[Bug 64610] New: apachectl -t fails to report variable syntax error
https://bz.apache.org/bugzilla/show_bug.cgi?id=64610 Bug ID: 64610 Summary: apachectl -t fails to report variable syntax error Product: Apache httpd-2 Version: 2.4.43 Hardware: PC OS: Mac OS X 10.1 Status: NEW Severity: normal Priority: P2 Component: All Assignee: bugs@httpd.apache.org Reporter: yumatechni...@gmail.com Target Milestone: --- For ages my Apache web server wound not work properly. Details: I spent a while configuring and installing web applications. Like I’ve done on several other machines. It was all working at this time. Then I decided to use variables rather than hard-coded values. Check: When problems first arrived, I ran “apachectl -t”, to test the syntax - it said Syntax OK. However each page that tried to visit (on my local machine) returned 403 Forbidden. Due to other projects I ignored this and decided to fix it later. (This also gave me space in the logs - so I could distinguish this issue.) The logs just said the "client denied by server configuration: ” then said my document root. This is quite general, but it indicated the problem was document root. So I looked for a .htaccess file - was none, so I created an empty one - no difference. Using trial and error, I swapped each variable with hard-coded values (reloading Apache, and testing each time). Fixed: Finally it worked after changing the “
[Bug 61355] DirectorySlash directive should use protocol in X-Forwarded-Proto header when available
https://bz.apache.org/bugzilla/show_bug.cgi?id=61355 --- Comment #12 from Eric Covener --- (In reply to Cyril B. from comment #11) > I've used the patch in production on 2.4.43 and had random Segmentation > faults. Fixed by replacing: > > +r->server->server_scheme = scheme; > > with: > > +r->server->server_scheme = "https"; This is a good catch, but I think both approaches are incorrect. It is not valid for a per-request header to change the state of anything like r->-server->* . In your case it will probably not lead to incorrect results since it does not actually vary per-request, but then you should also just use a scheme in the ServerName directive if it does not vary. I think the approach that fits w/ the Apache API is for something like mod_remoteip to implement the http_scheme callback to override the server scheme. Or, mod_dir should stop calling ap_construct_url and just return non fully qualified URL's and let the browser sort it out. This was not spec in the original HTTP/1.1 RFC but was in use forever and is valid now. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
[Bug 61355] DirectorySlash directive should use protocol in X-Forwarded-Proto header when available
https://bz.apache.org/bugzilla/show_bug.cgi?id=61355 Cyril B. changed: What|Removed |Added CC||c...@excellency.fr --- Comment #11 from Cyril B. --- I've used the patch in production on 2.4.43 and had random Segmentation faults. Fixed by replacing: +r->server->server_scheme = scheme; with: +r->server->server_scheme = "https"; -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
