On 12/04/13 03:03, Eric Blake wrote:
> Signed-off-by: Eric Blake
> ---
>
> I'm not pushing this without a decent review, particularly since
> I haven't personally tested it with the FreeBSD setup that Roman
> reported where gettext 0.18.3 doesn't play nicely with the
> autom4te wrapper.
Not ac
On Wed, Dec 04, 2013 at 03:38:33AM +, Pádraig Brady wrote:
> Ah so mallopt(M_CHECK_ACTION, 2) is documented to abort() on error.
> I suppose we need to suppress core dumps now with the gnulib
> ]GL_NOCRASH[ ... nocrash_init() guard?
> Actually no, that doesn't catch SIGABRT.
> So I suppose you
On Tue, Dec 03, 2013 at 08:55:17PM -0700, Eric Blake wrote:
> MALLOC_CHECK_ in the environment at the start of ./configure affects
> hundreds or even thousands of processes. mallopt() within a single
> conftest affects only the one process. I found it slightly easier to
> use mallopt() within the
On 12/03/2013 08:31 PM, Siddhesh Poyarekar wrote:
> On Tue, Dec 03, 2013 at 10:08:04AM -0700, Eric Blake wrote:
>> Michal complained that setting MALLOC_CHECK_ for the entire configure
>> slowed things down.
>
> That's funny because...
>
>> #endif
>> +#ifdef M_CHECK_ACTION
>> +mallop
On 12/03/2013 05:08 PM, Eric Blake wrote:
> On 11/28/2013 07:34 AM, Eric Blake wrote:
>
> [back from holidays, so reviving this thread]
>
>>> Michal, since you have the environment to test this, can you rerun:
>>>
>>> ./configure MALLOC_CHECK_=2
>>>
>>> and see if the hang goes away? If so, then
On Tue, Dec 03, 2013 at 10:08:04AM -0700, Eric Blake wrote:
> Michal complained that setting MALLOC_CHECK_ for the entire configure
> slowed things down.
That's funny because...
> #endif
> +#ifdef M_CHECK_ACTION
> +mallopt(M_CHECK_ACTION, 2);
> +#endif
> +
... this should have the s
With my help on IRC, Roman Bogorodskiy identified a bug where the
BSD autoconf wrapper script eats stdin[1], which in turn causes
autopoint from gettext 0.18.3 to fail to find AM_GNU_GETTEXT_VERSION
out of configure.ac[2], with this error:
./bootstrap: autopoint --force
autopoint: *** Missing vers
On 12/03/2013 06:28 PM, Pádraig Brady wrote:
> On 12/03/2013 05:47 PM, Paul Eggert wrote:
>> Perhaps I'm reading too much into the standards, but that's how
>> --with-jpeg etc. behave in GNU Emacs 'configure'.
>
> Well I agree, that's how I expect --with options to work generally.
> Making the exi
On 12/03/2013 11:08 PM, Eric Blake wrote:
> On 12/03/2013 03:44 PM, Eric Blake wrote:
>
>>
>> Libvirt would prefer a solution that uses nettle, at least when used in
>> RHEL.
>
> Correction: libvirt would prefer a solution that uses gnutls, and could
> live with a solution that uses openssl. Cer
Perhaps I'm reading too much into the standards, but that's how
--with-jpeg etc. behave in GNU Emacs 'configure'.
FWIW, I think you are right, and --with should fail if the specified
package is not installed, per the standards text. Alternatively, it
could configure "as if" the with-packa
On 12/03/2013 03:44 PM, Eric Blake wrote:
>
> Libvirt would prefer a solution that uses nettle, at least when used in
> RHEL.
Correction: libvirt would prefer a solution that uses gnutls, and could
live with a solution that uses openssl. Certification-wise, indirect
use of nettle via gnutls is
On 12/03/2013 03:19 PM, Nikos Mavrogiannopoulos wrote:
> On Tue, 2013-12-03 at 17:39 +0100, Ludovic Courtès wrote:
>> Pádraig Brady skribis:
>>
>>> The speed of md5 and sha* hashes has lagged a bit in gnulib.
>>> So to address that and to take advantage of the architecture
>>> specific assembly us
On Tue, 2013-12-03 at 17:39 +0100, Ludovic Courtès wrote:
> Pádraig Brady skribis:
>
> > The speed of md5 and sha* hashes has lagged a bit in gnulib.
> > So to address that and to take advantage of the architecture
> > specific assembly used in libcrypto, the attached gnulib patch
> > allows proj
With my help on IRC, Roman Bogorodskiy identified a bug where the
BSD autoconf wrapper script eats stdin[1], which in turn causes
autopoint from gettext 0.18.3 to fail to find AM_GNU_GETTEXT_VERSION
within configure.ac[2], with this error:
./bootstrap: autopoint --force
autopoint: *** Missing vers
ni...@lysator.liu.se (Niels Möller) writes:
> I haven't paid much attention to x86_64 assembly for sha1 and md5 (there
> is x86_64 assembly for sha1, which is a direct translation of an earlier
> version of the x86 assembly, and there's no md5-assembly at all for
> x86_64).
