Hi Brian, Thanks for the reply.
The only way I can see in the qpopper docs to tell it about MD5 is with --with-pam - added that configure option - set up a pam service as outlined in the qpopper install docs and ran it. The configure runs fine - but make blows up with alot of complaints about pam variables in pop_pass.c ---- -g -O2 -DHAVE_CONFIG_H -DLINUX -DUNIX pop_pass.c -o pop_pass.o pop_pass.c:377: warning: `struct pam_response' declared inside parameter list pop_pass.c:377: warning: its scope is only this definition or declaration, which is probably not what you want. pop_pass.c:377: warning: `struct pam_message' declared inside parameter list pop_pass.c: In function `PAM_qpopper_conv': pop_pass.c:385: `PAM_SYSTEM_ERR' undeclared (first use in this function) pop_pass.c:385: (Each undeclared identifier is reported only once pop_pass.c:385: for each function it appears in.) pop_pass.c:389: sizeof applied to an incomplete type pop_pass.c:391: `PAM_CONV_ERR' undeclared (first use in this function) pop_pass.c:396: dereferencing pointer to incomplete type pop_pass.c:398: `PAM_PROMPT_ECHO_ON' undeclared (first use in this function) pop_pass.c:399: arithmetic on pointer to an incomplete type pop_pass.c:399: dereferencing pointer to incomplete type pop_pass.c:399: `PAM_SUCCESS' undeclared (first use in this function) pop_pass.c:400: arithmetic on pointer to an incomplete type pop_pass.c:400: dereferencing pointer to incomplete type pop_pass.c:404: `PAM_PROMPT_ECHO_OFF' undeclared (first use in this function) pop_pass.c:405: arithmetic on pointer to an incomplete type pop_pass.c:405: dereferencing pointer to incomplete type pop_pass.c:406: arithmetic on pointer to an incomplete type pop_pass.c:406: dereferencing pointer to incomplete type pop_pass.c:410: `PAM_TEXT_INFO' undeclared (first use in this function) pop_pass.c:411: `PAM_ERROR_MSG' undeclared (first use in this function) pop_pass.c:412: arithmetic on pointer to an incomplete type pop_pass.c:412: dereferencing pointer to incomplete type pop_pass.c:413: arithmetic on pointer to an incomplete type pop_pass.c:413: dereferencing pointer to incomplete type pop_pass.c:416: default label not within a switch statement pop_pass.c: At top level: pop_pass.c:428: variable `PAM_conversation' has initializer but incomplete type pop_pass.c:429: warning: excess elements in struct initializer pop_pass.c:429: warning: (near initialization for `PAM_conversation') pop_pass.c:431: warning: excess elements in struct initializer pop_pass.c:431: warning: (near initialization for `PAM_conversation') pop_pass.c: In function `auth_user': pop_pass.c:439: `pam_handle_t' undeclared (first use in this function) pop_pass.c:439: `pamh' undeclared (first use in this function) pop_pass.c:439: invalid lvalue in assignment pop_pass.c:440: parse error before `int' pop_pass.c:447: invalid use of undefined type `struct pam_conv' pop_pass.c:449: `pamerror' undeclared (first use in this function) pop_pass.c:452: `PAM_SUCCESS' undeclared (first use in this function) pop_pass.c:455: `erc' undeclared (first use in this function) pop_pass.c:476: `errmsg' undeclared (first use in this function) pop_pass.c:491: `PAM_ESTABLISH_CRED' undeclared (first use in this function) pop_pass.c:501: `PAM_TTY' undeclared (first use in this function) pop_pass.c:509: `PAM_RHOST' undeclared (first use in this function) make[1]: *** [pop_pass.o] Error 1 make[1]: Leaving directory `/usr/local/src/qpopper4.0.4/popper' make: *** [popper_server] Error 2 --- Anybody got any ideas? Thanks Mike -----Original Message----- From: Brian C. Hill [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 21, 2002 11:32 AM To: Mike Pacheco Subject: Re: qpopper 4.0 install problems Maybe this is the md5 issue with redhat. I think you need to specify that on the configure line. Brian ====================================================================== On Wed, Aug 21, 2002 at 11:28:02AM -0400, Mike Pacheco wrote: > Hi All, > > Though I had this install down pretty good. Got a RedHat 7.1 box with > sendmail 8.11 and qpopper running fine - used the following options to > install it and it worked first try. > > ---------- > #!/bin/sh > # > # Install config switches used on first attempt > # > ./configure \ > --enable-specialauth \ > --enable-auth-file=/etc/pop/authuser \ > --enable-log-login \ > --enable-popuid=pop \ > --enable-servermode \ > --enable-shy > > make > make install > ---------- > > Now I am upgrading a seperate box to take its place - this one is RH 7.3 > with sendmail 8.12.5 - using qpopper 4.0 and the same install script - the > authfile is there - permission are OK - I even opened them up to 777 on > directory and file for testing once I started having problems. But > everytime I telnet localhost 110 and do > > -- > +OK ready > user xxx > +OK Password required for xxx > pass 123xyz > -ERR [AUTH] Password supplied for "xxx" is incorrect. > -- > > The password is OK - I've checked it and changed it, plus tried different > users - it OK's the user - just not the password - the users are in the > auth-file. I've deleted source and binary and recompiled 3 times - each > time I see > > -- > Found getspname (); using shadow passwords by default > Restricting access to only users listed in /etc/pop/authuser > -- > Plus the usual other confige words - no errors during configure, make or > make install. > > In short - I'm stumped - RH didn't change the shadow password system between > 7.1 and 7.3? Did it? I can't authenticate to pop - any ideas? > > Thanks - Mike Pacheco > -- _____________________________________________________________________ / Brian C. Hill [EMAIL PROTECTED] http://brian.bch.net \ | Unix Specialist BCH Technical Services http://www.bch.net |
