| Summary: | Add support for "catch throw" in backtrace generation | ||
|---|---|---|---|
| Product: | [Applications] drkonqi | Reporter: | Christoph Feck <cfeck> |
| Component: | backtraceparsing | Assignee: | George Kiagiadakis <mail> |
| Status: | RESOLVED FIXED | ||
| Severity: | wishlist | CC: | andresbajotierra, nate |
| Priority: | NOR | ||
| Version First Reported In: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Compiled Sources | ||
| OS: | Linux | ||
| Latest Commit: | https://invent.kde.org/plasma/drkonqi/-/commit/02c03ee13f6bb27e15d82cd363c9f83bbbc566b8 | Version Fixed/Implemented In: | |
| Sentry Crash Report: | |||
|
Description
Christoph Feck
2011-11-24 12:15:43 UTC
I also would like to have it, but is that really possible for drkonqi, which asks gdb to attach that process only after the crash/exception has happen? I think that is only possible when the applicaition is run by gdb and the those extra instructions are given before the crash/exception happens. The best that can be done, I think, is Dr.Konqi should detect that the backtrace implies uncaught exception, then instead of telling users to report the crash it should inform users : "Unfortunately, this backtrace is basically of no use due to uncaught exception in the code (not your fault). If you can reproduce the crash every time, please follow the guide at http://techbase.kde.org/some-guiding-page and use gdb to get a better backtrace, then report it to bugs.kde.org manually. Thanks for your help . " I was rather thinking about the "Restart Application" button. Maybe it could be possible to restart it with said option, then on next crash, gdb could show a better backtrace. Thank you for the bug report. As this report hasn't seen any changes in 5 years or more, we ask if you can please confirm that the issue still persists. If this bug is no longer persisting or relevant please change the status to resolved. A possibly relevant merge request was started @ https://invent.kde.org/frameworks/kcrash/-/merge_requests/53 A possibly relevant merge request was started @ https://invent.kde.org/plasma/drkonqi/-/merge_requests/205 Git commit 1cfbe46675cb5e5cab18ac70b3c09edd14e22445 by Harald Sitter. Committed on 08/12/2023 at 05:59. Pushed by sitter into branch 'master'. detect active exceptions and put them in the metadata technically very straight forward except for a needed refactor of how the INI writing works: instead of snprintfing into a preallocated string we now write out the line on the fly. this allows us to make sure value doesn't contain \n that would break the ini format. when a \n appears we'll write it out as string rather than escape sequence. M +3 -3 autotests/metadatatest.cpp M +3 -0 src/CMakeLists.txt A +57 -0 src/exception.cpp [License: LGPL(v2.0+)] A +19 -0 src/exception_p.h [License: LGPL(v2.0+)] M +10 -0 src/kcrash.cpp M +19 -9 src/metadata.cpp M +1 -2 src/metadata_p.h https://invent.kde.org/frameworks/kcrash/-/commit/1cfbe46675cb5e5cab18ac70b3c09edd14e22445 Git commit 02c03ee13f6bb27e15d82cd363c9f83bbbc566b8 by Harald Sitter. Committed on 08/12/2023 at 06:46. Pushed by sitter into branch 'master'. handle exception data kcrash can now forward exception metadata, let's handle them. we route the name through c++filt since it will likely be mangled M +20 -0 src/bugzillaintegration/reportinterface.cpp M +2 -0 src/drkonqi.h M +19 -1 src/main.cpp https://invent.kde.org/plasma/drkonqi/-/commit/02c03ee13f6bb27e15d82cd363c9f83bbbc566b8 |