Giuseppe "Cowo" Corbelli ...

A few simple words to express what I think...

Fuck code signing/security/whatsoever: let MY hardware execute what I want, huh?

I read somewhere that this can be done using Visual Studio 2008; don't know, as I don't (still) use the beast.
Here's what to do:

1. Download and install Device command shell from here http://cowo.supersized.org/archives/28-Visual-Studio-2005-Device-Command-Shell.html
2. Download Security-Off provisioning XML file from here: SecurityOff.xml or use the one distributed with Windows Mobile 6 SDK
3. Execute ce connect <YourMobileID>
4. Execute ce config SecurityOff.xml

Really tired of all this mobile shit; I'd say it's holiday time...

Device Command Shell (DCS) provides a command-line interface to several useful functions for developing Smart Device applications using Visual Studio 2005. It is implemented as a Visual Studio "Command Window" add-in for convenience and quick access.

This is the author description for DCS. Useful tools, indeed. Download URL used to be http://www.gotdotnet.com/workspaces/workspace.aspx?id=50618f79-c7b1-4588-9c0a-cf4ddae8092a but seems that M$ is moving everything to MSDN Code Gallery and this stuff got lost somewhere. OK, don't know if it's OK for me to redistribute, but many seem to need this...
I believe this should be version 0.98...

Download Device Command Shell here: DeviceCommandShellSetup.zip

Back in the good old times there was boost::build v1, which I was more or less able to comprehend. Now I've installed boost v1.35 it seems that only boost::build v2 is available; I'm sure it rocks and so on, but I've struggled TOO MUCH to make it work. Me being a dumbass? Quite likely.

But at long last I was able to put all the pieces together:

1. Find tools/build/v2/user-config.jam under boost root
2. Uncomment the using msvc ;
3. Find the stlport line and write something like using stlport : : STLPORTPATH\stlport STLPORTPATH\lib ;
4. Place bjam.exe into boost root
5. Run bjam msvc/variant=debug,release/threading=multi/stdlib=stlport/link=shared,static/runtime-link=shared stage

This shit will build debug and release, static and dynamic libraries linked to dynamic C runtime.

I have a few NICs on a machine, so I told Microsoft SQL Server 2005 - 9.00.3042.00 (Intel X86) not to bind on every interface but only one. So far so good.
Unfortunately I had the BAAAD idea to change this NIC IP Address. Old IP was 192.168.0.103, new IP is 192.168.6.65. When I try to start the service with commandline sqlservr.exe -sSQLEXPRESS where's what I get

Server failed to listen on 192.168.0.103  1433. Error: 0x2741. To proceed, notify your system administrator.

Standard useless error message. Ok, there must be this old address configured somewhere. In the registry, of course. Just searched throughout the whole registry and found:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft SQL Server\MSSQL.1\MSSQLServer\SuperSocketNetLib\Tcp

Here you'll find subkeys named IP1 to IPSomeNumber. If one refers to an unbindable address just set Enabled or Active to 0.

I just bought a GEM Equinox and I'm looking for some RAM to add. It requires 30 pin SIMMs, so I went through some rubbish. Here's the capacity of the RAM I found:

2X PM10244-70 T9210 1X M5M41000BJ 234GE20-7	1 MB
1X TMS4C1024DJ HEP 3312 24 2X TMS44400DJ OBI 4051 AH	1 MB
2X KOREA AO169315GGG HM514400AS7 1X KOREA A0029305GGGK HM511000AJP7	1 MB

I own a Yamaha PSR 5700 since mid nineties. Not the best keyboard around, but has worked nicely for a decade, so no complaints here.
However the floppy failed last week. It all started with some random "Disk Error" and then no disks would work. Floppy is a Mitsubishi MF353C-152MZ.
Ok, I thought, 10 years is enough. I grabbed a new floppy from a PC, opened the keyboard and hooked it up. And, surprise, it did not work!
To make it short: I have found a few references on the net, but no ready solution (I did not consider buying a new 130 USD replacement floppy a solution).
If you're reading this, you're likely sharing my past week worries; but light up your heart! Read on and find the solution.
I thought the whole floppy chain (power supply, floppy drive, controller) had to be build by standard components, so something could be done with acceptable effort. I was right (one of the few times :-)
I started here: http://pinouts.ru/Storage/InternalDisk_pinout.shtml
This document states Also many synthesizers that have floppy drives use the standard Shugart interface pinout; so, let's say this Shugart interface is what PSR 5700 has (controller is a HD63266F but I could not find much info about this).
Compare the two pinouts: the Shugart interface has lines for 4 devices and a single motor-enable line, while modern controllers support two devices, density select and independent motor-enable lines. Also notice that modern floppies are usually jumpered to work as "B" or "Device 1".
At this point I had to do some measures and test, but cut it short.
Here's the summary:

• PSR 5700 selects Device 0, and we have a floppy jumpered as Device 1
• PSR 5700 relies on a Drive Ready signal, which can be either located on the floppy drive or faked

And the long awaited solution: grab a floppy cable, cut wires 10, 12, 34 (remember that wire 1 is red-marked) and solder together

• CONTROLLER pin 10
• CONTROLLER pin 34
• FLOPPY pin 12

With "CONTROLLER" I mean the half wire that goes to the keyboard. Now you have 3 half wires dangling and 3 half wires soldered together. Remaining wires are untouched.
What you've done is:

• Routing the "Device Select 0" signal to "Device 1" so that floppy gets correctly selected
• Faking the "Drive Ready" signal, so that when the PSR 5700 selects the floppy it appears to be ready for operation

