integrit-users Mailing List for integrit file verification system (Page 6)
Brought to you by:
ecashin
Menu
▾
▴
integrit-users — mailing list for discussion of issues related to using integrit
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(9) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(10) |
Feb
(27) |
Mar
(5) |
Apr
(1) |
May
(1) |
Jun
(12) |
Jul
(5) |
Aug
(14) |
Sep
(6) |
Oct
(31) |
Nov
(6) |
Dec
(4) |
| 2002 |
Jan
(2) |
Feb
(13) |
Mar
(2) |
Apr
(3) |
May
(2) |
Jun
(2) |
Jul
(5) |
Aug
(6) |
Sep
(13) |
Oct
(1) |
Nov
(1) |
Dec
(1) |
| 2003 |
Jan
(8) |
Feb
|
Mar
(2) |
Apr
|
May
(2) |
Jun
|
Jul
(8) |
Aug
(5) |
Sep
(1) |
Oct
|
Nov
|
Dec
|
| 2004 |
Jan
|
Feb
(4) |
Mar
(1) |
Apr
(4) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2005 |
Jan
(1) |
Feb
|
Mar
|
Apr
(5) |
May
(5) |
Jun
|
Jul
|
Aug
|
Sep
(5) |
Oct
(5) |
Nov
|
Dec
(1) |
| 2006 |
Jan
|
Feb
|
Mar
(2) |
Apr
(3) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(15) |
Jun
(2) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(2) |
Dec
|
| 2008 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(6) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2022 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Ed L C. <ec...@ug...> - 2002-02-28 20:07:48
|
Ed L Cashin <ec...@ug...> writes: > /bin/ls /etc | grep -v system | sed -e 's/^/!/' >> integrit-system.conf That grep pattern should of course be '^system$', so that you won't match other files whose names contain the string, "system". /bin/ls /etc | grep -v '^system$' | sed -e 's/^/!/' >> integrit-system.conf -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: Ed L C. <ec...@ug...> - 2002-02-28 19:53:32
|
Derrick Judge <der...@in...> writes: > Hi, > > Is there a way to only check one file in a directory with > out making the config file a mile long? For example I only > to monitor /etc/system in the /etc directory. I don't think so, but you'd have to use a big font or a lot of files to really get a mile's worth. ;) > I have read the docs and web pages but have not found a way > to do this. integrit is able to keep its feature list sparse because UNIX is such a rich environment. The trick is to use integrit with other UNIX tools: /bin/ls /etc | grep -v system | sed -e 's/^/!/' >> integrit-system.conf You could automate the regeneration of the config file with a script or makefile. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: Ed L C. <ec...@ug...> - 2002-02-25 21:57:01
|
<eri...@pr...> writes: ... > > It goes to standard output, which must be piped to somewhere else as > > safely as your circumstances merit. For example, for simple > > script-kiddie detection it would be sufficient to simply pipe the > > output to something like sendmail. > > So just piping the output to /bin/sendmail will send the human-readable > output (including md5 checksum) to an e-mail address? e.g., integrit -c -C > /mnt/cdrom/integrit.conf | sendmail fo...@ba.... Yes. There's an example like that in the examples directory called "integrit_check". It uses "sendmail -t" and printf to create a nice header. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: matt <m...@tt...> - 2002-02-25 21:20:03
|
> > > 1. Is there any way to prevent prying eyes from viewing integrit's
> > > configuration file [besides filesystem permissions]? (e.g., Tripwire
I missed the start of this thread, but what I do is keep the databases
and the configuration files on a local computer. These are then
copied to the remote host and run by a local cron job. This keeps all
data and configuration off the remote host, so unless a hacker is
watching ports/processes as the check occurs - they wouldn't even know
that an IDS was being used/run.
If anyone is interested in the Perl scripts I wrote to achieve this,
I can put them on my website and post a link here.
Matt.
