<div dir="auto"> From a machine with internet, you can create a source-bundle, see Cerbero help and readme for details. This bundle exist for distributing the result in a LGPL compliant way. It can be used to build without internet access.<div dir="auto"> </div><div dir="auto">Nicolas</div></div><div class="gmail_extra"> <div class="gmail_quote">Le 30 juin 2017 11:59 AM, "Maurer, Martin" <<a href="mailto:martin2.maurer@zeiss.com">martin2.maurer@zeiss.com</a>> a écrit : <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> <div lang="DE" link="#0563C1" vlink="#954F72"> <div class="m_1362792297831668936WordSection1"> Hello, we want to compile Gstreamer for Windows on a build server without (!) a connection to the Internet. We can pre-download packages from Internet and store it somewhere on our own server. With python2.exe ./cerbero-uninstalled -c config/win64.cbc fetch --print-only I seem to be able to create a list of needed source code packages and URLs, which I can pre-download and store. But it looks like this is only a small part of the way to a solutions. Could it be possible to use a “cache directory” to avoid downloads on each compile? But there is e.g. bootstrap download intltool version 0.40 (later version 0.51 is used), but I am not able to switch off or fake the download. Also the packages which are downloaded are distributed to different directories. Some into MinGW installation, some in temp directory and named download.zip, some into build directory, were I assumed to get them. There are packages like the ones in build\sources\local, but there is also build\mingw\w64 where mingw-w64-gcc-4.7.3-windows-x86_64.tar.xz is stored. There is “wipe” command, but it waits for a “yes”, even twice. Is there a possibility to run it without asking the user, so you can run it via a batch file? Each package is compiled via different steps: fetch, extract, configure, compile, install, post_install, gen_library_file Can I do the fetch step of all packages in one go? Then remove the computer from Internet and do a python2.exe ./cerbero-uninstalled -c config/win64.cbc package gstreamer-1.0 to do the remaining work? (I already tried to patch some *.recipe to skip fetch and extract, which seems to work, perhaps a command line switch would be helpful, to enable/disable each step if not already available?) Another idea: Patching each URL/location to not connect anymore to servers in the Internet, but local ones. But I think a lot places in source code (and on the fly?) to change and perhaps changes from release to release. Has someone a build server setup and running for compiling Gstreamer for Windows? Is it somehow possible at all? Best regards, Martin </div> </div> _______________________________________________ gstreamer-devel mailing list <a href="mailto:gstreamer-devel@lists.freedesktop.org">gstreamer-devel@lists.freedesktop.org</a> <a href="https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" rel="noreferrer" target="_blank">https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a> </blockquote></div></div>