The OSX message is [SSL: CERTIFICATE_VERIFY_FAILED] while trying to download files.
While, the Linux Python message is less descriptive.
Error downloading URL 'https://xxxx.xxx.xxx': <urlopen error unknown url type: https>.
This hopefully reduces the stream of common issues raised during installation as a result of the change of Python TLS certs to 2048 or greater.
This change relates to #133#136#140#176#6275#6548#6723#9036 … and possibly more.
Installation can be accomplished by installing the latest Python certificates (pip install certifi), symbolic linking a new python version, and relying on the python shebang line to run $ ./emsdk.py install latest.
Remove obsolete docs links.
Remove downloading of the old releases list, and bundle the file in the repo.
Also the message in update-tags was wrong: we need git to do anything there now, as update-tags gets emscripten-releases using git currently (which is ok as update-tags is run by users that installed with git; otherwise they installed without git and just update the entire emsdk normally).
Previously we were asserting and generating a stack track if a
download failed. However the assert is only relevant if the download
succeeds.
Also:
- Remove an unused function
- Remove extra braces from tuple destructuring
Without setting this, if a non-releases build was previously activated, it will keep affecting EMSCRIPTEN_ROOT in the emitted config file.
Helps with #326
This change moves the python code for emsdk into a file ending in .py.
This script is then run via emsdk.bat on windows or emsdk (a shell
script) on non-windows.
This avoid the #!/bin/sh at the top of the python script and the "exec"
hack on the first line that re-runs it under python. Hopefully this
preserves the intent of #273 without jumping through so many hoops.
The code for construct_env expects the output file to be at `argv[2]`,
but it is actually invoked in emsdk_env.{sh,fish} with $@ there
instead. Usually this is not a problem because the emdsk_env.{sh,fish}
is `sourced` directly from the user's shell and $@ is empty, but this
can break in scripted environments.
The $@ was added two years ago in 3b6c6b86 for no discernible reason,
but it was entirely ignored in the code until last month when #307
added a meaningful argument to construct_env but incorrectly kept the
$@.
This means that `./emsdk` works on UNIX system after emsdk self-updates
from a zip file. Without this one would need to run `python ./emsdk`
which seems to be why the tests were doing it this way.
