Hi guys,
I wanted just ask about the latest fileinfo fixes as I see so far nothing
was merged.
This fixes should be merged into the current 5.3.11 and 5.4.1 branches.
The info about the changes is in the following tickets:
https://bugs.php.net/bug.php?id=61565
https://bugs.php.net/bug.php?id=61566
Regards
Anatoliy
Another important thing i didn't mention - there is a lot of test fixes
going to be commited, they should be merged into release branches too.
thanks )
Hi guys,
I wanted just ask about the latest fileinfo fixes as I see so far nothing
was merged.This fixes should be merged into the current 5.3.11 and 5.4.1 branches.
The info about the changes is in the following tickets:https://bugs.php.net/bug.php?id=61565
https://bugs.php.net/bug.php?id=61566Regards
Anatoliy
Hi!
Another important thing i didn't mention - there is a lot of test fixes
going to be commited, they should be merged into release branches too.
Test fixes are definitely not critical bugs. So I would prefer handling
them as usual fixes.
--
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
hi Stas,
This is zero code or tests change. This is pure restoring what we are
supposed to have, like it was in svn.
However, as you won't merge the tests fixes either, we won't 5.4.1's
tests at all to valid the build (way too many false positive), it does
not really matter anymore.
Cheers,
Hi!
Another important thing i didn't mention - there is a lot of test fixes
going to be commited, they should be merged into release branches too.Test fixes are definitely not critical bugs. So I would prefer handling
them as usual fixes.--
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227--
--
Pierre
@pierrejoye | http://blog.thepimp.net | http://www.libgd.org
Hi!
However, as you won't merge the tests fixes either, we won't 5.4.1's
tests at all to valid the build (way too many false positive), it does
not really matter anymore.
Wait, are we talking about new test fixes or about tests that are broken
in RC1 and need fixing? I was assuming it's the former but maybe I miss
the point entirely. I've run tests of RC1 on Linux and those were just
fine. If they fail somewhere else please tell me and I'll see if I
should add those.
I'm ok with pulling the LF fix, that seems innocent enough - it's not
even a code change, just a metadata change.
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
hi Stas,
Hi!
However, as you won't merge the tests fixes either, we won't 5.4.1's
tests at all to valid the build (way too many false positive), it does
not really matter anymore.Wait, are we talking about new test fixes or about tests that are broken
in RC1 and need fixing? I was assuming it's the former but maybe I miss
the point entirely. I've run tests of RC1 on Linux and those were just
fine. If they fail somewhere else please tell me and I'll see if I
should add those.
I'm ok with pulling the LF fix, that seems innocent enough - it's not
even a code change, just a metadata change.
btw, It is not a realy issue as we won't use 5.4.1 tests to test the
release but 5.4's branch ones. Too many false positive.
The lf should not be a problem either as the releases are packaged on
linux, so LF is the default there.
All in all, I do not see a critical stopping preventing us to go final
(and begin with next soonish :) ).
Cheers,
Pierre
@pierrejoye | http://blog.thepimp.net | http://www.libgd.org
Hi!
btw, It is not a realy issue as we won't use 5.4.1 tests to test the
release but 5.4's branch ones. Too many false positive.
Who are we? I'll definitely be using 5.4.1, since that's what I will be
releasing. There's no point in testing code different from one you're
releasing. Do you mean windows tests? Are they broken now?
The lf should not be a problem either as the releases are packaged on
linux, so LF is the default there.
I'll merge the .gitattributes patch, since it doesn't hurt anything, but
not sure about other stuff - still don't understand what's going on
there and if we need another RC.
--
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
Hi!
Hi!
btw, It is not a realy issue as we won't use 5.4.1 tests to test the
release but 5.4's branch ones. Too many false positive.Who are we?
My team and I, as always when we refer to that.
I'll definitely be using 5.4.1, since that's what I will be
releasing. There's no point in testing code different from one you're
releasing. Do you mean windows tests? Are they broken now?
Yes, see the commits, and the LF thing.
