Self Updating Exe par on Windows - windows

I am trying to create a script that will "self update" when it detects a new version on a server. Initially the idea I had was, when a new version is detected, download the file, then starts a DOS batch file that simply overwrites the original exe with the new one. I thought, that I could use sytem(start update.bat ) then immediately exit 0; The .bat file waits for a couple seconds, and then tries to delete the old exe. This fails, I guess because even when using system(start ... ), the new "process" is actually the same process, is this correct? Is there any way to launch a completely new process from perl in windows, that would allow me to delete the .exe? Or is there a different approach that would be better?
Thanks,
Eric Seifert

See the documentation on the Windows Restart Manager, an API introduced with Windows Vista that manages restarting applications, for updates, for you.

Ever heard of java WebStart? There is a PAR equivalent, PAR::WebStart
Basically, you configure it, then download code live from a trusted website
so your .exe would just be a frontend to PAR::WebStart, with the real program downloaded off your website

I know this is an old question, but I thought I'd put my solution here since I'm working on the same sort of thing. Your batch idea is the way I went, but the small change I made was to use exec() instead of system() since the former (according to this) "...executes a system command and never returns", which seems to release the running process. This appears to work both when running interpreted Perl files with Strawberry Perl, and when running compiled programs with PAR::Packer.
I've created test.pl, which was ultimately compiled to test.exe, as follows:
#!/usr/bin/perl
use 5.10.0;
use strict;
use warnings;
my $batch_file = <<"BATCH";
rem ### IT IS SAFE TO DELETE THIS FILE ###
\#echo off
ping 127.0.0.1 -n 2 -w 1000 > nul
echo Copying over update file...
copy testfile.exe test.exe
BATCH
open my $fh, '>', 'update.bat';
print $fh $batch_file;
close $fh;
exec( 'update.bat' );
testfile.exe is a standby "upgraded" version of test.exe, just to make sure that I can overwrite test.exe while it's running, and I can. The ping is a way to pause for 2 seconds to make sure the process has had a chance to exit before trying to overwrite the file.
Oddly enough, the created batch file update.bat can't delete itself, even though batch files can normally delete themselves. If I create a batch file with this:
start /b "" cmd /c del %0
It will delete itself without any errors. But if I include that snippet in update.bat, it complains that The process cannot access the file because it is being used by another process. I'm not sure why. Since this isn't particularly important in my application, I haven't pursued this. I just leave update.bat behind and let it be overwritten next time an update occurs. But I do notate in the file that it's safe to delete in case anyone looks at it later.

Related

How to work around ebook-convert.exe (in calibre portable) halting batch file execution?

I use ebook-convert.exe from a batch file, and on one of my machines (old Windows vista laptop), any calls ebook-convert.exe prevent further commands in a batch file from running (in that .bat file, or in that cmd shell instance for instance). The conversion to .mobi actually succeeds, prints no errors even in -v -v full verbose mode. I tried versions 1.20.0 and 0.9.6 of calibre portable. It seems to be related to the executable itself, and not to the specific conversion job (just running it with -h flag has same effect). I tried running cmd in administrator mode, wrapping it with 'call' command, nothing helps. Assuming the calibre issue is a black hole that can't be solved, is there a DOS trick I could possibly try to keep batch commands running no matter what odd thing happens with ebook-convert.exe on that machine?
Using call in a batch file should work on normal console programs but it is possible for programs to "escape" if they really want to.
Using Start instead might work but it does not wait by default and its syntax is not sane:
Start /B /WAIT c:\path\without\spaces.exe optionalParameter
or
Start "" /B /WAIT "c:\path\w i t h\spaces.exe" optionalParameter
Remove /B to run the application in a new console.
Another option is to call ping localhost and For (to get the file size with %%~zA) in a loop until the size of the destination file stops increasing. This is a bit of a hack and might not work, it depends on how the converter opens the file and how often it flushes the writes.

