Re: Tru64 and C2
At 06:37 PM 12/3/2007 -0800, Randall Gellens wrote: At 5:18 PM -0600 12/3/07, J Springer wrote: Randall Gellens wrote: Hmmm. I think I see the problem. There is some confusion in the script with no and no_ for pam, and only for pam. Problem ... ...snip... ... confirmation, if you don't mind, could you run the attached ./configure script (after another 'make realclean') and send me the output? Holler when you're ready for me to run another test. I have one [dumb] question, if configure fails and a Makefile isn't created, shouldn't 'make realclean' fail? Or do I need to copy the Makefile.in to Makefile first? If ./configure fails, then there is no Makefile, and hence 'make realclea' will fail. That's fine, as there shouldn't be a config.h or any of the other files created by ./configure. I get in the habit of recommending 'make realclean' just to be sure that the stuff isn't there. So, no need to copy a Makefile. ./configure --enable-temp-drop-dir=/usr/poptmp --enable-log-login --enable-server-mode --enable-standalone --enable-specialauth use_pam is no_ checking for dlopen in -ldl... no checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam OK, that confirms what I found the first time. The second ./configure I sent does indeed fix what I found the first time, which is why pam was enabled by default when it should be off by default. At 6:21 PM -0600 12/3/07, Vic Summerour wrote: At 03:47 PM 12/2/2007 -0800, Randall Gellens wrote: At 4:10 PM -0600 12/2/07, Vic Summerour wrote: I have used qpopper with Tru64, Digital Unix, OSF/1 for years. Sometime in the ... ...snip... ./configure --enable-specialauth --without-pam Was --enable-specialauth needed? If you left it off, what happened? With ./configure only, the following occurs: ./configure . . . checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam This is fixed by the first change to ./configure. With ./configure --without-pam, it will configure and compile correctly. Attempts to retrieve pop mail results in an invalid password message: With ./configure --enable-specialauth --without-pam, it will configure correctly, compile, and work. I believe this is because TRUE64 uses getprpwnam() instead of getspnam(), and ./configure was only checking for the later when deciding if specialauth should be enabled by default. I've added a check for the former, so hopefully a default ./configure will now correctly check passwords on True64. If you could verify this for me, I'd be very grateful. I'm attaching a new ./configure. Please be sure to do a 'make realclean' beforehand. Using the original qpopper4.1a5 distribution with the new configure supplied by Randall, the following occurs when ./configure is done. snip checking if ar chokes on -r... no checking for getspnam... no checking for getprpwnam... no checking for getspnam in -lshadow... no Will NOT compile in debugging code System name: OSF1; System version: V5.1 Set OS_DEFS = -DOSF1 -DUNIX snip checking for set_auth_parameters in -lsecurity... yes use_pam is _default_ checking for dlopen in -ldl... no checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam When using the original qpopper4.1a5 distribution with the new configure, the ./configure --without-pam results in: snip checking if ar chokes on -r... no checking for getspnam... no checking for getprpwnam... no checking for getspnam in -lshadow... no Will NOT compile in debugging code System name: OSF1; System version: V5.1 Set OS_DEFS = -DOSF1 -DUNIX snip checking for set_auth_parameters in -lsecurity... yes use_pam is no checking for dlopen in -ldl... no checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam Here's the configure results when new configure is used with the qpopper4.1a5 distrubution that contains patches supplied by Hujimu Umemoto. snip checking if ar chokes on -r... no checking for getspnam... no checking for getprpwnam... no checking for getspnam in -lshadow... no Will NOT compile in debugging code System name: OSF1; System version: V5.1 Set OS_DEFS = -DOSF1 -DUNIX snip checking for set_auth_parameters in -lsecurity... yes use_pam is _default_ checking for dlopen in -ldl... no checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam I am not sure why configure is not finding getprpwnam. This system is definitely configured for C2 security. If you want, I can supply config.log files. Vic Summerour
