Magnus Fromreide wrote:
> On Fri, Aug 25, 2006 at 04:15:55PM +0200, Thomas Anders wrote:
>> Robert Story wrote:
>>> $ cvs diff configure.in
>>> Index: configure.in
>> [...]
>>> +AC_CONFIG_HEADER(include/net-snmp/libtool-config.h)
>>> AC_CONFIG_HEADER(include/net-snmp/net-snmp-config.h)
>> I don't
Wes Hardaker wrote:
> What we *should* do is:
[...]
> 2) fix the incompatibilities when we can (removing the PACKAGE
>variables is an excellent example).
The attached patch uses AH_VERBATIM to make autoheader protect the PACKAGE_*
variables in net-snmp-config.h[.in] in a way suitable for all b
On Fri, Aug 25, 2006 at 04:15:55PM +0200, Thomas Anders wrote:
> Robert Story wrote:
> > $ cvs diff configure.in
> > Index: configure.in
> [...]
> > +AC_CONFIG_HEADER(include/net-snmp/libtool-config.h)
> > AC_CONFIG_HEADER(include/net-snmp/net-snmp-config.h)
>
> I don't think this'll work. autohe
On Fri, Aug 25, 2006 at 04:15:55PM +0200, Thomas Anders wrote:
> Robert Story wrote:
> > $ cvs diff configure.in
> > Index: configure.in
> [...]
> > +AC_CONFIG_HEADER(include/net-snmp/libtool-config.h)
> > AC_CONFIG_HEADER(include/net-snmp/net-snmp-config.h)
>
> I don't think this'll work. autohe
On Fri, 25 Aug 2006 16:15:55 +0200 Thomas wrote:
TA> > Index: configure.in
TA> [...]
TA> > +AC_CONFIG_HEADER(include/net-snmp/libtool-config.h)
TA> > AC_CONFIG_HEADER(include/net-snmp/net-snmp-config.h)
TA>
TA> I don't think this'll work. autoheader can only maintain a *single*
TA> FILE.in (from
Robert Story wrote:
> $ cvs diff configure.in
> Index: configure.in
[...]
> +AC_CONFIG_HEADER(include/net-snmp/libtool-config.h)
> AC_CONFIG_HEADER(include/net-snmp/net-snmp-config.h)
I don't think this'll work. autoheader can only maintain a *single*
FILE.in (from a single acconfig.h). Splitting
On 23/08/06, Magnus Fromreide <[EMAIL PROTECTED]> wrote:
> The problem with shipping of the HAVE_s is that they reflect the state on
> the build machine as opposed to the install machine and in those days of
> package managers the build and install machine often do differ.
Good point.
> This lead
On Tue, Aug 22, 2006 at 09:26:03AM +0100, Dave Shield wrote:
> > TA> It's a start, but I think our current approach of shipping all
> > TA> autoconf variables in net-snmp-config.h is even more flawed than
> > TA> that. See some random Debian bug (wrt. some other software) on the
> > TA> topic:
> >
On 22/08/06, Wes Hardaker <[EMAIL PROTECTED]> wrote:
> DS> Something like
> DS> #ifndef OMIT_PROBLEMATICAL_SETTINGS
>
> If we don't use those, I wouldn't include them by default...
Fine.
I was assuming that the default would be to provide the same set of
definitions that we've always done (for bac
> "DS" == Dave Shield <[EMAIL PROTECTED]> writes:
DS> Is it really that unimaginable that a different package might also
DS> define LOGFILE?
Sure!
No, just kidding of course.
DS> a) Net-SNMP-specific settings - stuff that's only meaningful to our code.
DS> (e.g. DEFAULT_MIBDIRS, USE_REVERS
> "DS" == Dave Shield <[EMAIL PROTECTED]> writes:
DS> Something like
DS> #ifndef OMIT_PROBLEMATICAL_SETTINGS
If we don't use those, I wouldn't include them by default...
--
Wes Hardaker
Sparta, Inc.
-
Using Tomcat but n
Robert Story wrote:
> On Wed, 16 Aug 2006 00:01:33 +0200 Thomas wrote:
> TA> Another proposal that came up on IRC was to split net-snmp-config.h into
> an
> TA> external and an internal header file (with the PACKAGE_* variables being
> TA> in the internal one, of course).
>
> I would like to see
On 22/08/06, Thomas Anders <[EMAIL PROTECTED]> wrote:
> Agreed in general. Still, where would you suggest to put the PACKAGE_*
> variables in that picture?
c) They're clearly generically named, but the values are Net-SNMP specific.
> Actually, they look too problematic to be put anywhere. :-(
Dave Shield wrote:
> As I see it, there are three general types of entry in the config.h file:
>
>a) Net-SNMP-specific settings - stuff that's only meaningful to our code.