Starting independent batch files through ssh and process ownership

I've used batch files for many things in the past... but I've always had this problem. I'm sorry if this is a repeat question, I'm not entirely sure I know how to phrase it for searching purposes. The problem is this:
1) Batch file starts some process.
2) command window closed by user.
3) process started by batch file ends.
I imagine this is due to the fact that the started process is "called" by the batch file, and is thus it's child. Specifically what I'm trying to do is login to a server through ssh, run a batch file located on that server which then starts a java program. I need the batch to either stay open, or allow the java program to own itself somehow. That way, when I leave the SSH session, the program will continue to run. Any ideas how I can do this?
I'm running a windows XP x64 server with MobaSSH.
You could try using the psexec tools from sysinternals.
Some possible helpful commands:
at
schtasks
sc
wmic
I'm not sure that any of the above commands will be of any help, but I think they're worth checking out.
Question is not clear, but looks like what you are looking for is a way to "detach" the script from the terminal so that it will continue to run even when the terminal is closed.
You can do:
nohup <your-script> &
Or:
<your-script> &
disown

CMD: Bat to Exe Converter - Temp directory problem

i am using 'Bat to Exe Converter' to convert my batch files to exe format.
Now, i am running into some problems.
Whenever i convert something, and i set 'Working Directory' to 'Current Directory', and i start my exe in echo on mode, this is what i end up with to check if there is a specific file in the directory of my exe:
the actual command: if not exist "%~dp0\file.txt" goto :nofile
output: if not exist "C:\Users\MyUser\AppData\Local\Temp\4CBC\\file.txt" goto :nofile
Can anyone help me with this? I don't want it in the temp directory, i want it to be in the directory of my exe.
Thanks.
Without having Bat to Exe changed by the author, I think you have two options:
Remove the need for accessing %~dp0
Perhaps you can merge file.txt with the include option of Bat to Exe into the EXE file. If so, "file.txt" will automatically be unpacked in the current directory when running your compiled exe, and you can it access by %CD%\file.txt.
Get %~dp0 from outside and pass it to the exe as a command line parameter.
This can be done by a simple starter bat file that resides in the same directory as your compiled main batch file. This script schould contain the line
YourCompiled.Exe %~dp0% %%*
Your compiled exe then gets its directory from %1. So you cannot pack everything into one exe, but the main portion of it, perhaps that is sufficient for you.
Well, apparently your batch to exe converter simply packs the batch file and extracts it to a temporary directory prior to execution. Very simplistic, hard to get wrong (compared to actually understand the batch file) but it introduces errors such as the one you're describing.
Your best bet is probably to use another batch to exe converter; some of them are actually a little more sophisticated.
Generally, this is not a good idea. firstly, its prone to errors and instability of the converter on different cmd features. secondly, a determined hacker can still decode what you are doing with the batch. My suggestion, if you are so afraid of people looking into your batch,
1) let only the people who are authorized to use your batch to use it
2) give them the correct permissions.
OR, don't use batch at all
1) create a central interface such as a web interface, where all tasks to be done goes through that interface, like using an ATM machine where only buttons are allowed and all the available user options can be done by pushing buttons...etc..
2) authenticate your users through a central authentication system, eg Active Directory, or LDAP or a database.
This is an 2.5 yr old subject but there is an answer to this so I'm posting for anyone else that happens to find this in a search.
B2EC written by Fatih Kodak, has an option to "Submit current directory".
