I'd like to see the command-line arguments that FlexBuilder is using to compile my application. This is so that I can build them into the ANT script I'm working on. Is there any way to view the command line compilation step?
The reason I'm asking for this is that when I compile my app using Ant/Flex SDK vs. FlexBuilder my app behaves differently.
So I figured out the answer.
First of all, you can get a better idea of how FlexBuilder compiles your app by adding a -dump-config=C:\myConfig.xml to the compile arguments in FlexBuilder. This outputs an xml file containing configuration settings used in the compilation step. You can also use this file as an argument to compc or mxmlc if you'd like. Read more about it
here...
But, here's what actually solved my problem. I was using the regular old Flex SDK installed on our integration server to compile my apps using Ant. This is the free SDK you can download from Adobe's site. I then took the FlexBuilder directory from my local machine and copied it up to the integration server and pointed my build script to use that version of the SDK (and changed my path env variable).
When I compiled using the FlexBuilder version of the SDK, all was well and the strange bugs I was seeing in my app disappeared.
Moral of the story, make sure you are using the same version of the SDK for your automated builds as you use to build locally.
Related
I'm working on a cross-platform command line app that converts a user's text files into HTML output (similar to Markdown), and I want to provide the users an option to customize the output style with SASS/Less. Is there any CSS preprocessor/compiler library out there that I can use to compile stylesheets at runtime?
LibSASS requires installing things on Linux, which I'm trying to avoid because this needs to be a simple "one-click, run anywhere" utility without any dependencies apart from .NET Core. If I remember correctly, it also had conflicts when trying to import multiple platform NuGet packages in the same project.
I started with dotless but I've already stumbled on like 3 serious bugs and that project hasn't been updated since 2018 so I don't think they're getting fixed anytime soon. And I don't want to tie my project to a dead library for a pretty central functionality.
Looks like SharpScss is what I'm looking for. By adding
<PropertyGroup>
<RuntimeIdentifiers>win-x86;win-x64;linux-x64;osx-x64</RuntimeIdentifiers>
</PropertyGroup>
to the project file, the SASS parsing runs in both Windows and Linux (haven't tested on OSX) without having to install or configure anything.
We plan on using SASS instead of plain CSS for our SharePoint project very soon. While testing and trying to set everything up, I ran into some problems:
We're using Visual Studio 2015 and on my developer machine I installed the Web Compiler Extension to compile the .scss-files and partial files to a regular .css-file.
That worked very nicely but the problem is, that there will be a few developers working simultaneously on the styles. I want to avoid merging the resulting css-file each time someone tries to check in something into source control (we're using Team Foundation Server).
Since there is a build running every time someone is checking in their changes, and to deploy the resulting solution to the nightly build machine, the idea was to somehow include the SASS compiler in the build definition. This way the more readable scss-files get merged and the build creates the resulting css-file to include it in the solution.
Maybe I'm thinking too complicated, but I just couldn't get that to work so far.
Any ideas how I can achieve that?
(Maybe I should also mention that none of the dev machines got any internet connection)
If you're building an MVC app, you can use MVC's bundling feature along with the SASS NuGet package. And, be sure to enable minification. There's a UseNativeMinification property on SassAndScssSettings. That way you don't need to deal with merging the css file when you get latest or check in. Reference this thread: SASS/TFS best practice
Another way is running a script (e.g with PowerShell task) on the server that to install the gulp components and then call the sass compile task to compile the SASS. Refer to Powershell build - compiling SASS for details.
I am using Visual Studio 15.9.14 and Qt 5.13.0 on multiple machines. When I check out my source from version control:
If I open the VS IDE to build my solutions, everything compiles and links correctly.
If I build the solutions from the command line using devenv.exe, there are multiple compile and link errors in the Qt projects.
The problem is that when building from the command line, the .vcxproj.user files are NOT generated, and therefore $QTDIR is not defined for use in my projects. The result is my automated/nightly builds fail.
I can build a tool to create the files and integrate it into my build process, but I shouldn't have to. This problem seems to be related to the VS/Qt integration. I have also encountered a similar problem in the IDE where I had to force the files to regenerate by touching the projects.
Any suggestions/help would be appreciated.
I would like to deploy for the first time my first app made with PyQt5. I found pyqtdeploy (link) to be probably what I need to do this for Windows and other platforms.
Anyway, I am following this tutorial, but, as I am completely new to deploying applications, I am not sure on one thing: do I need to download Qt for the target specific platform I need to deploy my app to? If so, am I supposed to use qmake from that Qt installation against the result of pyqtdeploy? For example: if I want to deploy to Windows, do I have to download Qt from here (and which version???)?
From the same tutorial page, this is not so clear, maybe it's obvious, but as I am a newbie in this, not for me.
No you don't need to download Qt.
When building your 'sysroot.toml' file, pyqtdeploy will download qt-everywhere-src and builds it from source.
In the built sysroot directory you can find qmake executable in "...\Qt\bin" (qmake.exe) depends on where your project files (sysroot.toml file) are.
For example if your host machine is Windows and C:\Users\username\Desktop\My_PyQtDeploy_Proj is your project folder which contains sysroot.toml, after building sysroot you can find qmake in "C:\Users\username\Desktop\My_PyQtDeploy_Proj\sysroot-win-64\Qt\bin\qmake.exe"
I've been googling for a solution to this issue and although I've found many people sharing my problem none of their solutions work for me.
I wrote a C++ application using Qt framework using Visual Studio 2010. I built and ran the application in "Release" mode from Visual Studio without issue, but when I copy that exe from the Release folder to a new destination (pretend its a new PC) it fails to run providing this error:
---------------------------
TestApplication
---------------------------
This application failed to start because it could not find or load the Qt platform plugin "windows".
Available platform plugins are: windows.
Reinstalling the application may fix this problem.
Within the executable directory I have the following file structure:
./TestApplication.exe
./libGLESv2.dll
./Qt5Core.dll
./Qt5Gui.dll
./QtWidgets.dll
./platforms/qwindows.dll
./qt.conf
All dll files were taken from the 5.0.0 build of Qt in the Qtbase folder where the libraries reside.
The qt.conf file is:
[Paths]
Plugins=.
Without it, the launch error is the same except it says "minimal" and "offscreen" are available platforms as well.
For all other people experiencing this error it seemed to be solved by creating the platforms folder and putting in the qwindows.dll. But doing that myself doesn't change any behavior.
Is there something I've done wrong? Perhaps my method of generating the .exe in the first place is wrong?
If you are using libGLESv2.dll, then you must include libEGL.dll, too.
You can't see that in depends.exe, don't know how the Qt developers managed to hide this.
If your Qt is out of the box, then both Dlls are necessary even if you are not using OpenGL.
Also, if your Qt is out of the box, you need to include also the three ic*.dll, which contain information for Unicode handling.
You can see which DLLs are needed by looking at which ones are invoked when running Debug (F5) in Qt Creator.
=Carl
The release is likely missing a library/plugin or the library is in the wrong directory and or from the wrong directory.
Qt intended answer: Use windeployqt.
Qt comes with platform console applications that will add all dependencies (including ones like qwindows.dll and libEGL.dll) into the folder of your deployed executable. This is the intended way to deploy your application, so you do not miss any libraries (which is the main issue with all of these answers). The application for windows is called windeployqt. There is likely a deployment console app for each OS.