majaxadobe / teajs Goto Github PK

http://code.google.com/p/v8cgi/issues/detail?id=16

exception: /usr/local/src/teajs# scons -h scons: Reading SConscript files ...  
scons: *** Path for option v8_path does not exist: ../v8 File 
"/usr/local/src/teajs/SConstruct", line 372, in <module>

What steps will reproduce the problem?
1. Send a request with 'Cookie: aaa=bbb;ccc=ddd";
2. request.cookie['ccc'] does not exist.
3.

What is the expected output? What do you see instead?
ddd.  undefined.

What version of the product are you using? On what operating system?
ubuntu, teajs 0.9.4

Please provide any additional information below.


To fix this, change...

http.js:498
        var all = this._headers["HTTP_COOKIE"].split("; ");
        for (var i=0, len=all.length;i<len;i++) {
                var row = all[i];
                var eq = row.indexOf("=");
                var name = row.substring(0,eq);


to...


http.js:498
        var all = this._headers["HTTP_COOKIE"].split(";");
        for (var i=0, len=all.length;i<len;i++) {
                var row = all[i];
                var eq = row.indexOf("=");
                var name = row.substring(0,eq);
                name=name.trim();

http://code.google.com/p/v8cgi/issues/detail?id=104

http://code.google.com/p/v8cgi/issues/detail?id=117

http://code.google.com/p/v8cgi/issues/detail?id=116

-*- mode: compilation; default-directory: "~/teajs/" -*-
Compilation started at Fri Nov 23 21:19:53

scons v8_path=/usr/lib64 binary_b=1 pgsql=1 
scons: Reading SConscript files ...
Checking for C header file sys/mman.h... (cached) yes
Checking for C function sleep()... (cached) yes
Checking for C function inet_pton()... (cached) yes
Checking for C function inet_ntop()... (cached) yes
Checking for C library v8... (cached) yes
scons: done reading SConscript files.
scons: Building targets ...
g++ -o src/lib/binary-b/bytestorage-b.os -c -Wall -O3 -fPIC -DHAVE_MMAN_H 
-DHAVE_SLEEP -DHAVE_PTON -DHAVE_NTOP -DCONFIG_PATH=/etc/teajs.conf 
-DVERSION=0.9.5~dev -Dposix -DDSO_EXT=so -Isrc -I/usr/lib64/include 
src/lib/binary-b/bytestorage-b.cc
g++ -o src/lib/pgsql/pgsql.os -c -Wall -O3 -fPIC -DHAVE_MMAN_H -DHAVE_SLEEP 
-DHAVE_PTON -DHAVE_NTOP -DCONFIG_PATH=/etc/teajs.conf -DVERSION=0.9.5~dev 
-Dposix -DDSO_EXT=so -Isrc -I/usr/lib64/include "-I`/usr/bin/pg_config 
--includedir`" src/lib/pgsql/pgsql.cc
src/lib/pgsql/pgsql.cc:40:29: fatal error: postgres_fe.h: No such file or 
directory
compilation terminated.
scons: *** [src/lib/pgsql/pgsql.os] Error 1
scons: building terminated because of errors.

Compilation exited abnormally with code 2 at Fri Nov 23 21:19:55

Since the only part of pgsql you’re intending to use is the client library, 
you must follow the rules for using the client library. The docs say: 
http://www.postgresql.org/docs/9.2/interactive/libpq-build.html , or 
“#include <libpq-fe.h>”—nothing about postgres_fe.h, which isn’t 
installed into pg_config --includedir anyways.

ohnobinki@gentoodcxdelllappy ~/teajs $ qlist -e postgresql-base | grep -e 
postgres_fe\\.h
/usr/include/postgresql-9.2/internal/postgres_fe.h
/usr/include/postgresql-9.2/server/postgres_fe.h
ohnobinki@gentoodcxdelllappy ~/teajs $ pg_config --includedir
/usr/include/postgresql-9.2

So, I suggest to use define TRUE and FALSE following the convention anyone else 
does.

