[1.55][log] Last minute bug fix
Hi, Is it too late for a fix in Boost.Log in 1.55? Here's the commit: https://svn.boost.org/trac/boost/changeset/86600 This fixes garbage output when using dump() manipulator on CPUs with AVX2 support. The bug didn't appear on any of our testers because apparently none of them have Haswell CPUs. I tested this change locally on Linux, gcc 4.8, but it should be valid on all platforms. If it's too late already then maybe a note with the patch in the release notes would be possible?
On 9 November 2013 16:07, Andrey Semashev
Hi,
Is it too late for a fix in Boost.Log in 1.55? Here's the commit:
https://svn.boost.org/trac/boost/changeset/86600
This fixes garbage output when using dump() manipulator on CPUs with AVX2 support. The bug didn't appear on any of our testers because apparently none of them have Haswell CPUs. I tested this change locally on Linux, gcc 4.8, but it should be valid on all platforms.
If it's too late already then maybe a note with the patch in the release notes would be possible?
I don't know what the current release plans are, but adding a patch to the release notes is always possible.
On Sunday 10 November 2013 23:11:39 Daniel James wrote:
On 9 November 2013 16:07, Andrey Semashev
wrote: Hi,
Is it too late for a fix in Boost.Log in 1.55? Here's the commit:
https://svn.boost.org/trac/boost/changeset/86600
This fixes garbage output when using dump() manipulator on CPUs with AVX2 support. The bug didn't appear on any of our testers because apparently none of them have Haswell CPUs. I tested this change locally on Linux, gcc 4.8, but it should be valid on all platforms.
If it's too late already then maybe a note with the patch in the release notes would be possible?
I don't know what the current release plans are, but adding a patch to the release notes is always possible.
Release managers are not responding, and the release is behind the schedule, so I'm guessing I missed 1.55 (did I?). Could you add the attached patch and the following description to the release notes? The patch fixes incorrect output of the [@http://www.boost.org/doc/libs/1_55_0/libs/log/doc/html/log/detailed/utilitie... dump] manipulator, when used on AVX2-enabled CPU (e.g. Intel Haswell).
On 11 November 2013 08:05, Andrey Semashev
Release managers are not responding, and the release is behind the schedule, so I'm guessing I missed 1.55 (did I?).
Looks like it.
Could you add the attached patch and the following description to the release notes?
The patch fixes incorrect output of the [@http://www.boost.org/doc/libs/1_55_0/libs/log/doc/html/log/detailed/utilitie... dump] manipulator, when used on AVX2-enabled CPU (e.g. Intel Haswell).
I've just added it.
On Monday 11 November 2013 20:23:14 Daniel James wrote:
On 11 November 2013 08:05, Andrey Semashev
wrote: Release managers are not responding, and the release is behind the schedule, so I'm guessing I missed 1.55 (did I?).
Looks like it.
Could you add the attached patch and the following description to the release notes?
The patch fixes incorrect output of the [@http://www.boost.org/doc/libs/1_55_0/libs/log/doc/html/log/detailed/util ities.html#log.detailed.utilities.manipulators.dump> dump] manipulator, when used on AVX2-enabled CPU (e.g. Intel Haswell).
I've just added it.
Thanks. I've applied the patch to the maintenance branch. Do you think that we should mention the maintenance branch in the release notes?
On 12 November 2013 02:35, Andrey Semashev
Do you think that we should mention the maintenance branch in the release notes?
No, not really. In the last release I just used the maintenance branch to manage the patches. And there's only a week before subversion is turned off.
participants (2)
-
Andrey Semashev
-
Daniel James