--
#!/usr/bin/perl
@t=split//,"Just Another Perl Hacker,";sub J{chr rand(90)+4*8};sub X{($s,$t)=@
_;$$=0;$:=(grep{$$s[$$++]eq$_}@$t)+rand .5;}map{push@P,&J;}@t;$"="";$|=1;while
($:<@t){$;++;my@z;eval'push@z,[map{int rand 30?$_:&J}@P];'x15;@s=sort{&X($a,\@
t)<=>&X($b,\@t)}@z;@P=@{$s[$#s]};print"\r@P";}print" bred in $; generations\n"
|
|
From: <eri...@pr...> - 2002-02-25 05:41:06
|
On 23 Feb 2002, Ed L Cashin wrote: > <eri...@pr...> writes: > > 1. Is there any way to prevent prying eyes from viewing integrit's > > configuration file [besides filesystem permissions]? (e.g., Tripwire > > encodes its configuration files - But I don't know how that would > > work with cron-jobs...) Or, is it nonsense to think that if > > filesystem permissions are not enough. i.e., If an intruder has > > gained root access and can thereby read the config file, it's > > already too late because intruder can then just remove the cron job. > > No, integrit provides no feature for preventing root from viewing > configuration files or database files. I think you're right that this > feature would not be very useful. I was thinking that creating a passphrase-encrypted config file would help slow-down intruders... But I suppose that setting the config file's permissions as `400` would be effective enough. > > 2. Where does "human-readable output" come-from/go-to when integrit is run > > by a cron-job? > > It goes to standard output, which must be piped to somewhere else as > safely as your circumstances merit. For example, for simple > script-kiddie detection it would be sufficient to simply pipe the > output to something like sendmail. So just piping the output to /bin/sendmail will send the human-readable output (including md5 checksum) to an e-mail address? e.g., integrit -c -C /mnt/cdrom/integrit.conf | sendmail fo...@ba.... -- Eric P. Los Gatos, CA |
|
From: Ed L C. <ec...@ug...> - 2002-02-24 01:46:26
|
<eri...@pr...> writes: > Hello, All: > > I'm new to integrity checking, so please forgive my naivete: > > 1. Is there any way to prevent prying eyes from viewing integrit's > configuration file [besides filesystem permissions]? (e.g., Tripwire > encodes its configuration files - But I don't know how that would > work with cron-jobs...) Or, is it nonsense to think that if > filesystem permissions are not enough. i.e., If an intruder has > gained root access and can thereby read the config file, it's > already too late because intruder can then just remove the cron job. No, integrit provides no feature for preventing root from viewing configuration files or database files. I think you're right that this feature would not be very useful. In the worst case, where a sophisticated intruder has gained root access, an off-host integrit database could be valuable for after-the-fact forensics on a cracked system; e.g., you could take the hard drive from the cracked machine and use integrit to examine it on a secure host, perhaps with i-viewdb and i-ls if necessary. > 2. Where does "human-readable output" come-from/go-to when integrit is run > by a cron-job? It goes to standard output, which must be piped to somewhere else as safely as your circumstances merit. For example, for simple script-kiddie detection it would be sufficient to simply pipe the output to something like sendmail. > 3. When running integrit from /mnt/cdrom as a cron-job (and reading > integrit's known database from /mnt/cdrom): Is anything special needed > besides configuring `-c /mnt/cdrom/integrit.cdb`? You'll need a configuration file, too, "-C /mnt/cdrom/hosty.conf". (The -c option doesn't take an argument but just tells integrit to check the filesystem against the known database.) > e.g., Where will the results of the check go - The file specified in > the `current=` parameter of the config file? The results of the check go to standard output for you to read. The results of the update are in the current database -- update means update the current database to reflect the current state of the filesystem. > 4. When running from a cron-job: Where will the md5 hash of the > current database go? (So that - when I do finally find the time to > review the current .cdb - I know that it hasn't been tampered with.) It's at the bottom of the report that integrit produces on its standard output. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: <eri...@pr...> - 2002-02-24 00:26:51
|
