I need to search through all sub directories of a directory and find all files containing a "VERSION" string including a number.
I need to increment this number, so 1.1.2 will be incremented to 1.1.3 etc. and save it in the file again.
I need to run this on Windows machines only, if it makes any difference.
Can I do this with cmd commands or do I need to use a program for this ?
I would like to run this without installing anything if possible.
I ended up writing a Java program to go through the entire structure.
I found that using CMD commands was too complicated for me to structure, and using a Java program I could also easily reuse parts of the program in different but similar structures elsewhere in my file system.
The CMD usage could my the files for me, but extract a single line, manipulating it and putting it back inside the file was not easy using CMD commands, where as using Java was much easier.
Related
It's just one of those days...
I want to create a shell script (naive version could be a simple alias), but I want to make it the good way from the bottom up. Writing script itself isn't a problem at all, but since it's possible a couple of other people would like to use this script as well I want to make it accessible and maintainable as well. That's why I have two questions:
Are there any guides regarding structure of the repository with shell script inside? I mean, some conventions that, for example, we put a script in some directory, manpage, should I split it into multiple files etc.
How to make this script be accessible to other people? I know that if we want to install command line tool, we move it to /usr/bin or /usr/local/bin, but what with docs or some dependencies?
There are a lot of tutorials regarding writing command line tools in Ruby or Python with some templates how to manage everything, but I haven't found anything about writing a command line tool as a shell script.
In general, you would follow the conventions in man hier for the distribution. A shell script would be treated the same as a binary file.
I'd like to write an AppleScript for replacing three system files with ones I've modified. I'd like to do this with an AppleScript instead of manually replacing them because I'll have to replace three files every time there's an OS X update. Specifically, I'll be replacing stock graphics drivers with ones I've modified to support a graphics card which is connected via Thunderbolt. Is it possible to write an AppleScript for replacing one file with another? I ask because I know that when you replace a file, a dialog pops up with three options, and I don't know how to address that.
You can do this with Finder:
set freshFile to choose file
tell application "Finder"
move freshFile to desktop replacing yes
end tell
All you need to do is work out the source and destination paths to completely automate the script.
Many scripters do not like working with Finder, for a variety of reasons. If you want something that is incredibly fast, you would use the do shell script inside of your AppleScript:
do shell script " mv -f ~/Desktop/ArlandaTilUppsala.pdf ~/Documents/Employ.pdf"
I want to implement an autoupdater in my Python application. This is no problem with the source version; Python doesn't care at all that the script it's running is trying to overwrite itself.
Windows, however, does care if an EXE tries to overwrite itself. My question is, does my EXE even have to overwrite itself? Or is the EXE just an interpreter, and I only have to overwrite library.zip?
If it does, is there any alternative to starting an updater application and shutting down the main EXE?
The cx_Freeze exe is compiled along with cx_Freeze, so in most cases you can safely leave it alone and just update library.zip. However, you should make sure that you prepare your updates using the same version of cx_Freeze that you froze the application with originally, in case it expects specific things about the files around it.
Also, the exe gets stamped with a version number relating to your application (you can see it in the file properties). If you don't replace it, that version number won't change.
If you do need to replace the exe, I believe the trick is to copy it to a temporary folder and re-run from there, so that it can replace the original. You can also look into update frameworks like Esky, which aim to handle these kinds of details for you. I haven't used that, so I don't know how well it works.
Assume a Visual C++ solution that outputs several executables. These executables are meant to be run in a certain order and with certain parameters -- and for this purpose there already is an ant build.xml script.
What would be a decent approach to integrating this ant script with VC++, such that the ant script will point against the recently output executables (.\Debug and .\Release folders) and ideally could be run directly from VC++, and dare I say with remote debugging.
I was thinking of using build post-events that populate a build.properties file with the output location of each executable, and let the ant script use this .properties file.
Any help on the matter would be great.
I'm not sure if there is a good answer for this. Perhaps you are not asking the right questions. From C++ you can launch anything, including scripts. I'm not sure what you mean by VC++ integration.
The generic answer would be:
save the output locations somewhere, doesn't matter where (file, registry, environment variables etc.)
retrieve them in the script before use
But depending on what you need, you could also try:
Output the same executables in the same folder structure. This way you can use relative paths.
Use a post-build event which copies the script in the output folder and make it use the relative path.
Instead of a script you can also try handling everything from the first EXE. Instead of an ANT script it could use a configuration file which specifies execution order and parameters.
I would like to control a subset of files and only allow some of them to be installed if run with a command line switch for instance.
Is this possible?
For example
if (some condition)
install full set of files
else
Install other set of files
Alternatively I can just run another installer but then I have to pass the file/path location to that second installer. There is also the issue of bundling that second installer with the first one. I think that part is not that difficult though
Yes, it is even rather easy. There are several ways to do this, all of which depend on Pascal scripting.
Method 1
You can use the GetCmdTail, ParamCount, and ParamStr functions to obtain the entire or parts of the command-line.
Then you can use the Check parameter on separate files. Hence, each file will be installed if and only if the called function returns true.