[opensuse-factory] may coreutils soon need libcrypto?
Hi ... maybe Coolo? The next coreutils release is coming out in a few days. Besides many other changes, it comes with a performance boost for the md5sum + sha*sum programs up to 50% due to now linking against "openssl" (per default). This is cool per se, but I see potential new build cycles: $ ldd /usr/bin/md5sum md5sum-new /usr/bin/md5sum: linux-vdso.so.1 (0x00007fffdc520000) libc.so.6 => /lib64/libc.so.6 (0x00007f4c917aa000) /lib64/ld-linux-x86-64.so.2 (0x00007f4c91b59000) md5sum-new: linux-vdso.so.1 (0x00007fff457fe000) libcrypto.so.1.0.0 => /lib64/libcrypto.so.1.0.0 (0x00007f33b0603000) libc.so.6 => /lib64/libc.so.6 (0x00007f33b0254000) libdl.so.2 => /lib64/libdl.so.2 (0x00007f33b0050000) libz.so.1 => /lib64/libz.so.1 (0x00007f33afe3a000) /lib64/ld-linux-x86-64.so.2 (0x00007f33b09ee000) Should I use that new default setting, or should I configure --with-openssl=no ? Thanks & have a nice day, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 04.12.2013 18:29, schrieb Bernhard Voelker:
Should I use that new default setting, or should I configure --with-openssl=no ?
Hmm, openssl is indeed not in the cycle, so it would be a change. openssl should be a single addition though from what I see. If it's a problem, we can of course also work with update-alternatives and offer a coreutils-fast. But for now my gut feeling would be to add ssl in the inner loop. Greetings, Stephan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Wednesday 2013-12-04 19:54, Stephan Kulow wrote:
Am 04.12.2013 18:29, schrieb Bernhard Voelker:
Should I use that new default setting, or should I configure --with-openssl=no ?
Hmm, openssl is indeed not in the cycle, so it would be a change. openssl should be a single addition though from what I see. If it's a problem, we can of course also work with update-alternatives and offer a coreutils-fast. But for now my gut feeling would be to add ssl in the inner loop.
Just thinking out loud here - maybe coreutils should be split again at the upstream level. textutils et al :) Then the cryptoutils could be a more leaf-like package. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 12/04/2013 10:39 PM, Jan Engelhardt wrote:
Just thinking out loud here - maybe coreutils should be split again at the upstream level. textutils et al :) Then the cryptoutils could be a more leaf-like package.
I don't think that will happen. And it won't solve the problem. I think you need md5sum, sha1sum at a very core level. Have a nice day, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 04.12.2013 23:02, schrieb Bernhard Voelker:
On 12/04/2013 10:39 PM, Jan Engelhardt wrote:
Just thinking out loud here - maybe coreutils should be split again at the upstream level. textutils et al :) Then the cryptoutils could be a more leaf-like package.
I don't think that will happen. And it won't solve the problem. I think you need md5sum, sha1sum at a very core level.
While it's rather academical - where do you see *sum at a very core level? I don't see any value while booting, while building other packages, basically nowhere but in cases where you want to put things in the internet. Greetings, Stephan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 12/05/2013 07:25 AM, Stephan Kulow wrote:
Am 04.12.2013 23:02, schrieb Bernhard Voelker:
I don't think that will happen. And it won't solve the problem. I think you need md5sum, sha1sum at a very core level.
While it's rather academical - where do you see *sum at a very core level? I don't see any value while booting, while building other packages, basically nowhere but in cases where you want to put things in the internet.
Okay, perhaps not for booting, but I could imagine that they're used for building. I could imagine e.g. 'make check' comparing expected vs. actual results via md5sum; ... not to mention the typical md5sum check after downloading a package from upstream. Anyway, I personally consider the *sum tools rather core because I'm using them on a daily basis. My personal view, though. Have a nice day, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 05.12.2013 10:47, Bernhard Voelker wrote:
On 12/05/2013 07:25 AM, Stephan Kulow wrote:
Am 04.12.2013 23:02, schrieb Bernhard Voelker:
I don't think that will happen. And it won't solve the problem. I think you need md5sum, sha1sum at a very core level.
While it's rather academical - where do you see *sum at a very core level? I don't see any value while booting, while building other packages, basically nowhere but in cases where you want to put things in the internet.
Okay, perhaps not for booting, but I could imagine that they're used for building. I could imagine e.g. 'make check' comparing expected vs. actual results via md5sum; ... not to mention the typical md5sum Other people use cmp or diff for that ;)
Greetings, Stephan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 12/05/2013 10:58 AM, Stephan Kulow wrote:
On 05.12.2013 10:47, Bernhard Voelker wrote:
On 12/05/2013 07:25 AM, Stephan Kulow wrote:
Am 04.12.2013 23:02, schrieb Bernhard Voelker:
I don't think that will happen. And it won't solve the problem. I think you need md5sum, sha1sum at a very core level.
While it's rather academical - where do you see *sum at a very core level? I don't see any value while booting, while building other packages, basically nowhere but in cases where you want to put things in the internet.
Okay, perhaps not for booting, but I could imagine that they're used for building. I could imagine e.g. 'make check' comparing expected vs. actual results via md5sum; ... not to mention the typical md5sum Other people use cmp or diff for that ;)
Sure. Ironically, there's even one example in coreutils itself [1] ;-) the 'factor' tests factorize ranges of numbers created by seq(1), and the SHA1 of the output is compared to a previously known number. [1] http://git.savannah.gnu.org/cgit/coreutils.git/tree/tests/factor The universe is literally more than diverse. ;-) Have a nice day, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 12/04/2013 07:54 PM, Stephan Kulow wrote:
Am 04.12.2013 18:29, schrieb Bernhard Voelker:
Should I use that new default setting, or should I configure --with-openssl=no ?
Hmm, openssl is indeed not in the cycle, so it would be a change. openssl should be a single addition though from what I see. If it's a problem, we can of course also work with update-alternatives and offer a coreutils-fast. But for now my gut feeling would be to add ssl in the inner loop.
Okay, I can try. If there's a problem, I can still fall back to --with-openssl=no. Thanks! Have a nice day, Berny -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (3)
-
Bernhard Voelker -
Jan Engelhardt -
Stephan Kulow