Hello, All: I'm new to integrity checking, so please forgive my naivete: 1. Is there any way to prevent prying eyes from viewing integrit's configuration file [besides filesystem permissions]? (e.g., Tripwire encodes its configuration files - But I don't know how that would work with cron-jobs...) Or, is it nonsense to think that if filesystem permissions are not enough. i.e., If an intruder has gained root access and can thereby read the config file, it's already too late because intruder can then just remove the cron job. 2. Where does "human-readable output" come-from/go-to when integrit is run by a cron-job? 3. When running integrit from /mnt/cdrom as a cron-job (and reading integrit's known database from /mnt/cdrom): Is anything special needed besides configuring `-c /mnt/cdrom/integrit.cdb`? e.g., Where will the results of the check go - The file specified in the `current=` parameter of the config file? 4. When running from a cron-job: Where will the md5 hash of the current database go? (So that - when I do finally find the time to review the current .cdb - I know that it hasn't been tampered with.) -- Eric P. Los Gatos, CA |
|
From: Ed L C. <ec...@ug...> - 2002-02-12 14:43:02
|
Franky Van Liedekerke <lie...@pa...> writes: > Yup, the less options, the easier the use gets :) and RD...@ga... (Mr. Ross Druker) writes: > I only use them on occasion, but the change is fine by me. Great. 3.00.05-beta seems pretty stable to me, so there ought to be a new stable version coming out soon, and its utils won't have "-s" but will instead print any available SHA1 info by default. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: <RD...@ga...> - 2002-02-12 13:28:12
|
I only use them on occasion, but the change is fine by me. On Feb 11, 5:46pm, Ed L Cashin wrote: > Subject: [Integrit-users] poll: always show SHA1 for i-viewdb, i-ls? > Hey, folks. > > I find it a bit annoying to have to specify the "-s" flag to get SHA1 > checksums in the output of i-viewdb and i-ls. If I'm going to the > trouble of using these utilities, odds are I want all the info. > > If any of you use the supplemental utilities, what are your feelings > about making the default behavior of the two utilities be to show the > SHA1 checksum? Then the "-s" option could be removed, making them > even simpler than they are now. > > -- > --Ed L Cashin | PGP public key: > ec...@ug... | http://noserose.net/e/pgp/ > > > _______________________________________________ > Integrit-users mailing list > Int...@li... > https://lists.sourceforge.net/lists/listinfo/integrit-users >-- End of excerpt from Ed L Cashin -- Ross Druker Rohm and Haas Co. RD...@Ro... Philadelphia, PA (215) 592-3281 The opinions expressed are mine and not those of Rohm and Haas Company. |
|
From: Franky V. L. <lie...@pa...> - 2002-02-11 22:59:08
|
Yup, the less options, the easier the use gets :) On 11 Feb 2002 17:46:29 -0500 Ed L Cashin <ec...@ug...> wrote: > Hey, folks. > > I find it a bit annoying to have to specify the "-s" flag to get SHA1 > checksums in the output of i-viewdb and i-ls. If I'm going to the > trouble of using these utilities, odds are I want all the info. > > If any of you use the supplemental utilities, what are your feelings > about making the default behavior of the two utilities be to show the > SHA1 checksum? Then the "-s" option could be removed, making them > even simpler than they are now. > > -- > --Ed L Cashin | PGP public key: > ec...@ug... | http://noserose.net/e/pgp/ > > > _______________________________________________ > Integrit-users mailing list > Int...@li... > https://lists.sourceforge.net/lists/listinfo/integrit-users > |
|
From: Ed L C. <ec...@ug...> - 2002-02-11 22:46:27
|
Hey, folks. I find it a bit annoying to have to specify the "-s" flag to get SHA1 checksums in the output of i-viewdb and i-ls. If I'm going to the trouble of using these utilities, odds are I want all the info. If any of you use the supplemental utilities, what are your feelings about making the default behavior of the two utilities be to show the SHA1 checksum? Then the "-s" option could be removed, making them even simpler than they are now. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: Ed L C. <ec...@ug...> - 2002-02-04 05:45:32
|
Hi. After a bit of orienteering through the forest of make compatability, integrit now builds on all the platforms I tested, and should build on many others too. It will build in a directory other than the source directory, like this: tar xvfz ~/packages/integrit-3.00.05.tar.gz mkdir integrit-build cd integrit-build ../integrit-3.00-beta/configure && make && make utils It includes the Windows compatability changes (getting rid of files with names like "aux"). It includes the cmp-like exit statuses -- hence the new major version number. I may have forgotten some changes, but they should all be documented in the Changes file. -- --Ed L Cashin | PGP public key: ec...@ug... | http://noserose.net/e/pgp/ |
|
From: Jos B. <jo...@cn...> - 2002-01-05 21:45:01
|
On Sat, Jan 05, 2002 at 02:11:53PM -0500, Ed L Cashin wrote:
> I've made the necessary changes, and the CVS version now supports
> separate build directories.
