Installation fails with corrupt files [Windows and Linux]

[GeneThomas]

The download node .zip is corrupted, using curl to download it is ok?

python -V
Python 3.11.9

git clone "https://github.com/emscripten-core/emsdk.git"
Cloning into 'emsdk'...

cd emsdk

git rev-parse HEAD
b665079

emsdk install latest

Resolving SDK alias 'latest' to '4.0.5'
Resolving SDK version '4.0.5' to 'sdk-releases-d7f8ff5e2ca3539c33fae81e98f7c56ef9fa1239-64bit'
Installing SDK 'sdk-releases-d7f8ff5e2ca3539c33fae81e98f7c56ef9fa1239-64bit'..
Installing tool 'node-20.18.0-64bit'..
Downloading: C:/Users/Pemrt/Documents/Projects/ext/emsdk/downloads/node-v20.18.0-win-x64.zip from https://storage.googleapis.com/webassembly/emscripten-releases-builds/deps/node-v20.18.0-win-x64.zip, 31379384 Bytes
Unpacking 'C:/Users/Pemrt/Documents/Projects/ext/emsdk/downloads/node-v20.18.0-win-x64.zip' to 'C:/Users/Pemrt/Documents/Projects/ext/emsdk/node/20.18.0_64bit'
Unzipping file 'C:/Users/Pemrt/Documents/Projects/ext/emsdk/downloads/node-v20.18.0-win-x64.zip' failed due to reason: File is not a zip file! Removing the corrupted zip file.
error: installation failed!

curl https://storage.googleapis.com/webassembly/emscripten-releases-builds/deps/node-v20.18.0-win-x64.zip --output downloads\node-v20.18.0-win-x64.zip

unzip /d c:\temp\node-tmp downloads\node-v20.18.0-win-x64.zip

is ok, showing that the .zip curl downloaded is ok?

[GeneThomas]

owner@owner:~/emsdk$ python3 -V
Python 3.12.3

wner@owner:~/emsdk$ ./emsdk install latest
Resolving SDK alias 'latest' to '4.0.5'
Resolving SDK version '4.0.5' to 'sdk-releases-d7f8ff5e2ca3539c33fae81e98f7c56ef9fa1239-64bit'
Installing SDK 'sdk-releases-d7f8ff5e2ca3539c33fae81e98f7c56ef9fa1239-64bit'..
Installing tool 'node-20.18.0-64bit'..
Downloading: /home/owner/emsdk/downloads/node-v20.18.0-linux-x64.tar.xz from https://storage.googleapis.com/webassembly/emscripten-releases-builds/deps/node-v20.18.0-linux-x64.tar.xz, 25738964 Bytes
Unpacking '/home/owner/emsdk/downloads/node-v20.18.0-linux-x64.tar.xz' to '/home/owner/emsdk/node/20.18.0_64bit'
xz: (stdin): Unexpected end of input
tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now
['tar', '-xf', '/home/owner/emsdk/downloads/node-v20.18.0-linux-x64.tar.xz', '--strip', '1'] failed with error code 2!
error: installation failed!

[GeneThomas]

A work around is to use curl to download the archives by hand (into the downloads/ directory) and setting EMSDK_KEEP_DOWNLOADS=1 which re-uses archives that have been downloaded already.

I have a slow internet connection, perhaps storage.googleapis.com is truncating the download and curl, used by the python to perform the downloads, does not notice!

Isn’t wasm’s loop/block structure odd!

[sbc100]

Were you able to reproduce this or was it a one off?

[GeneThomas]

Yes, I can reproduce it.

…

On Tue, 18 Mar 2025 at 05:12, Sam Clegg ***@***.***> wrote: Were you able to reproduce this or was it a one off? — Reply to this email directly, view it on GitHub <#1538 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAD5SOAPDEP4V7JAMVAPDU32U3X4VAVCNFSM6AAAAABZENYO62VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZQGA4TGNJTGY> . You are receiving this because you authored the thread.Message ID: ***@***.***> [image: sbc100]*sbc100* left a comment (emscripten-core/emsdk#1538) <#1538 (comment)> Were you able to reproduce this or was it a one off? — Reply to this email directly, view it on GitHub <#1538 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAD5SOAPDEP4V7JAMVAPDU32U3X4VAVCNFSM6AAAAABZENYO62VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZQGA4TGNJTGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

-- Gene Thomas 021 436384 https://genethomas.com ***@***.***

[sbc100]

The number of bytes read looks correct (31379384 Bytes).

Can you modify the emsdk.py script to remove the rmfile lines right below the Removing the corrupted zip file. line? Then attach take the zip file from the emsdk/downloads directory and attach it to this ticket?

[JohnMcLellan]

I have found the same issue with emsdk on both Debian And Windows 24H2 with as slow internet (don't know if that's relevant) - the problem is it does for each download - is their a way around it ?

[sbc100]

@JohnMcLellan, you can try setting EMSDK_KEEP_DOWNLOADS=1 as mentioned above.

However it would be really great to get the bottom of this issue. If anyone can reliably reproduce the issue perhaps you can you save the corrupted file and take a look at its contents? Perhaps you could attach the corrupted file here? I have no idea how/why python's urllib could result in a corrupted download.

[JohnMcLellan]

It seemed to timeout before the end of the download. I no longer want to try I was at this for 3 days - extremely reproduce-able. I used curl to download each of node, python, sdk wasm-binaries and it's working fine now on Windows 11.

[sbc100]

Very strange. If a timeout occurs then I would expect the download function to fail, and we shouldn't try to extract the archive at all.

Any anyone ever wants to help debug this and they can reproduce it locally please comment here and we can try to track it down. As it stands we don't have any way to verify/reproduce this.

[JohnMcLellan]

It turned out I was using the Git Bash version not the Git GUI even though it installed - when using Git GUI all is fine. Sorry for any confusion.

[ployt0]

python-3.13.3-0-win-amd64.zip

Here you go. Downloaded using git bash like all the above users. It does open with 7zip, but errors are reported when I try to extract all. All my downloads are corrupted using this script.

If it was me I would replace urllib with requests.

I tried another shell, CMD, and have attached that's download additionally.

python-3.13.3-0-win-amd64-using-cmd.zip