The attached code, along with formdef.txt which contains 156kB of base 64 text 
will lock up with both cpus pegged at 100%
Some debug in the base64.js code shows that it dies repeatedly at the same 
point in the loop.


I am using teajs 0.9.4 on Windows 7 32-bit on a core2 duo PC

in http.js line 65, change:

ServerRequest.prototype.header = function(name) {
/*--*/  var n = "HTTP_"+name.toUpperCase().replace(/-/g,"_");
    return (n in this._headers ? this._headers[n] : null);
}

To

ServerRequest.prototype.header = function(name) {
/*--*/  var n = name.toUpperCase().replace(/-/g,"_");
    return (n in this._headers ? this._headers[n] : null);
}

http://code.google.com/p/v8cgi/issues/detail?id=76

The buffer class should force a GC and retry when buffer malloc fails.  It is 
entirely feasible that the reason for the malloc failure is many dead but 
uncollected buffers.  If there is room to allocate the buffer object but not 
enough to allocate the storage then the allocation will fail even though there 
may be potentially huge amounts of free memory.

It also seems odd that it is possible to have upwards of 2GB in buffer 
allocations before garbage collection occurs.  This is incredibly unfriendly to 
other applications (let alone the buffer malloc call.) 

-*- mode: compilation; default-directory: "~/teajs/" -*-
Compilation started at Fri Nov 23 21:14:01

scons v8_path=/usr/lib64 binary_b=1 pgsql=1 
scons: Reading SConscript files ...
Checking for C header file sys/mman.h... (cached) yes
Checking for C function sleep()... (cached) yes
Checking for C function inet_pton()... (cached) yes
Checking for C function inet_ntop()... (cached) yes
Checking for C library v8... (cached) yes
scons: done reading SConscript files.
scons: Building targets ...
g++ -o src/lib/binary-b/bytestorage-b.os -c -Wall -O3 -fPIC -DHAVE_MMAN_H 
-DHAVE_SLEEP -DHAVE_PTON -DHAVE_NTOP -DCONFIG_PATH=/etc/teajs.conf 
-DVERSION=0.9.5~dev -Dposix -DDSO_EXT=so -Isrc -I/usr/lib64/include 
src/lib/binary-b/bytestorage-b.cc
src/lib/binary-b/bytestorage-b.cc: In member function 'unsigned char 
ByteStorageB::getByte(size_t)':
src/lib/binary-b/bytestorage-b.cc:102:38: warning: converting to non-pointer 
type 'unsigned char' from NULL [-Wconversion-null]
scons: done building targets.

Compilation finished at Fri Nov 23 21:14:02

If the very last line in a source file is a line comment then the file will 
fail to run with an 'unexpected end of input' error.

Adding a blank line after the comment will fix it.


Teajs 0.9.4 on Windows 7 32-bit

Clone the teajs repository on a machine with CentOS 6. Execute scons in that 
repository.

Expecting successful compile. Getting apr.h-related errors.

Changing SConstruct, setting apr_include to /usr/include/apr-1 instead of 
/usr/include/apr-1.0, proves a working workaround.

http://code.google.com/p/v8cgi/issues/detail?id=85

What steps will reproduce the problem?
1. Make a file in /tmp owned by the web server's user.
2. use new Session()  lots of times for an hour.
3. Eventually the file that was created will be removed & the script slows down 
alot cause there're lots of files to clean up in /tmp

What version of the product are you using? On what operating system?
teajs

Please provide any additional information below.
Here is a patch to fix these problems.
Also, the default path for the session in v8cgi.conf should be something other 
than /tmp

http://code.google.com/p/v8cgi/issues/detail?id=92

hi, i use teajs to develep a app , but i don't know how to debug the javascript 

Does have a tool like node-inspector the debug the javascript ?

Thanks