Whoopee :-)
> VPATH isn't necessary because of the way the makefiles work. The new
> makefile system is compatible with more versions of make and handles
> the build/source directories explicitly.
Yeah, I see, dep.mak now uses qualified paths.
Fyi, the FreeBSD build now only shows these warnings, which I think we have
discussed before (it's the long != time_t issue).
../integrit/xml.c: In function `xml_start_report':
../integrit/xml.c:119: warning: long int format, time_t arg (arg 3)
gcc -I. -I../integrit -I../integrit/hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o eachfile.o -c ../integrit/eachfile.c
../integrit/eachfile.c: In function `show_diffs':
../integrit/eachfile.c:263: warning: assignment from incompatible pointer type
--
Jos Backus _/ _/_/_/ Santa Clara, CA
_/ _/ _/
_/ _/_/_/
_/ _/ _/ _/
jo...@cn... _/_/ _/_/_/ use Std::Disclaimer;
|
|
From: Ed L C. <ec...@te...> - 2002-01-05 19:14:43
|
(copied to integrit-users) Jos Backus <jo...@cn...> writes: ... > How hard would it be to adapt the auto* stuff to generate configure/Makefile > such that builddir != srcdir? This is very convenient when building for > multiple platforms. This way one can do > > tar xzvf integrit.tar.gz > mkdir obj > cd obj > ../integrit/configure > make I've made the necessary changes, and the CVS version now supports separate build directories. > I tried adding > > VPATH = /home/jos/src/integrit/integrit VPATH isn't necessary because of the way the makefiles work. The new makefile system is compatible with more versions of make and handles the build/source directories explicitly. ... > Also, with today's CVS I am seeing > > lizzy:~/src/integrit/integrit-3.0% make > make: don't know how to make integrit.h. Stop > > integrit.h doesn't appear to exist anywhere... For the list folks, this file was omitted at the time of my first "version three in CVS" announcement but was added after Jos' report. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: Ed L C. <ec...@te...> - 2001-12-29 18:48:20
|
Ed L Cashin <ec...@te...> writes: > Please let me know if you find something that isn't working. ... but I'm going to be away from email until January third. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: Ed L C. <ec...@te...> - 2001-12-29 05:45:11
|
"breech[mdo]" <br...@mi...> writes: > Hello all, > > I'm having an issue compiling integrit-2.03.02 on an OpenBSD-stable sparc64 > system: > > [breech@sparc64 integrit-2.03]$ sudo gmake > >>> making library for shared integrit code: gmake -f integrit.mak > libintegrit.a > gmake[1]: Entering directory `/usr/src/integrit-2.03' > gcc -I. -I ./hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o gnupg/sha1.o -c > gnupg/sha1.c > gnupg/sha1.c: In function `transform': > gnupg/sha1.c:237: virtual memory exhausted > gmake[1]: *** [gnupg/sha1.o] Error 1 > gmake[1]: Leaving directory `/usr/src/integrit-2.03' > gmake: *** [libintegrit.a] Error 2 > [breech@sparc64 integrit-2.03]$ uname -a > OpenBSD sparc64 3.0 sparc64-122501#0 sparc64 > > Any ideas? Thanks in advance ... It says that gcc's out of virtual memory during the build -- maybe it really ran out of memory or there's some problem with the build tools. Unless someone here has seen that problem, an openbsd newsgroup or mailing list might have more answers for you. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: Ed L C. <ec...@te...> - 2001-12-28 22:55:16
|