> (e.g. DEFAULT_MIBDIRS, USE_REVERSE_ASNENCODING, etc)
[...]
> b) Generic system-related settings
> (e
> TA> It's a start, but I think our current approach of shipping all
> TA> autoconf variables in net-snmp-config.h is even more flawed than
> TA> that. See some random Debian bug (wrt. some other software) on the
> TA> topic:
>
> TA> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=363173
Wes> We
Magnus Fromreide wrote:
> On Tue, Aug 22, 2006 at 12:03:05AM +0200, Thomas Anders wrote:
>> Robert Story wrote:
>>> On Fri, 18 Aug 2006 10:29:17 -0700 Wes wrote:
>>> WH> And that's included automatically by what file? Or is included
>>> WH> explicitly by what files?
>>>
>>> Well, as near as I can
On Tue, Aug 22, 2006 at 12:03:05AM +0200, Thomas Anders wrote:
> Robert Story wrote:
> > On Fri, 18 Aug 2006 10:29:17 -0700 Wes wrote:
> > WH> And that's included automatically by what file? Or is included
> > WH> explicitly by what files?
> >
> > Well, as near as I can tell from cursory checking
> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
TA> It's a start, but I think our current approach of shipping all
TA> autoconf variables in net-snmp-config.h is even more flawed than
TA> that. See some random Debian bug (wrt. some other software) on the
TA> topic:
TA> http://bugs.debian.
Robert Story wrote:
> On Fri, 18 Aug 2006 10:29:17 -0700 Wes wrote:
> WH> And that's included automatically by what file? Or is included
> WH> explicitly by what files?
>
> Well, as near as I can tell from cursory checking, none of our code uses the
> PACKAGE_* macros, so we could just not includ
On Fri, 18 Aug 2006 10:29:17 -0700 Wes wrote:
WH> And that's included automatically by what file? Or is included
WH> explicitly by what files?
Well, as near as I can tell from cursory checking, none of our code uses the
PACKAGE_* macros, so we could just not include it anywhere until we need it.
> "RS" == Robert Story <[EMAIL PROTECTED]> writes:
RS> $ cat include/net-snmp/libtool-config.h.in
And that's included automatically by what file? Or is included
explicitly by what files?
--
Wes Hardaker
Sparta, Inc.
-
On Thu, 17 Aug 2006 20:45:10 -0700 Wes wrote:
WH> > "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
WH>
WH> TA> Then, I'd start with putting PACKAGE_* into the internal header
WH> TA> file and work from there. Suggestions for additional entries are
WH> TA> most welcome.
WH>
WH> I think one
> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
TA> Then, I'd start with putting PACKAGE_* into the internal header
TA> file and work from there. Suggestions for additional entries are
TA> most welcome.
I think one of the things you should do is propose your total
solution here... some o
Wes Hardaker wrote:
>> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
>
> TA> The name of the additional header file is probably the last thing
> TA> I'd object against if we solved the issue in question. Let's agree
> TA> on splitting the two first.
>
> That's ok I'll object then. I obj
> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
TA> The name of the additional header file is probably the last thing
TA> I'd object against if we solved the issue in question. Let's agree
TA> on splitting the two first.
That's ok I'll object then. I object to using config.h. Lets make
Robert Story wrote:
> On Wed, 16 Aug 2006 00:23:07 +0200 Thomas wrote:
> TA> Hmmh, I actually prefer your original net-snmp-config-internal.h proposal
> TA> since it's much more clear what to expect from this file, then. I'd be
> TA> reluctant about re-using the old config.h name for something not
On Wed, 16 Aug 2006 00:23:07 +0200 Thomas wrote:
TA> Robert Story wrote:
TA> > I'd propose going back to the original 'config.h' for the 'internal' file
TA> > name.
TA>
TA> Hmmh, I actually prefer your original net-snmp-config-internal.h proposal
TA> since it's much more clear what to expect from
Robert Story wrote:
> On Wed, 16 Aug 2006 00:01:33 +0200 Thomas wrote:
> TA> Another proposal that came up on IRC was to split net-snmp-config.h into
> an
> TA> external and an internal header file (with the PACKAGE_* variables being
> TA> in the internal one, of course).
>
> I would like to see
On Wed, 16 Aug 2006 00:01:33 +0200 Thomas wrote:
TA> Another proposal that came up on IRC was to split net-snmp-config.h into an
TA> external and an internal header file (with the PACKAGE_* variables being
TA> in the internal one, of course).
I would like to see this resolved as well, and think th
-Coders,
net-snmp-config.h currently contains a number of problematic variables like all
the PACKAGE_* that shouldn't be defined in an external header file because it
may easily interfere with the PACKAGE_* variables (even if internal) of another
package that's using net-snmp. I'd really like to g
30 matches
Mail list logo