That's it! New floppy on the PSR and no money drain!
BTW: remembering what said before, it is possible, if not likely, that this method would work on many other keyboards. Until now this stuff has been tested on:

• Yamaha PSR 5700 by Giuseppe Corbelli - works OK (see disk change detection issue below)
• Yamaha Clavinova CVP-70 by Greg Moseley - works OK
• Yamaha Clavinova CVP-55 by Gage Johnson - works OK
• Yamaha Clavinova CVP-65 by Stephen Marshall - works OK

Disk changes are not detected!. I think this can be solved by routing controller pin 2 to floppy pin 34 but I'm too lazy to reopen the keyboard :-)

I'm an happy user of Boost libraries and recently I had to mess with QT libraries.
I didn't really like the QString-to-number conversion features of QString class because of the clumsy error checking mechanism, so I wrote a little wrapper for boost::lexical_cast.
Unfortunately boost::lexical_cast is not written with extensibility in mind so a temporary std::basic_string variable is required to perform the magic (or at least I couldn't find a better, unobtrusive way).

namespace boost { namespace detail {

#ifndef DISABLE_WIDE_CHAR_SUPPORT
inline bool operator>>(boost::detail::lexical_stream<QString, std::wstring>& in, QString& output)
{
    std::wstring tmp;
    in >> tmp;
    output = QString::fromStdWString(tmp);
    return true;
}
#endif

template<typename Source>
inline bool operator>>(boost::detail::lexical_stream<QString, Source>& in, QString& output)
{
#if defined(BOOST_NO_STRINGSTREAM)
    in << '\0';
#endif
    std::string tmp;
    in >> tmp;
    output = QString::fromStdString(tmp);
    return true;
}

}}

This allows you to write something like:

QString x("123");
int id(boost::lexical_cast<int>(x));

I have to do a simple backup task to DVD-R using flexbackup on Linux 2.6. My first idea was to use packet writing: setup the media, mount as rw somewhere and write to it.

So I installed udftools (Debian package, version 1.0.0b3-12) and started to read docs. README.Debian states:

3) Write an empty session spanning the whole medium. It seems that without this step, any attempt to create the UDF filesystem will fail. DVD-RW: growisofs -Z /dev/hdc=/dev/zero

If applied to a non-rewritable media THIS PRODUCES A NICE COASTER!

Ok, just skip this step and go to the next, which should be

mkudffs --udfrev=0x0150 --spartable=2 --media-type=dvd /dev/pktcdvd/0

but this does not work, with error trying to change type of multiple extents.

In short, unless I'm missing something, this means YOU CANNOT USE UDF FILESYSTEM ON DVD-R MEDIA ON LINUX 2.6. Go for cdrecord/growisofs and ISO.

Had to install SLED 10 on recent ASUS laptop with this ethernet board. Standard SLED 10 kernel does not provide a driver; basically you have to

1. Download driver package from Realtek website
2. Build the driver for the default distribution and, optionally, for SMP
3. Create a driver disk for the default distribution, by following these instructions at Novell website
4. During installation process load the driver disk, then load module.
5. The installer won't copy the module to target system, but anyway the target system kernel may be different from installation one. You need to setup some RPMs or an update.tar.gz file with modules for all kernel versions you need to support.

Module name is r8168

Here you can find a tarball of the driver disk. Just unpack to a filesystem accessible during the installation process. This driver disk supports r8168 for the following kernels:

• 2.6.16.21-0.8-default
• 2.6.16.21-0.8-smp

As an happy user of the SciTE editor I was missing this feature in Komodo Edit which by the way is an excellent product.

Comment/uncomment means that a "special" comment string is used, i.e. #~ for Python. If this string is present at beginning of line the macro removes it (uncomment); if not it is inserted (comment). Python and PHP only for now, and poor code for sure. Any help is welcome.

import logging

logger = logging.getLogger('macro')
h = logging.FileHandler("/tmp/somefile", "w")
logger.addHandler(h)

ke = komodo.editor
kd = komodo.document
kv = komodo.view
commentMap = {
    u'python': '#~ ',
    u'php': '//~ ',
    u'javascript': '//~ ',
    u'c++': '//~ ',
}

commentTxt = commentMap[kd.language.lower()]

def EVAL(stmt):
    res = eval(stmt)
    logger.error("%s --> %s" % (stmt, res))

def getLineNo(position=None):
    pos = position or ke.currentPos
    return ke.lineFromPosition(pos)
    
def getLine(lineNo=None):
    line = lineNo or getLineNo()
    eol = ke.getLineEndPosition(line)
    return ke.getTextRange(ke.currentPos, eol)

def vCHome(lineNo=None):
    if (lineNo is not None):
        ke.gotoLine(lineNo)
    
    ke.vCHome()
    if (getLine().strip() == ''):
        return
    
    if (ke.getTextRange(ke.currentPos, ke.currentPos+1).isspace()):
        ke.vCHome()
        
def doLine(lineNo=None):
    vCHome(lineNo)
    pos = ke.currentPos
    s = ke.getTextRange(pos, pos+len(commentTxt))
    
    if (s == commentTxt):
        ke.setSel(pos, pos+len(commentTxt))
        ke.replaceSel('')
    else:
        ke.insertText(pos, commentTxt)

komodo.view.setFocus()

try:
    pStart = ke.anchor
    pEnd = ke.currentPos
    lStart = ke.lineFromPosition(pStart)
    lEnd = ke.lineFromPosition(pEnd)
    if (lEnd < lStart):
        tmp = lStart
        lStart = lEnd
        lEnd = tmp
    for line in range(lStart, lEnd+1):
        doLine(line)
    
except:
    logger.error("Error", exc_info=True)

return