We are keeping a qmake project file in source control. Since every user can check out the project to a different path, it's causing some problems in version controlling.
For instance Every time project is updated, we have to go into project file manually and fix the paths in the file.
What is the best solution to deal with it? I am thinking about using an environment variable and assigning environment variable to variables in project file but I also wonder if there is a better approach.
The safest solution is to only use relative paths in the qmake files. That way, it doesn't matter where users check out to.
In my experience, depending on users setting environment variables causes lots of fragilities, and should be avoided at all costs, e.g.
User moves to a different build space, but forgets to update the variable, so they end up building the wrong thing
User thinks that environment variables are a good idea, so adds another environment variable that changes some other aspect of the behaviour of the build (e.g. compiler flags) and your builds end up not reproducible.
Related
I am fairly unused to compilation and building projects so pardon me if my approach to compilation and build seems a bit odd. Any tip is welcome.
I am currently working on a 3D geometry C++ project (Which is a dll). This project uses external dll's such as BOOST. So when building the project I have to define the directories in which the .dll, .lib and .h/.hpp files are.
Currently I am using scons to build the project and define those paths straight in the SConstruct file.
However those paths are later reused for other build operations. (In the present case compiling the C++ code in MEX for matlab but that's not really the point here).
Thus I currently have to define the same path in different places which is inefficient. In addition the project has to be easy to set up for another user. So having to change and update path in many different files is something I would like to avoid.
From where I stand I see two alternatives:
First I could ask the user to define environment variables and read them from inside my various build scripts. However I am not really satisfied with this solution for it asks the user for additional manipulation and, as far as I've understand, I lose the cross-platform portability that scons offers. (I know it might still possible but requires some extra steps and I'd like to keep things as simlpe as possible)
Second I could define all my path in a single .txt (or something similar) file at the root of my project and read it from my various build scripts. However this makes the process sensitive to typos and parsing errors which is not really to my taste.
So my question is the following:
Is there a better way or good practice to have the user input the paths necessary for compilation that satisfies the following:
Has the user only input once every path
Is done within the project folder through a file or something else
Is as foolproof as possible
Does not require too much additional download/installing (I don't really want to have the user install a brand new software for this. However I'm fine with something like a simple light .exe that I can add in my project files)
SCons's Variables are likely your best choice here.
See: https://scons.org/doc/production/HTML/scons-user.html#sect-command-line-variables
It allows reading defaults from a file:
vars = Variables('custom.py')
You'd have to craft some logic to save any variables specified on the command line.
I have developed an application that I plan to deploy on Windows, Mac, and Linux. The program requires access to some files (scripts and the like) at run-time.
The installation process should install the files to a location that my application can later determine without user-intervention (or perhaps just a prompt allowing the user to change the location, if desired).
What is the best way to achieve this? I can't seem to find any way to:
1. Use a "standardized path" variable in the project file's INSTALLS statement. (e.g., my application could use QStandardPaths to initialize the location, but I can't figure out how to access this path from the INSTALLS statement)
2. Save the path to my project's QSettings (.plist, registry, whatever) for later retrieval
That leaves me with creating a custom project file and INSTALLS command for each environment, and then I still can't install to the user's directory because I don't know the user's name when I deploy the make command. It seems as if there must be a better way, but I can't seem to find any documentation for this. Am I just using the wrong keywords in my searches? Thanks in advance!
What standard directory? What type of getting that standard directory?
For instance, you can put such thing in your windows branch of .pro file:
win32 {
APPDATA_DIR = $$system(echo %APPDATA%) # should be %LOCALAPPDATA% as requested
message($$APPDATA_DIR)
}
Just unsure of what exact kind of standartized path you are talking about. QStandardPaths knows many. It makes sense to be more concrete to find the correspondence with concrete OS.
Also somewhat relative reply on mine, on how to check the correspondence with certain variable, etc: Qt .pro file - how to add conditioning on OSX version?
Maybe this class will help you
QStandardPaths documentation
But your problem is still little bit unclear for me.
