[Bug 59163] New: Adding debug channel +crypt makes Total Commander behave different/fail.
http://bugs.winehq.org/show_bug.cgi?id=59163 Bug ID: 59163 Summary: Adding debug channel +crypt makes Total Commander behave different/fail. Product: Wine Version: 11.0-rc3 Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: crypt32 Assignee: wine-bugs@list.winehq.org Reporter: bernhardu@mailbox.org Distribution: --- When trying to further investigate bug 56887 I wanted to add a debug log with +crypt. But then it fails before that bugs orignal issue. I could track it down to the channel +crypt calling `gnutls_global_set_log_level` with a value of equal or above 2. Unfortunately +crypt sets it by default to 4. This different behaviour can be avoided by additionally using the environment GNUTLS_DEBUG_LEVEL set to a value 1 or 0. As this difference seems to happen inside gnutls it may be important to note this happens inside a Debian 13 Trixie VM. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 --- Comment #1 from Bernhard Übelacker <bernhardu@mailbox.org> --- Created attachment 80018 --> http://bugs.winehq.org/attachment.cgi?id=80018 instructions to setup and replicate the issue; one failing run; one succeeding run; This attachement contains a failing +crypt run with GNUTLS_DEBUG_LEVEL=2, which immediately shows the "Connect failed!" message. And the top of a "succeeding" +crypt run with GNUTLS_DEBUG_LEVEL=1, which asks about confirmation about the self signed certificate. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 Bernhard Übelacker <bernhardu@mailbox.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Distribution|--- |Debian URL| |https://totalcommander.ch/1 | |103/tcmd1103x64.exe Keywords| |download -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 --- Comment #2 from Bernhard Übelacker <bernhardu@mailbox.org> --- Created attachment 80021 --> http://bugs.winehq.org/attachment.cgi?id=80021 +crypt_rr-debugging I tried to debug this further and found the failure seen with +crypt alone disappears when using +crypt,+secur32. And when I attached a debugger I found a unix-side segfault, because the logging function set by gnutls_global_set_log_function seems no longer in memory. This seems to be a result of several loads and unloads of dll's and so's in CRYPT_LoadProvider. Unfortunately it looks like libgnutls.so itself doesn't get unloaded, therefore the log function pointer stays still set. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 --- Comment #3 from Bernhard Übelacker <bernhardu@mailbox.org> --- Submitted this merge request to set the logging function inside gnutls to NULL, when the dll gets unloaded: https://gitlab.winehq.org/wine/wine/-/merge_requests/9821 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 Bernhard Übelacker <bernhardu@mailbox.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |c2f2877c933619aedaae6cab14b | |8312f97b92b69 --- Comment #4 from Bernhard Übelacker <bernhardu@mailbox.org> --- The three patches, setting the logging function to null on dll unload, got committed. Marking resolved fixed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 Gijs Vermeulen <gijsvrm@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |c2f2877c933619aedaae6cab14b | |8312f97b92b69 Regression SHA1|c2f2877c933619aedaae6cab14b | |8312f97b92b69 | Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #5 from Gijs Vermeulen <gijsvrm@gmail.com> --- Resolving. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=59163 Alexandre Julliard <julliard@winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #6 from Alexandre Julliard <julliard@winehq.org> --- Closing bugs fixed in 11.0-rc4. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (1)
-
WineHQ Bugzilla