When this is used, you can reference %1 in your batch file to get the path of the EXE that was executed (instead of the path of the extracted BAT that is really being run).
Hovering over that option in the UI shows "Submit the current working directory as the last parameter". The "last parameter" in my use has always been %1 but you can test your code to be sure.
The latest version, 2.1.4 at time of writing, of Bat to Exe by Fatih Kodak creates an Environmental Variable at runtime that can be substituted in place of %~dp0 to reference the Exe's path. Therefore, you can simply replace %-dp0 with %b2eprogrampathname% in the original batch file.
You can use external folders with f2ko's batch to exe converter. Having
a separate folder for subroutines can neaten up a project folder.
To call mysubroutine that is located in mysubroutinesfolder\mysubroutine,
...
pushd mysubroutinesfolder
call mysubroutine
popd
...
The call can be made a one liner:
call xqt mysubroutine
where xqt.cmd is a program that does the call for you:
pushd mysubroutinesfolder
call %*
popd
exit /b
(the %* means "all of the arguments").
In this way your batch programs run as batch, and UNMODIFIED they will
compile with the bat to exe converter, creating a completely folder independent executable. Select "temporary directory",
and include all of the subroutines/executables in your mysubroutines folder
by "selecting them all" with your cursor as usual, then hit "copy".
Be sure to include the xqt.cmd program too; place it "outside" of your mysubroutines folder. Make sure that is is accessible by your main program. Remember to select x64 if you
are runnning on a x64 machine, or the executable will not find SYSTEM32
files. You can find f2k0's batch to exe converter at:
http://www.f2ko.de/programs.php?pid=b2e
Try this development environment for batch scripts, Batch Compiler . It has everything you need to develop a batch program.And compile into stable stand alone executable (Exe).
Friendly user interface.
Debugger, Check your code for syntax errors.
Powerful, versatile compiler.
Allows mouse input in batch files.
Use Windows Common Dialog Boxes.(BrowseFiles,BrowseFolders)
Draw graphics in batch files.
Reverse engineering proof encryption of source code.
Include Company name, Copyright info and Version info.
Make invisible(silent) executables.
Executables with administrator privileges.
Run & debug your script while editing.
Embed resources with executable.(music,images,files)
Advance Commands (BrowseFiles,LaunchSilent,MouseCMD)
Stand-alone executables.No dependencies needed.
Executables are woking on almost all windows operating systems.(98 to 10)
Quick download : http://bc.gotek.info/files/BatchCompiler159.zip
Cheers!
%cd% will give you the current directory:
if not exist "%CD%\file.txt" goto :nofile
Use %CD% instead of %~dp0.
EDIT:
B2EC is not a real converter. Creation location of equipped .cmd file was chosen to be %TEMP% and this is a good choice. Application just lacks 3rd option for working directory of the script - .exe file directory. I advise you to mail the author about adding this one.
Different paths for .exe and created .cmd lead to information loss, i.e. we are unable to know .exe directory and current directory at the same time without providing additional information to the script (e.g. using environment variable or passing it as first/last argument to the script). This script would need to handle it and we would end writing cmd scripts tailored for this converter, which is bad.
%~dp0 - script directory (%TEMP%/.../) - practically useless
%cd% - working directory (as set up in the converter) - currently there are only 2 options: current directory (working directory of .exe) and temporary directory (actually equal to %~dp0, but without trailing backslash)
I think it can be solved by patching cmd.exe instance in memory to change the script path, but that's B2EC developer's duty.
Side note: Normal executable files can be easily executed with specified 0th argument by providing appriopriate lpApplicationName and lpCommandLine to CreateProcess function. Command files are executed via cmd.exe, so 0th argument cannot be set this way.

