Graeme Perrow's questions

I have been using Google's OAuth 2.0 authentication on my site for years. Recently it stopped working because the redirect_uri I had used http:// and Google now apparently requires https. My site supports https so I changed my local client_secrets.json file to use https and also went to the Google Cloud Platform page on credentials, selected the OAuth 2.0 Client ID, and made sure the entry under Authorized redirect URIs had https:// as well. After restarting my web server, authentication still fails because the redirect_uri is still using http://.

Is this just a caching issue and will go away in an hour/day? Is there something else I have to do to update the redirect URI or to tell Google to refresh it?

I have sendmail running on a RedHat 7.1 machine. I want to configure it to authenticate using either PLAIN or CRAM-MD5. I have it working with PLAIN (so I know the client works, I have the username/password correct, etc.) but CRAM-MD5 always returns "535 5.7.0 authentication failed". For testing I'm using a simple python script which uses the smtplib module; I'm assuming their implementation of CRAM-MD5 is correct.

It responds to EHLO with "AUTH DIGEST-MD5 CRAM-MD5 LOGIN PLAIN" and sends a challenge when I send "AUTH CRAM-MD5", so it does accept CRAM-MD5 requests but it refuses to authenticate them.

I know that CRAM-MD5 requires the server to have access to the plaintext passwords; is it possible that it doesn't? How can I tell and if not, how can I configure the machine so that it does?

I've added the username / password into /etc/mail/authinfo and run map hash /etc/mail/authinfo < /etc/mail/authinfo. There are no error messages in the /var/log/maillog file.

I am running Apache httpd 2.4.4 on a RedHat 5 machine. I am using Django and running through mod_wsgi. I have Python 2.4.3 installed in /usr/bin and Python 2.7.6 installed in /usr/local/bin and I want to use 2.7.6. I build the 2.7.6 from source and also built mod_wsgi-3.4.

I have created a virtualenv using 2.7.6 and from a command shell, python -V tells me it's using 2.7.6 (after doing a . virtpy/bin/activate). Running import sys; sys.version shows me 2.7.6 as well, and sys.executable shows me /www/sqla/virtpy/bin/python.

When I run it from httpd, I get problems loading libraries and when I look at sys.version, I see 2.7.3. The weird thing is that sys.executable still shows me /www/sqla/virtpy/bin/python, exactly the same one as I see from the command shell.

Specifically, when python tries to load _socket, I see this error from httpd:

ImportError: /www/sqla/virtpy/lib/python2.7/lib-dynload/_socket.so: undefined symbol: _PyInt_AsInt

If I try to import _socket from a command shell, no problem. I have . /www/sqla/virtpy/bin/activate in my /etc/sysconfig/httpd and have restarted Apache with that change.

How can the same executable give different versions of python? How can I convince it to use the right one?

I would like all mail sent to a particular alias on a Linux machine to be processed by a perl script but sendmail is giving me a "Service unavailable" error and I don't understand what I'm missing.

I created my handle_email.pl script (right now, this just prints its parameters to /tmp/email.txt) and added a symbolic link to it in /etc/smrsh.

Then I added this to /etc/aliases:

mailtest: |handle_email.pl

and ran newaliases afterwards. Note that I've also tried mailtest: |/etc/smrsh/handle_email.pl which gave the same results. I've also tried copying the actual script into /etc/smrsh rather than a link, same results again.

When I send mail to the mailtest alias, the file in /tmp never gets created, and I get this in /var/log/maillog:

Nov  1 14:43:59 localhost sendmail[24839]: qA1IhwHm024839: from=<xxx@xxx>, size=284, class=0, nrcpts=1, msgid=<[email protected]>, proto=ESMTP, daemon=MTA, relay=[10.7.160.180]
Nov  1 14:43:59 localhost smrsh: uid 8: attempt to use "handle_email.pl"
Nov  1 14:43:59 localhost sendmail[24850]: qA1IhwHm024839: to=|handle_email.pl, ctladdr=<mailtest@xxx> (8/0), delay=00:00:00, xdelay=00:00:00, mailer=prog, pri=30570, dsn=5.0.0, stat=Service unavailable
Nov  1 14:43:59 localhost sendmail[24850]: qA1IhwHm024839: qA1IhxHm024850: DSN: Service unavailable
Nov  1 14:44:00 localhost sendmail[24850]: qA1IhxHm024850: to=<xxx@xxx>, delay=00:00:01, xdelay=00:00:01, mailer=esmtp, pri=31594, relay=sy-int-mx.xxx. [<IP addr>], dsn=2.0.0, stat=Sent (qA1Ii4127508 Message accepted for delivery)

I get an email message describing the failure, which says:

The original message was received at Thu, 1 Nov 2012 14:58:16 -0400
from [<ip address>]

   ----- The following addresses had permanent fatal errors -----
|/etc/smrsh/handle_email.pl
    (reason: Service unavailable)
    (expanded from: <mailtest@xxx>)

   ----- Transcript of session follows -----
smrsh: "handle_email.pl" not available for sendmail programs
554 5.0.0 Service unavailable

Why will smrsh not execute my script?

Update: Answering questions in comments.

I've copied the script to /etc/smrsh and removed the link. The file permissions are 0700 (also tried 0755) and the first line is #!/usr/local/bin/perl which is correct. There is only one file in /etc/smrsh. The ls -l output:

[main:g:64] xxx:/etc/smrsh# ls -l
total 4
-rwx------. 1 root root 281 Nov  1 16:35 handle_email.pl

I just installed Apache 2.2.17, PHP 5.2.14, and MySQL 5.1.51 on Windows XP. Apache is working and is loading PHP properly, but when I try to connect to MySQL, I get "Fatal error: Call to undefined function mysql_connect()". After some googling, I found that I need to load the php_mysql.dll extension, which requires libmysql.dll. I found the libmysql.dll file in the MySQL distribution, but php_mysql.dll is nowhere to be found.

Why is it not included with PHP or MySQL? Where can I find it?