Fixed

  Commited in r692.

  Status: Fixed
• Dec 02, 2009

  r692 (mgw fixes) committed by ondrej.zara   -   mgw fixes

  mgw fixes
• Dec 02, 2009

  r691 (#46 usleep) committed by ondrej.zara   -   #46 usleep

  #46 usleep

Earlier this year

• Dec 01, 2009

  issue 46 (system.usleep) changed by ondrej.zara   -  

  Status: Accepted

  Owner: ondrej.zara

  Status: Accepted

  Owner: ondrej.zara
• Dec 01, 2009

  issue 46 (system.usleep) reported by vanMuiswinkel   -   Since there is no retry on lock in sqlite, having a system.usleep would be useful.

  Since there is no retry on lock in sqlite, having a system.usleep would be useful.
• Nov 28, 2009

  issue 45 ('v8cgi' module.) Status changed by ondrej.zara   -   Implemented in r690. Apache module displays "?" in executableName, because I don't know how to get the module binary path :)

  Status: Fixed

  Implemented in r690. Apache module displays "?" in executableName, because I don't know how to get the module binary path :)

  Status: Fixed
• Nov 28, 2009

  r690 (#45 - v8cgi free variable. Apache module displays '?' for ex...) committed by ondrej.zara   -   #45 - v8cgi free variable. Apache module displays '?' for executableName

  #45 - v8cgi free variable. Apache module displays '?' for executableName
• Nov 27, 2009

  issue 24 (freebsd7.2 can't install) commented on by changlon   -   Freebsd need -lexecinfo -I/usr/local/include -L/usr/local/lib and mysql is /usr/local/mysql

  Freebsd need -lexecinfo -I/usr/local/include -L/usr/local/lib and mysql is /usr/local/mysql
• Nov 27, 2009

  issue 38 (Relative directory toString() value) Status changed by ondrej.zara   -   Path-based functions will be part of the FS/IO module, currently being specified by CommonJS.

  Status: WontFix

  Path-based functions will be part of the FS/IO module, currently being specified by CommonJS.

  Status: WontFix
• Nov 27, 2009

  issue 45 ('v8cgi' module.) changed by ondrej.zara   -  

  Status: Accepted

  Owner: ondrej.zara

  Status: Accepted

  Owner: ondrej.zara
• Nov 27, 2009

  issue 45 ('v8cgi' module.) reported by markobc...@gmail.com   -   Feature request: a 'v8cgi' module that contains several members describing the engine. -- Example -- var v8cgi = require('v8cgi'); v8cgi.version; // version descriptor (eg. '0.7.5' for release, '0.7.6dev' for trunk). v8cgi.executableName; // path to the v8cgi executable. v8cgi.instanceType; // 'standalone' if invoked directly or 'module' if an apache module -- Prior Art -- flusspferd's 'flusspferd' module - http://flusspferd.org/docs/js/flusspferd%20core/flusspferd.html

  Feature request: a 'v8cgi' module that contains several members describing the engine. -- Example -- var v8cgi = require('v8cgi'); v8cgi.version; // version descriptor (eg. '0.7.5' for release, '0.7.6dev' for trunk). v8cgi.executableName; // path to the v8cgi executable. v8cgi.instanceType; // 'standalone' if invoked directly or 'module' if an apache module -- Prior Art -- flusspferd's 'flusspferd' module - http://flusspferd.org/docs/js/flusspferd%20core/flusspferd.html
• Nov 27, 2009

  r689 (updated dom unit test, added doctype to dom/parser/serialize...) committed by ondrej.zara   -   updated dom unit test, added doctype to dom/parser/serializer

  updated dom unit test, added doctype to dom/parser/serializer
• Nov 26, 2009

  r688 (xmldom parsing) committed by ondrej.zara   -   xmldom parsing

  xmldom parsing
• Nov 26, 2009

  API_DOM (Module "dom") Wiki page edited by ondrej.zara   -   Revision r687 Edited wiki page through web user interface.

  Revision r687 Edited wiki page through web user interface.
• Nov 26, 2009

  r686 (xmldom parsing) committed by ondrej.zara   -   xmldom parsing

  xmldom parsing
• Nov 25, 2009

  issue 44 (Can't set two cookies at a time) Status changed by ondrej.zara   -   Commited in r685, thanks!

  Status: Fixed

  Commited in r685, thanks!

  Status: Fixed