Update sentry react native to 0.36.0

[byCedric]

We have an issue where the @sentry/cli package is incompatible with the standard Docker linux distro Alpine. In the latest version of this cli package they added an option to use a custom CDN for prebuilt binaries. With this we can work around the incompatibility issue. It might be wise to just update to the latest version to include this ability and other fixes from the repo.

[byCedric]

There has been some progress on the incompatible issue regarding @sentry/cli on Alpine distributions. Binaries from the official repo have been patched starting from >=1.31.0 and it looks like the issue is resolved. Can we get this merged asap to have sentry-expo work out-of-the-box on Alpine?

[brentvatne]

hey @byCedric - you can give this a try by installing sentry-expo@1.9.0 explicitly. I didn't make that latest yet on npm because I don't have time to test it over the weekend, but wanted to get it out there so you could use it.

[byCedric]

You are the best @brentvatne! I will try that version tomorrow, thanks!

[byCedric]

@brentvatne works like a charm!

[serhiipalash]

Hi @byCedric !
How did you make it work? I use your image bycedric/ci-expo and set sentry-expo@1.9.0 in package.json, but I still get error

[byCedric]

I approached @serhiipalash on the Expo Slack to get the issue resolved. I could vaguely remember something like this happening in one of our projects too. Turns out it's an issue in Alpine combined with Docker Node, as Bitbucket Support described below.

For all future crawlers, you have to unset NPM_CONFIG_USER in your Bitbucket Pipeline before executing npm install or npm ci.

Here is the full post, with the name of the supporter removed due to privacy.

Hey Cedric,
Thank you very much for contacting Atlassian Support!

While we are working on performance upgrades to improve build performance (specifically to decrease the build time variation), we notice a bug in Alpine which causes widespread issues for npm users.
Hence, we've set an environment variable called NPM_CONFIG_USER which seems to be causing these permission errors when reading node_modules (as we set the config user as "nobody) to work as a workaround till the bug is resolved.

We have Github issues open tracking these bugs:
https://github.com/nodejs/docker-node/issues/813
https://github.com/npm/npm/issues/20861

In the meanwhile, we would suggest running the following command at the start of your build as a workaround till the bug is resolved:

unset NPM_CONFIG_USER
We'll be investigating additional ways to prevent having to use this workaround, but for the time being, we recommend this solution to proceed further.

Thank you!

Regards, 
<redacted>
Atlassian Support