Build systems like make use timestamp check if a dependency is changed during two build. Here are a few common issue I run into with timestamp
Open a file, make some change, but later, I decided it is not good. I revert the change, for example, git checkout -- file if I am using git for the project.
Open a file, I just accidentally hit keyboard shortcut for save of the editor
Either way, the file's timestamp is changed. If now I want to build the project, everything depending on that file need to be rebuilt. This often means the whole project.
Is there anyway around these issues? For example, a build system using a version control for checking file change, preferably git. Or any other solutions to the above issues are welcome.
Many thanks in advance.
SCons uses checksums, not timestamps by default. However, checksumming requires reading all the contents of all the source files on disk, and that is much slower than simply reading directory entries, which is why most build systems use timestamps.
Software Build Systems gives a good overview of these issues.
I'm not sure if this is too general, so if it is I'll say that I'm on Linux using qmake, but I'd like to be able to switch from Linux to Windows with my project whenever I need to, as well as, possibly other PCs.
In order to do this, I'd like to know how some of the programmers on here have gotten around using resource directories without using absolute path definitions. With Qt, it seems like the runtime working directory is the build directory of the application, and not the source directory.
Ideally, I think the best solution would be to somehow get the Resource directory as it resides in the source directory and copy that to the relative build directory (i.e., Debug or Release, depending on development stage) so that the application can access that via run time.
This can introduce some complication, however (at least, I think it can).
Anyway, what would be a good solution to do this?
If you are using Qt. I would suggest using deploy process.
http://doc.qt.digia.com/qtcreator/creator-building-running.html
Basically, you just need to declare which directories need to be copied.
The qt creator will copy those dirs to build dir(release/debug) after build process is done.Then you simply run the executable.
Here is one of example.
https://github.com/longwei/incubator-cordova-qt.
in the pro file
wwwDir.source = www
xmlDir.source = xml
qmlDir.source = qml
DEPLOYMENTFOLDERS = wwwDir xmlDir qmlDir
second
include(deployment.pri)
qtcAddDeployment()
then it is done.
Its not clear what exactly you're trying to achieve, but perhaps a simple solution would be for the build scripts to pass the necessary path via a compilation definition (-D with gcc). Then depending on if its a Debug, Release, etc build, the definition would be set accordingly, then the corresponding binary would have the correct path.
As a side note, I tried qmake for a while, but found SCons to be much more versatile.
Is it possible to combine the following properties, and if so, how?
Store in our version control system some Visual Studio 2008 native C++ (VCPROJ) project files for the developers in our team that use this IDE.
Allow some of those developers to tweak their projects (e.g. using debug version of third-party libraries instead of the usual ones).
Make sure these modifications are done in files that are not versioned.
In other words, I would like to allow developers to tweak some settings in their projects without risking that these changes are committed.
An 'optional VSPROP' file approach seems doomed to fail, as VS2008 refuses to load projects that refer to non-existent VSPROP files...
Any other suggestion? Is this possible with VS2010?
You may not be able to do this but using a solution that generates the vcproj like CMake for example would let you do this. Scripts all your project with CMake and literally conditionally include a config file(if present for example) that developers can change on their setup.
Branches could solve this problem: you create a branch, play with different versions of third-party, merge changes to trunk if results are good.
Well, as a preliminary solution you could put the project file into something like .hgignore or .gitignore after its initial commit.
This way changes to it can't be done accidentally.
At least that's how I handle .hgignore itself.
We use a versionned "common_configuration" folder, and a script which copies project files from this "common_configuration" folder towards the "project" folder.
We have another script to copy the configuration backwards, so the developpers need to make a conscious action to commit their local changes to the global version control system.
It answers partly your needs :
The upside : we have a way to keep a common configuration for everyone, and no accidental committing of local configuration
The downside : blindly copying the files actually crushes local changes. We live with it. We could write some more clever merger tool (using diff, or xml specific manipulations), but don't want to spend to much time on supporting the deployment tools.