Re: Qpopper and Tru64 (was Re: Tru64 and C2)
Hi, On Mon, 3 Dec 2007 18:50:22 -0800 Randall Gellens [EMAIL PROTECTED] said: randy At 6:21 PM -0600 12/3/07, Vic Summerour wrote: Currently with Tru64 V5.1A with PK6, I have successfully compiled and run 4.1a4 with the following compiler: alpha cc -V Compaq C V6.4-014 on Compaq Tru64 UNIX V5.1A (Rev. 1885) Compiler Driver V6.4-215 (sys) cc Driver Unfortunately, 4.1a5 will not successfully compile based on the same configure. It stop with the error: cc -c -I.. -I.. -I. \ -I../mmangle -I../common \ -g -DHAVE_CONFIG_H -DOSF1 -DUNIX main.c -o main.o cc: Error: main.c, line 150: Missing type specifier or type qualifier. (missingtype) socklen_t ai_addrlen; ^ make[1]: *** [main.o] Error 1 make[1]: Leaving directory `/dsk1/vic/qpopper/qpopper4.1a5/popper' make: *** [popper_server] Error 2 Oops, it seems my fault. The socklen_t is used in an offhand. Please try this patch: Index: config.h.in diff -u config.h.in.orig config.h.in --- config.h.in.orig2007-04-27 06:59:36.0 +0900 +++ config.h.in 2007-12-05 03:46:16.0 +0900 @@ -197,6 +197,9 @@ /* Define if you have the iruserok_sa function. */ #undef HAVE_IRUSEROK_SA +/* do we have socklen_t? */ +#undef HAVE_SOCKLEN_T + /* Define if you have the dbm.h header file. */ #undef HAVE_DBM_H Index: configure.in diff -u configure.in.orig configure.in --- configure.in.orig 2007-04-27 06:59:36.0 +0900 +++ configure.in2007-12-05 03:43:57.0 +0900 @@ -991,6 +991,7 @@ IPv6_CHECK_FUNC(iruserok_sa) fi +AC_EGREP_HEADER(socklen_t, sys/socket.h, AC_DEFINE(HAVE_SOCKLEN_T)) AC_ARG_ENABLE(tempnam, [ --enable-tempnam Use tempnam() instead of mkstemp()], Index: popper/main.c diff -u -p popper/main.c.orig popper/main.c --- popper/main.c.orig 2007-04-27 07:11:00.0 +0900 +++ popper/main.c 2007-12-05 03:51:07.0 +0900 @@ -138,6 +138,10 @@ # endif /* HAVE_PAM_PAM_APPL_H */ #endif /* USE_PAM */ +#ifndef HAVE_SOCKLEN_T +typedef unsigned int socklen_t; +#endif /* HAVE_SOCKLEN_T */ + #ifdefINET6 # include netdb.h #else /* not INET6 */ Sincerely, -- Hajimu UMEMOTO @ Internet Mutual Aid Society Yokohama, Japan [EMAIL PROTECTED] [EMAIL PROTECTED],jp.}FreeBSD.org http://www.imasy.org/~ume/
Re: Tru64 and C2
At 03:47 PM 12/2/2007 -0800, Randall Gellens wrote: At 4:10 PM -0600 12/2/07, Vic Summerour wrote: I have used qpopper with Tru64, Digital Unix, OSF/1 for years. Sometime in the not so distant past but I can't remember when or the details of why, I have had to configure with the following: ./configure --enable-specialauth --without-pam Was --enable-specialauth needed? If you left it off, what happened? With ./configure only, the following occurs: ./configure . . . checking for pam_authenticate in -lpam... no Can't use PAM: can't find libpam With ./configure --without-pam, it will configure and compile correctly. Attempts to retrieve pop mail results in an invalid password message: With ./configure --enable-specialauth --without-pam, it will configure correctly, compile, and work. Currently with Tru64 V5.1A with PK6, I have successfully compiled and run 4.1a4 with the following compiler: alpha cc -V Compaq C V6.4-014 on Compaq Tru64 UNIX V5.1A (Rev. 1885) Compiler Driver V6.4-215 (sys) cc Driver Unfortunately, 4.1a5 will not successfully compile based on the same configure. It stop with the error: cc -c -I.. -I.. -I. \ -I../mmangle -I../common \ -g -DHAVE_CONFIG_H -DOSF1 -DUNIX main.c -o main.o cc: Error: main.c, line 150: Missing type specifier or type qualifier. (missingtype) socklen_t ai_addrlen; ^ make[1]: *** [main.o] Error 1 make[1]: Leaving directory `/dsk1/vic/qpopper/qpopper4.1a5/popper' make: *** [popper_server] Error 2 This appears to be associated with 4.1a4 to 4.1a5 enhancements associated with ipv6 support. Since Tru 5.1A supports ipv6 even though it is not being used, I tried to configure with: ./configure --enable-specialauth --without-pam --enable-ipv6 but that made no difference. By defining INET6 in config.h, 4.1a5 will successfully compile and run. Do you know in which header file this is normally defined? If it is a missing .h on Tru64, fixing ./configure to add it or add it in the right place might be the proper solution. For Tru64 5.1A PK6, I could not find INET6 defined in any header file in any of the obvious include directories. I also looked briefly through the Network Programmer's Guide for mention of it without success. -- Randall Gellens Opinions are personal;facts are suspect;I speak for myself only -- Randomly-selected tag: --- Computers ... are not designed, as we are, for ambiguity. --Thomas
