[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 4 Aug 2016 13:51:03 +0300
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: specify home ($JOHN) at compiletime
On Wed, Aug 03, 2016 at 02:52:14PM +0200, Matus UHLAR - fantomas wrote:
> to explain:
> I have started with packaged john 1.7.9, then backported 1.8.0 because of
> fork option. they both used ~/.john/john.pot
>
> then i went to jumbo because of extra formats, and bleeding-jumbo because of
> faster algorithms...
>
> I want to make them all use the same john.conf and john.pot files...
Oh, OK. It is not typical to want that from your user-local builds.
Personally, I am more comfortable with each build using its own files.
> I use debian, but yes, I was trying to create packages, so I could easily
> distribute them to different systems.
If you're the only user for those packages, it might be easier to use
tarballs instead (of the "run" and "doc" directories).
> unfortunately files from debian 1.8.0 package don't work with jumbo:
>
> make[2]: *** No rule to make target 'linux-x86-64'. Stop.
> make[2]: Leaving directory '/home/uhlar/src/debian/john/bleeding-jumbo/src'
>
> I have no problem to build it manually, but the process seems to have
> problems:
>
> env CFLAGS='-DJOHN_SYSTEMWIDE=1 -DCFG_FULL_NAME="/etc/john/john.conf"'
> ./configure
>
> 1. configure warning:
>
> configure: WARNING: No recognized optimization option present in CFLAGS
>
> maybe I should add "-O2" or "-O3" to CFLAGS? (debian adds -funroll-loops
> which makes -O3 from -O2)
Yes, you should have "-O2 -fomit-frame-pointer" in CFLAGS, if you
override CFLAGS at all. Also, by default "configure" will detect and
enable use of your build host CPU's features. This is very important
for performance (e.g., AVX2 is often a 2x speedup), but it makes your
binary non-portable to machines with less capable CPUs. To disable this
detection, use "./configure --disable-native-tests", and then you'll
want to make separate builds for each of your target CPU categories -
e.g., plain x86-64/SSE2, SSSE3, SSE4.1, AVX, XOP, AVX2. With
non-ancient systems, perhaps you can require at least SSSE3, but
anything below AVX is only a minor improvement over SSE2.
As to "-funroll-loops", I think our current "configure" will make use of
it by default where appropriate (not for all source files). Ditto for
"-finline-functions".
(And no, merely adding "-funroll-loops" doesn't make "-O3" from "-O2".
There are more differences between these optimization levels.)
> 2. compile warnings:
>
> params.h:152:0: warning: "CFG_FULL_NAME" redefined
> #define CFG_FULL_NAME "$JOHN/john.conf"
> ^
> <command-line>:0:0: note: this is the location of the previous definition
>
> apparently the #define should to #ifndef CFG_FULL_NAME block
> (debian patch adds that).
> Otherwise, seems that the latter is being used.
I think Debian overrides CFG_FULL_NAME because of their policy to have
all config files under /etc. This policy is inconsistent with JtR's,
and vice versa, where JtR checks a set of directories for several of its
files, including the config file and more. By overriding CFG_FULL_NAME
like above, if it worked, you'd break some desirable functionality -
namely, being able to override the system-wide config file at runtime by
placing a local one in ~/.john/john.conf. Just how do Debian package
users make edits to their john.conf? Edit the global file as root?
That's nasty. Here's how I do it on Owl, as non-root:
cp /usr/share/john/john.conf ~/.john
vi ~/.john/john.conf
This works due to our default CFG_FULL_NAME.
There's no "#ifndef CFG_FULL_NAME" on purpose. We don't support this,
and proper support would amount to more than just the #ifndef.
I suggest that you simply drop this Debian'ism, especially as it doesn't
even build for you.
> ...I'm out of ideas. Debian uses backslash escaping:
>
> CFLAGS='-DJOHN_SYSTEMWIDE=1 -DCFG_FULL_NAME=\"/etc/john/john.conf\"'
> ./configure
>
> but that causes configure errors:
>
> configure: creating ./fmt_externs.h
> <command-line>:0:16: warning: missing terminating " character
I guess the exact escaping needed varies between Debian's invocation
from a script and your invocation from the command-line.
> any idea how to continue?
Just don't override CFG_FULL_NAME. Also, seriously consider whether you
actually want system-wide builds.
And I guess by now you realize that you're doing something weird, and/or
that we don't support your use case well (sorry!) Hopefully, you have
at least tens of similar machines to make this exercise worthwhile.
Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