Why doesn't cl.exe generate any output when I call it from Perl?

I'm having a weird problem with running cl.exe that has me stumped. In a large VS2008 solution consisting of C/C++ projects, I have one project that runs some scripts to do some extra processing. The project consists of a pre-build event, which calls a Perl script (ActiveState Perl is on the machine). This Perl script then calls cl.exe with /E to generate preprocessed output which gets redirected to a file. The line in Perl looks like this:
my $foo = `"\path\to\cl.exe" #args.rsp >out.txt 2>err.txt`;
args.rsp is a plain text file that contains a bunch of command line args for cl.exe, including /E to get pre-processor output on stdout.
This exact command line works as expected when run from the VS2008 command prompt. Building the project also works fine on my Windows XP machine. However, on my new Windows 7 box, when I build the project, out.txt ends up blank. I should also add that on some of my coworker's Windows 7 boxes, it works fine, and on some others it doesn't.
Clearly there's some kind of configuration difference going on, but I'm at a loss as to what it may be. We've checked matching versions of VS2008 SP1 and ActiveState Perl. I've tried myriad workarounds within the perl script - using system() instead of backticks, using cl.exe /P to output to a file and then moving the file (the file is blank), unsetting the VS_UNICODE_OUTPUT environment variable (no effect). Nothing has changed the behavior - output is generated when the command line is run manually, but not when it's run inside the pre-build event for this project.
Any ideas on what kind of configuration problem may be causing this? I'm pretty much out of avenues to pursue.
Sounds like an ACL issue to me. You can change windows to log access issues and then check the event log to see what user is getting access denied errors.
I believe the setting is in Local Policy | Audit Policy | Audit Object Access
Wow, the solution to this ended up being a lot stranger than I expected. The machine I'm working on (and the other co-workers who are also experiencing the problem) is a Mac Pro with bootcamp and Windows 7 installed. That causes C: to have the windows drive and E: to have the mac drive. This causes a problem because the pre-build event has a couple lines that test each drive letter to see if there's a drive there, and if there is, adds X:\Perl\bin to the path. Even though E:\Perl\bin doesn't exist, it gets added to the path. Later, the perl script runs and then calls cl.exe, and for some reason, having a directory in the mac drive causes cl.exe to fail. Why? I have no idea. Anyway, removing the mac drive directory from the path fixes the problem!
Thanks for your eyes everyone.
Check out the exit code of your program. You may want to build your executable name in a portable way using something like File::Spec. Also, check that #args is not interpolating. You may want to print your command line before executing to check if that's what you want. What is left your err.txt file?

How to capture and display output from a task via Windows CMD

I've got a PHP script which I'm running from a command line (windows) that performs a variety of tasks, and the only output it gives is via 'print' statements which output direct to screen.
What I want to do is capture this to a log file as well.
I know I can do:
php-cli script.php > log.txt
But the problem with this approach is that all the output is written to the log file, but I can't see how things are running in the mean time (so I can stop the process if anything dodgy is happening).
Just to pre-empt other possible questions, I can't change all the print's to a log statement as there are far too many of them and I'd rather not change anything in the code lest I be blamed for something going fubar. Plus there's the lack of time aspect as well. I also have to run this on a windows machine.
Thanks in advance :)
Edit: Thanks for the answers guys, in the end I went with the browser method because that was the easiest and quickest to set up, although I am convinced there is an actual answer to this problem somewhere.
You can create a powershell script that runs the command, reads the data from the command's STDOUT then outputs the output to both the log file and the terminal for you to watch. You can use the commands Write-Output and Write-Host.
Microsoft's site: http://www.microsoft.com/technet/scriptcenter/topics/msh/cmdlets/tee-object.mspx
Another option would be use find a tee program that will read input and divert it to two different outputs. I believe I have seen these for windows but I don't think they are standard.
Wikipedia: http://en.wikipedia.org/wiki/Tee_(command)
I have always opened the log file up in my web browser. This allows me to refresh it easily and does not interrupt any writing to the file that windows does. It isn't particularly elegant but it does work!
You want the "tee" command for Windows. See http://en.wikipedia.org/wiki/Tee_(command)
Powershell includes a tee command, and there are also numerous versions of tee for Windows available, for instance:
http://unxutils.sourceforge.net/
http://www.chipstips.com/?p=129
Also can be implemented in VBScript if you prefer.
EDIT: Just occurred to me I should also mention the tail command: http://en.wikipedia.org/wiki/Tail_(Unix). Tail allows you to read the last N lines of a file, and also includes a "file monitor" mode that just continually displays the end of the file in real-time. This is perfect for log file monitoring since it allows you to watch the log in real-time without interfering with the process that's writing to the log. There are several implementations of tail for Windows, both command line and GUI based. Microsoft's Services For UNIX packages (or whatever they're calling it now) also include a version of tail. Some examples:
mTail
Tail for Win32
WinTail
MakeLogic Tail
Some of these go far beyond just displaying the file in real-time as it updates and can send email alerts and colorize string matches, monitor multiple files at once, etc.
Slow:
for /f "delims=" %a in ('php-cli script.php') do #echo %a&echo %a>>log.txt
or in a batch file:
for /f "delims=" %%a in ('php-cli script.php') do #echo %%a&echo %%a>>log.txt

Resources