The lf should not be a problem either as the releases are packaged on
linux, so LF is the default there.I'll merge the .gitattributes patch, since it doesn't hurt anything, but
not sure about other stuff - still don't understand what's going on
there and if we need another RC
We do not need another RC for the tests fixes only. It is not optimal
but not critical enough to delay the release any longer. We will
simply use 5.4's branch tests for now as there are far less failures
(related to tests fixes, not code changes).
Also we have to release 5.3.11 final as well (same security fixes) but
we did not hear anything yet from Johannes so far. Johannes?
Cheers,
Pierre
@pierrejoye | http://blog.thepimp.net | http://www.libgd.org
Hi!
This fixes should be merged into the current 5.3.11 and 5.4.1 branches.
The info about the changes is in the following tickets:https://bugs.php.net/bug.php?id=61565
https://bugs.php.net/bug.php?id=61566
Are those critical bugs that break all fileinfo functionality?
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
Hi,
Hi!
This fixes should be merged into the current 5.3.11 and 5.4.1 branches.
The info about the changes is in the following tickets:https://bugs.php.net/bug.php?id=61565
https://bugs.php.net/bug.php?id=61566Are those critical bugs that break all fileinfo functionality?
those are not critical, but they break some tests. Thease bugs seem to be
already present before, nonetheless they've been fixed after libmagic
security upgrade to 5.11
Regards
Anatoliy
Hi,
those are not critical, but they break some tests. Thease bugs seem to be
already present before, nonetheless they've been fixed after libmagic
security upgrade to 5.11
They are critical as they break BC with previous versions. The
breakages are introduced with the previous security fix. Please merge
it.
Thanks,
Pierre
@pierrejoye | http://blog.thepimp.net | http://www.libgd.org
Hi!
They are critical as they break BC with previous versions. The
breakages are introduced with the previous security fix. Please merge
it.
Could you please describe what exactly was broken?
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
hi,
On another area, we cruelly need:
http://nebm.ist.utl.pt/~glopes/misc/lf_phpt
applied to all branches, incl. 5.3.11/5.4.1.
That restores the force LF EOL we did in SVN.
Cheers,
Hi!
They are critical as they break BC with previous versions. The
breakages are introduced with the previous security fix. Please merge
it.Could you please describe what exactly was broken?
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
--
Pierre
@pierrejoye | http://blog.thepimp.net | http://www.libgd.org
Hi!
On another area, we cruelly need:
http://nebm.ist.utl.pt/~glopes/misc/lf_phpt
applied to all branches, incl. 5.3.11/5.4.1.
I'm not sure I understand what is that and why we can't release 5.4.1
without it? I see a list of files in that link - what are these files,
what is supposed to be done with them?
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227
On Tue, 10 Apr 2012 18:32:37 +0100, Stas Malyshev smalyshev@sugarcrm.com
wrote:
On another area, we cruelly need:
http://nebm.ist.utl.pt/~glopes/misc/lf_phpt
applied to all branches, incl. 5.3.11/5.4.1.
I'm not sure I understand what is that and why we can't release 5.4.1
without it? I see a list of files in that link - what are these files,
what is supposed to be done with them?
Those were the files with svn:eol-style = LF in SVN. I've committed a
change to .gitattributes that disables EOL conversion in those files. I'm
told this could also be accomplished with a eol=lf attribute.
See
https://github.com/php/php-src/commit/112a476b683a634390b23fe7509d5b73632d0829
(5.4 and master)
and
https://github.com/php/php-src/commit/1c8fccdf6d6d04d8e6a5dedd7c9d6b1afda8839f
(5.3)
--
Gustavo Lopes
Hi!
Those were the files with svn:eol-style = LF in SVN. I've committed a
change to .gitattributes that disables EOL conversion in those files. I'm
told this could also be accomplished with a eol=lf attribute.See
https://github.com/php/php-src/commit/112a476b683a634390b23fe7509d5b73632d0829
(5.4 and master)
OK, I see now, thanks.
--
Stanislav Malyshev, Software Architect
SugarCRM: http://www.sugarcrm.com/
(408)454-6900 ext. 227