Hi. Version three is in CVS. The major version number change is on account of the change in the way integrit's exit status works. The exit status is now like cmp: zero for no change; one indicates changes; and two indicates error. It will break old scripts that take non-zero exit status as indicating error. I hope HP-UX users will check out the new version, since it has many changes suggested by David Doubrava and others that increase compatability with HP-UX (and SCO too!). FreeBSD users and others whose system make couldn't build integrit before should be pleased to find that they no longer need to use GNU make. The new makefile structure is more simple and more portable. Please let me know if you find something that isn't working. Instructions for anonymous CVS: http://sourceforge.net/cvs/?group_id=15369 (There's also an undocumented alpha-state test suite and bug-report-generator in the test directory for the adventurous to try.) -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: breech[mdo] <br...@mi...> - 2001-12-28 05:55:00
|
Hello all, I'm having an issue compiling integrit-2.03.02 on an OpenBSD-stable sparc64 system: [breech@sparc64 integrit-2.03]$ sudo gmake >>> making library for shared integrit code: gmake -f integrit.mak libintegrit.a gmake[1]: Entering directory `/usr/src/integrit-2.03' gcc -I. -I ./hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o gnupg/sha1.o -c gnupg/sha1.c gnupg/sha1.c: In function `transform': gnupg/sha1.c:237: virtual memory exhausted gmake[1]: *** [gnupg/sha1.o] Error 1 gmake[1]: Leaving directory `/usr/src/integrit-2.03' gmake: *** [libintegrit.a] Error 2 [breech@sparc64 integrit-2.03]$ uname -a OpenBSD sparc64 3.0 sparc64-122501#0 sparc64 Any ideas? Thanks in advance ... |
|
From: <RD...@ga...> - 2001-11-02 20:56:46
|
Yes, that seems to have fixed those problems. Thanks. Ross On Nov 2, 3:35pm, Ed L Cashin wrote: > Subject: Re: [Integrit-users] install program > RD...@ga... (Mr. Ross Druker) writes: > > > I'm trying to install integrit on HP-UX 10.20 and have run into a problem with > > the "install" program used in the Makefiles. From what I can see this problem > > also applies also to HP-UX 11.0 and AIX 4.2 (and maybe others). SGI IRIX > > appears to provide compatibility. I can work around the problem by manually > > creating needed directories. > > There's a install-sh script included with integrit and a configure > option that tells the build process to use it. > > It sounds like autoconf's test in the configure script that is > supposed to determine whether your system's install command is > BSD-compatable is registering a false positive. > > Try this: > > cd path/to/integrit > ./configure --with-install=`pwd`/install-sh > > You need an absolute path, hence pwd in backticks. > > -- > --Ed Cashin PGP public key: > ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ >-- End of excerpt from Ed L Cashin -- Ross Druker Rohm and Haas Co. RD...@Ro... Philadelphia, PA (215) 592-3281 The opinions expressed are mine and not those of Rohm and Haas Company. |
|
From: <RD...@ga...> - 2001-11-02 20:40:37
|
While compiling on HP-UX 10.20, with the gmake 3.79.1, gcc 3.0.1 (and the required binutils for HP, 2.11.2), I got the following compiler warnings. I don't know if they're particularly important. gcc -I. -I ./hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o xml.o -c xml.c xml.c: In function `xml_start_report': xml.c:119: warning: long int format, time_t arg (arg 3) gcc -I. -I ./hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o eachfile.o -c eachfile.c eachfile.c: In function `show_diffs': eachfile.c:264: warning: assignment from incompatible pointer type gcc -I. -I ./hashtbl -g -Wall -O2 -DHAVE_CONFIG_H -o gnupg/md5.o -c gnupg/md5.c gnupg/md5.c:341:5: warning: pasting "->" and "A" does not give a valid preprocessing token gnupg/md5.c:341:5: warning: pasting "->" and "A" does not give a valid preprocessing token gnupg/md5.c:341:5: warning: pasting "->" and "A" does not give a valid preprocessing token gnupg/md5.c:341:5: warning: pasting "->" and "A" does not give a valid preprocessing token gnupg/md5.c:342:5: warning: pasting "->" and "B" does not give a valid preprocessing token gnupg/md5.c:342:5: warning: pasting "->" and "B" does not give a valid preprocessing token gnupg/md5.c:342:5: