I’m trying to load minidumps generated from local uses of our Electron application into gdb.
To do so, I’ve intercepted the minidump files (*.dmp) before they are uploaded to sentry using beforeSend in the electron Sentry library.
On windows, this works great - the dmp files are located and we have had some success loading them with visual studio to get more debugging information than is available on the sentry event page.
On Linux, the minidump is also easy to find. However, when trying to use minidump-2-core to convert the file back to a core file to debug with gdb, minidump-2-core segfaults.
This might be a better suited question for electron’s crashReporter, however since sentry is able to successfully extract and symbolicate the stack trace I wanted to ask how sentry does it.
Our current approach:
- crash application
- locate new dmp file, example
/tmp/MyApp Crashes/1d63820f-da56-4904-ffad9297-5501f18a.dmp - pass that dmp file to
minidump-2-corebuilt from the chromium source tree.
minidump-2-core then segfaults.
The crash occurs in a native library loaded by the electron main process, that’s what we’re trying to debug.
This would help a lot! I realize this isn’t an issue with Sentry’s product, so totally understand if this goes unanswered.