I just checked in a so
On 12/03/2013 05:47 PM, Paul Eggert wrote:
> On 12/03/2013 05:45 AM, Pádraig Brady wrote:
>> I'll probably do this in coreutils configure.ac before gl_INIT,
>> so as to at least set the default as coreutils wants and caters for,
>> and allowing users to --without-openssl if they want.
>>
>> dnl Ena
Fine by me.
On Tue, Dec 3, 2013 at 4:29 PM, Pádraig Brady wrote:
> On 12/03/2013 04:19 PM, Ivailo wrote:
> > Thanks, but I'm afraid you will have to add -Wformat-security to the
> > pragmas. I've tested the patch adapted for diffutils and the gnulib
> > test suite still fails the same way, unle
On 12/03/2013 05:45 AM, Pádraig Brady wrote:
> I'll probably do this in coreutils configure.ac before gl_INIT,
> so as to at least set the default as coreutils wants and caters for,
> and allowing users to --without-openssl if they want.
>
> dnl Enable use of libcrypto by default
> AS_VAR_SET_IF([w
On 12/03/2013 10:31 AM, Pádraig Brady wrote:
>>
>> Any preferences on which one I should use in gnulib?
>>
>
> mallopt seems preferable to me.
Thanks, I went ahead and pushed the mallopt version of the patch.
--
Eric Blake eblake redhat com+1-919-301-3266
Libvirt virtualization library
l...@gnu.org (Ludovic Courtès) writes:
> Pádraig Brady skribis:
>
>> The speed of md5 and sha* hashes has lagged a bit in gnulib.
>> So to address that and to take advantage of the architecture
>> specific assembly used in libcrypto, the attached gnulib patch
>> allows projects to configure --wit
On 12/03/2013 05:08 PM, Eric Blake wrote:
> On 11/28/2013 07:34 AM, Eric Blake wrote:
>
> [back from holidays, so reviving this thread]
>
>>> Michal, since you have the environment to test this, can you rerun:
>>>
>>> ./configure MALLOC_CHECK_=2
>>>
>>> and see if the hang goes away? If so, then
On 12/03/2013 10:08 AM, Pádraig Brady wrote:
> Now it should be fairly easy to map the gnulib interface
> to the nettle one, so we might in future provide a --with-nettle
> option to gnulib similar to the current --with-openssl
Or maybe more like the threads interface, where we have
--with-crypto
On 12/03/2013 04:57 PM, Niels Möller wrote:
> l...@gnu.org (Ludovic Courtès) writes:
>
>> Pádraig Brady skribis:
>>
>>> The speed of md5 and sha* hashes has lagged a bit in gnulib.
>>> So to address that and to take advantage of the architecture
>>> specific assembly used in libcrypto, the attach
On 11/28/2013 07:34 AM, Eric Blake wrote:
[back from holidays, so reviving this thread]
>> Michal, since you have the environment to test this, can you rerun:
>>
>> ./configure MALLOC_CHECK_=2
>>
>> and see if the hang goes away? If so, then I know how to patch gnulib
>> to ensure that this part
Pádraig Brady skribis:
> The speed of md5 and sha* hashes has lagged a bit in gnulib.
> So to address that and to take advantage of the architecture
> specific assembly used in libcrypto, the attached gnulib patch
> allows projects to configure --with-openssl to use that if
> available or fall ba
On 12/03/2013 04:19 PM, Ivailo wrote:
> Thanks, but I'm afraid you will have to add -Wformat-security to the
> pragmas. I've tested the patch adapted for diffutils and the gnulib
> test suite still fails the same way, unless I add the pragma my self
> that is.
Thankgs for testing. Going with this
Thanks, but I'm afraid you will have to add -Wformat-security to the
pragmas. I've tested the patch adapted for diffutils and the gnulib
test suite still fails the same way, unless I add the pragma my self
that is.
Cheers!
On Tue, Dec 3, 2013 at 2:30 PM, Pádraig Brady wrote:
> On 12/02/2013 09
On 12/02/2013 09:51 PM, Ivailo wrote:
> And btw, the same failure is triggered on gettext (which I had to update
> to test the latest diffutils) for the exact same reason - gnulib's xvasprintf
> test. It's not just diffutils that's affected by this.
I'm about to push the attached in your name.
It'
On 12/03/2013 06:24 AM, Paul Eggert wrote:
> Pádraig Brady wrote:
>> Seems the handiest way to do this is to do the following
>> in configure.ac before gl_INIT:
>>
>> dnl Enable use of libcrypto by default
>> AC_ARG_WITH([openssl],
>> [AS_HELP_STRING([--with-openssl],
>> [use libcrypto hash r
29 matches
Mail list logo