warning: pasting "->" and "B" does not give a valid preprocessing token gnupg/md5.c:342:5: warning: pasting "->" and "B" does not give a valid preprocessing token gnupg/md5.c:343:5: warning: pasting "->" and "C" does not give a valid preprocessing token gnupg/md5.c:343:5: warning: pasting "->" and "C" does not give a valid preprocessing token gnupg/md5.c:343:5: warning: pasting "->" and "C" does not give a valid preprocessing token gnupg/md5.c:343:5: warning: pasting "->" and "C" does not give a valid preprocessing token gnupg/md5.c:344:5: warning: pasting "->" and "D" does not give a valid preprocessing token gnupg/md5.c:344:5: warning: pasting "->" and "D" does not give a valid preprocessing token gnupg/md5.c:344:5: warning: pasting "->" and "D" does not give a valid preprocessing token gnupg/md5.c:344:5: warning: pasting "->" and "D" does not give a valid preprocessing token -- Ross Druker Rohm and Haas Co. RD...@Ro... Philadelphia, PA (215) 592-3281 The opinions expressed are mine and not those of Rohm and Haas Company. |
|
From: Ed L C. <ec...@te...> - 2001-11-02 20:38:03
|
RD...@ga... (Mr. Ross Druker) writes: > I'm trying to install integrit on HP-UX 10.20 and have run into a problem with > the "install" program used in the Makefiles. From what I can see this problem > also applies also to HP-UX 11.0 and AIX 4.2 (and maybe others). SGI IRIX > appears to provide compatibility. I can work around the problem by manually > creating needed directories. There's a install-sh script included with integrit and a configure option that tells the build process to use it. It sounds like autoconf's test in the configure script that is supposed to determine whether your system's install command is BSD-compatable is registering a false positive. Try this: cd path/to/integrit ./configure --with-install=`pwd`/install-sh You need an absolute path, hence pwd in backticks. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: <RD...@ga...> - 2001-11-02 20:07:20
|
I'm trying to install integrit on HP-UX 10.20 and have run into a problem with
the "install" program used in the Makefiles. From what I can see this problem
also applies also to HP-UX 11.0 and AIX 4.2 (and maybe others). SGI IRIX
appears to provide compatibility. I can work around the problem by manually
creating needed directories.
The exact line in the main Makefile (a similar one occurs in other Makefiles)
is:
$(INSTALL) -d -m 755 $(SBINDIR); \
The output to the screen when running "gmake install" is:
creating directory /opt/integrit-2.03/sbin
cp: illegal option -- d
Usage: cp [-f|-i] [-p] [-e warn|force|ignore] source_file target_file
cp [-f|-i] [-p] [-e warn|force|ignore] source_file ... target_directory
cp [-f|-i] [-p] -R|-r [-e warn|force|ignore] source_directory ...
target_directory
The problem is the -d option. HP's don't have it. Neither does AIX 4.2. As
far as I can see, you can't use "install" under those systems to create a
directory; only to install files in existing directories.
There is also a problem with the -s option. I've excerpted the man pages for
install from a few systems to give more info.
Under SGI IRIX 6.5:
BSD/GNU compatibility (where the target file or directory is the last
argument) is supported in case none of the mandatory arguments (-f, -F,
or -dir) is provided. In this case the -s option means strip.
the -s option normally has this meaning:
-s Be silent. Older versions of install printed verbose
information by default when installing. Install is now
silent by default, but this option remains for
compatibility.
the -c and -d options:
-c BSD old compatibility mode (ignored)
-dir (or -d for BSD/GNU compatibility)
Create directories named by concatenating $ROOT to the
file arguments. When -dir is given, all the file
arguments are taken as directories; files and directories
may not both be installed with the same command.