Re: Tru64 and C2
I have used qpopper with Tru64, Digital Unix, OSF/1 for years. Sometime in the not so distant past but I can't remember when or the details of why, I have had to configure with the following: ./configure --enable-specialauth --without-pam Currently with Tru64 V5.1A with PK6, I have successfully compiled and run 4.1a4 with the following compiler: alpha cc -V Compaq C V6.4-014 on Compaq Tru64 UNIX V5.1A (Rev. 1885) Compiler Driver V6.4-215 (sys) cc Driver Unfortunately, 4.1a5 will not successfully compile based on the same configure. It stop with the error: cc -c -I.. -I.. -I. \ -I../mmangle -I../common \ -g -DHAVE_CONFIG_H -DOSF1 -DUNIX main.c -o main.o cc: Error: main.c, line 150: Missing type specifier or type qualifier. (missingtype) socklen_t ai_addrlen; ^ make[1]: *** [main.o] Error 1 make[1]: Leaving directory `/dsk1/vic/qpopper/qpopper4.1a5/popper' make: *** [popper_server] Error 2 This appears to be associated with 4.1a4 to 4.1a5 enhancements associated with ipv6 support. Since Tru 5.1A supports ipv6 even though it is not being used, I tried to configure with: ./configure --enable-specialauth --without-pam --enable-ipv6 but that made no difference. By defining INET6 in config.h, 4.1a5 will successfully compile and run. Vic Summerour At 09:57 AM 12/1/2007 -0800, Randall Gellens wrote: At 5:17 PM -0600 11/30/07, J Bacher wrote: Does qpopper no longer support Tru64 5.1 enhanced security? It's been awhile since I've been asked to upgrade qpopper on Tru64 and previous versions (at least up to 4.0.4) supported enhanced security with the --enable-specialauth configuration option. The --enable-specialauth option is still there, although ./configure figures out if it is needed for many platforms. Are you having a problem with Tru64? -- Randall Gellens Opinions are personal;facts are suspect;I speak for myself only -- Randomly-selected tag: --- When we remember we are all mad, the mysteries disappear and life stands explained. --Mark Twain
Re: Tru64 and C2
At 4:10 PM -0600 12/2/07, Vic Summerour wrote: I have used qpopper with Tru64, Digital Unix, OSF/1 for years. Sometime in the not so distant past but I can't remember when or the details of why, I have had to configure with the following: ./configure --enable-specialauth --without-pam Was --enable-specialauth needed? If you left it off, what happened? Currently with Tru64 V5.1A with PK6, I have successfully compiled and run 4.1a4 with the following compiler: alpha cc -V Compaq C V6.4-014 on Compaq Tru64 UNIX V5.1A (Rev. 1885) Compiler Driver V6.4-215 (sys) cc Driver Unfortunately, 4.1a5 will not successfully compile based on the same configure. It stop with the error: cc -c -I.. -I.. -I. \ -I../mmangle -I../common \ -g -DHAVE_CONFIG_H -DOSF1 -DUNIX main.c -o main.o cc: Error: main.c, line 150: Missing type specifier or type qualifier. (missingtype) socklen_t ai_addrlen; ^ make[1]: *** [main.o] Error 1 make[1]: Leaving directory `/dsk1/vic/qpopper/qpopper4.1a5/popper' make: *** [popper_server] Error 2 This appears to be associated with 4.1a4 to 4.1a5 enhancements associated with ipv6 support. Since Tru 5.1A supports ipv6 even though it is not being used, I tried to configure with: ./configure --enable-specialauth --without-pam --enable-ipv6 but that made no difference. By defining INET6 in config.h, 4.1a5 will successfully compile and run. Do you know in which header file this is normally defined? If it is a missing .h on Tru64, fixing ./configure to add it or add it in the right place might be the proper solution. -- Randall Gellens Opinions are personal;facts are suspect;I speak for myself only -- Randomly-selected tag: --- Computers ... are not designed, as we are, for ambiguity. --Thomas
Tru64 and C2
Does qpopper no longer support Tru64 5.1 enhanced security? It's been awhile since I've been asked to upgrade qpopper on Tru64 and previous versions (at least up to 4.0.4) supported enhanced security with the --enable-specialauth configuration option.