4 ms·

You're referencing a bash script that is run during the Github CI build: https://github.com/audacity/audacity/blob/407be925e5226ddb49441f2815e1ff19e11df57c/.gi

You're referencing a bash script that is run during the Github CI build: https://github.com/audacity/audacity/blob/407be925e5226ddb49441f2815e1ff19e11df57c/.github/workflows/cmake_build.yml https://github.com/audacity/audacity/blob/407be925e5226ddb49... This is not for crash reporting.

viraptor 5y ago
I'm not sure why you say it's not for crash reporting. Sentry can be used both is CI and in the distributed application.
- orf 5y ago
  > I'm not sure why you say it's not for crash reporting. Because, as the comment you’re replying to said, that bash script is only used in the CI process. That’s also completely obvious from only a cursory glance at the script itself.
  - viraptor 5y ago
    But it can be used for error/crash reporting in CI.
    - orf 5y ago
      No, it cant. Read the damn file you linked. It’s a script that uploads release files to Sentry so it can be used while correlating crash reports, when a release is made. And nobody in their right mind would ever send crash reports to Sentry from a CI. That’s completely nonsensical.
      - ljm 5y ago
        It would be an expensive way to report on a test suite, I'll give you that.
      - viraptor 5y ago
        It depends on what you're trying to achieve. It's totally reasonable to report stack traces from the CI, especially from the master branch for a quick overview on flakey tests. I don't know anything about their setup specifically, but that approach does have use cases.
        orf 5y ago
        It’s not “totally reasonable” to send CI stack traces to Sentry. It’s nonsensical for a number of reasons.
        viraptor 5y ago
        For me it solves the issue of tracking flakey tests. It's the easiest way to do it if you're already using the same error reporting platform in production. Please consider others' use cases before declaring this nonsensical.
      - untitaker_ 5y ago
        viraptor was not really responding to the discussion about the shell script at all (which is weird in its own right), but was rather talking about stuff like https://github.com/getsentry/pytest-sentry https://github.com/getsentry/pytest-sentry
aspaceman 5y ago
LMAO. But reading code is ,hard, especially when you want to contribute to FUD.

viraptor 5y ago

I'm not sure why you say it's not for crash reporting. Sentry can be used both is CI and in the distributed application.

orf 5y ago
> I'm not sure why you say it's not for crash reporting. Because, as the comment you’re replying to said, that bash script is only used in the CI process. That’s also completely obvious from only a cursory glance at the script itself.
- viraptor 5y ago
  But it can be used for error/crash reporting in CI.
  - orf 5y ago
    No, it cant. Read the damn file you linked. It’s a script that uploads release files to Sentry so it can be used while correlating crash reports, when a release is made. And nobody in their right mind would ever send crash reports to Sentry from a CI. That’s completely nonsensical.
    - ljm 5y ago
      It would be an expensive way to report on a test suite, I'll give you that.
    - viraptor 5y ago
      It depends on what you're trying to achieve. It's totally reasonable to report stack traces from the CI, especially from the master branch for a quick overview on flakey tests. I don't know anything about their setup specifically, but that approach does have use cases.
      - orf 5y ago
        It’s not “totally reasonable” to send CI stack traces to Sentry. It’s nonsensical for a number of reasons.
        viraptor 5y ago
        For me it solves the issue of tracking flakey tests. It's the easiest way to do it if you're already using the same error reporting platform in production. Please consider others' use cases before declaring this nonsensical.
    - untitaker_ 5y ago
      viraptor was not really responding to the discussion about the shell script at all (which is weird in its own right), but was rather talking about stuff like https://github.com/getsentry/pytest-sentry https://github.com/getsentry/pytest-sentry

aspaceman 5y ago

LMAO. But reading code is ,hard, especially when you want to contribute to FUD.