-
-
Notifications
You must be signed in to change notification settings - Fork 10.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix some CMP related msan failures #12275
Conversation
In HTTP_new_bio(), if the host has a trailing '/' we took a copy of the hostname but failed to terminate it properly.
If an empty password was supplied we ignored it and were trying to use the fallback method to read the password instead (i.e. read from stdin). However if that failed (which it always does if the cmp option -batch is used) then we were reporting that we had successfully read the password without actually setting one. Instead, if an empty password is explicitly provided we should use it. If no password is supplied explicitly and we have no fallback method then we assume the empty password. [extended tests]
976d58a
to
18d3b31
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the fixes; LGTM.
I mentioned a couple of potential coding style enhancements, but these are optional and unrelated to the given fixes.
Extended test still fail for the Travis CI build witth configuration It turns out that this is (as in #12145) due to some CMP verification steps being omitted to allow for more aggressive fuzzing in case I had already tried disabling such tests in
but this turns out to be insufficient here. Can we somehow check using Perl in that script if the configuration has been done with |
This PR is currently blocked by the following question:
|
Not sure the best way to do it, but this probably works:
|
Thanks @richsalz - the idea is good, @mattcaswell, if you add to
by
the exclusion of fuzz-related builds should work fine. |
[extended tests]
Updated with a new commit that should address the travis failures. I was able to tweak @richsalz's variant a bit more to get it to work. |
I just realised I forgot to address @DDvO's style comments. Fix up pushed addressing those. |
@mattcaswell, this is ready to merge. Shall I do? |
It's not been re-approved since the latest changes - so unfortunately not yet ready for merge. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oops, overlooked the need for renewed approval.
Still LGTM.
Unfortunately it is not clear whether all the extended tests are fixed because the fixup commit makes them not to run. |
Darn...I remembered to add "extended tests" to my earlier fixup - but not to the subsequent one. Will add shortly. |
I've restarted the Travis build on the previous commit to find out as the fixup commit should not change anything in regards to tests. |
Ah! Ok - perfect. Thanks. |
The Travis run passed. |
In HTTP_new_bio(), if the host has a trailing '/' we took a copy of the hostname but failed to terminate it properly. Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from #12275)
If an empty password was supplied we ignored it and were trying to use the fallback method to read the password instead (i.e. read from stdin). However if that failed (which it always does if the cmp option -batch is used) then we were reporting that we had successfully read the password without actually setting one. Instead, if an empty password is explicitly provided we should use it. If no password is supplied explicitly and we have no fallback method then we assume the empty password. [extended tests] Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from #12275)
[extended tests] Reviewed-by: David von Oheimb <david.von.oheimb@siemens.com> Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org> (Merged from #12275)
Pushed. Thanks. |
The cmp_cli test is failing in some travis builds on master due to msan failures. There are two separate problems fixed by this PR:
In HTTP_new_bio(), if the host has a trailing '/' we took a copy of the hostname but failed to terminate it properly.
If an empty password was supplied we ignored it and were trying to use the fallback method to read the password instead (i.e. read from stdin). However if that failed (which it always does if the cmp option -batch is used) then we were reporting that we had successfully read the password without actually setting one.
Instead, if an empty password is explicitly provided we should use it. If no password is supplied explicitly and we have no fallback method then we assume the empty password.
[extended tests]