Under AIX 4.2, they provide two commands, install and installbsd. However,
neither of them has a -d option for creating directories.
install Command
Syntax
/usr/bin/install [ -c DirectoryA ] [ -f DirectoryB ] [ -i ]
[ -m ] [ -M Mode ] [ -O Owner ] [ -G Group ] [ -S ] [ -n
DirectoryC ] [ -o ] [ -s ] File [ Directory ... ]
-s Suppresses the display of all but error messages.
-S Causes the binary to be stripped after installation.
installbsd Command
Purpose
Installs a command (BSD version of the install command).
Syntax
/usr/bin/installbsd [-c ] [-g Group ] [-m Mode ] [-o Owner
] [-s ] BinaryFile Destination
-s Causes the file specified by the BinaryFile parameter to be stripped
after installation.
Under HP-UX 10.20 and 11.0 the man page has these options:
install(1M) install(1M)
NAME
install - install commands
SYNOPSIS
/usr/sbin/install [-c dira] [-f dirb] [-i] [-n dirc] [-o] [-g group]
[-s] [-u user] file [dirx ...]
--
Ross Druker Rohm and Haas Co.
RD...@Ro... Philadelphia, PA
(215) 592-3281
The opinions expressed are mine and not those of Rohm and Haas Company.
|
|
From: Ed L C. <ec...@te...> - 2001-11-01 21:53:06
|
Wes Byne <fb...@en...> writes: > Just an FYI: it's the gcc that comes bundled with solaris 8's > software supplement. I grabbed the gcc on sunfreeware and > recompiled and integrit works fine. Thanks very much for this helpful information. I'll create a README.Solaris for version three. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
|
From: Wes B. <fb...@en...> - 2001-11-01 21:02:42
|
Just an FYI: it's the gcc that comes bundled with solaris 8's software supplement. I grabbed the gcc on sunfreeware and recompiled and integrit works fine. thanks, Wes On Wed, Oct 31, 2001 at 02:31:05PM -0500, Ed L Cashin wrote: > Wes Byne <fb...@en...> writes: > > > With debugging output: > > > > Starting program: /opt/integrit2/sbin/integrit -C > > /opt/integrit/conf/integrit.conf > > > > Program received signal SIGSEGV, Segmentation fault. > > 0x8073f6a in _init () > > (gdb) step > > Single stepping until exit from function _init, > > which has no line number information. > > > > Program terminated with signal SIGSEGV, Segmentation fault. > > The program no longer exists. > > > > > > I've tinkered with it some, but haven't gotten any further than > > this. Any thoughts? > > This sounds really low-level, since the segfault looks like it's > happenning outside of the domain of the C program, in the C runtime > initialization. > > Have you looked at groups.google.com for Solaris-on-intel users > talking about similar problems? I'm guessing that it's a compiler or > libc (or glibc) problem. > > -- > --Ed Cashin PGP public key: > ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ > > > _______________________________________________ > Integrit-users mailing list > Int...@li... > https://lists.sourceforge.net/lists/listinfo/integrit-users -- ***** Wes Byne Water Research and Consumption fb...@en... (voice):706-542-6041 http://watershed.engr.uga.edu ***** |
|
From: Ed L C. <ec...@te...> - 2001-10-31 19:33:54
|
Wes Byne <fb...@en...> writes: > With debugging output: > > Starting program: /opt/integrit2/sbin/integrit -C > /opt/integrit/conf/integrit.conf > > Program received signal SIGSEGV, Segmentation fault. > 0x8073f6a in _init () > (gdb) step > Single stepping until exit from function _init, > which has no line number information. > > Program terminated with signal SIGSEGV, Segmentation fault. > The program no longer exists. > > > I've tinkered with it some, but haven't gotten any further than > this. Any thoughts? This sounds really low-level, since the segfault looks like it's happenning outside of the domain of the C program, in the C runtime initialization. Have you looked at groups.google.com for Solaris-on-intel users talking about similar problems? I'm guessing that it's a compiler or libc (or glibc) problem. -- --Ed Cashin PGP public key: ec...@te... http://www.terry.uga.edu/~ecashin/pgp/ |
1 message has been excluded from this view by a project administrator.
