Currently this problem belongs to QSerialPort which internally uses
QLockFile to indication of employment of the /dev/ttyXYZ resources.> is this
still a problem on modern unixes?
> is this still a problem on modern unixes?
Yes, of course, it still is a problem...
There are one TIOCEXCL flag t
Here you go :)
https://codereview.qt-project.org/166278
Simon
From: Development on
behalf of Denis Shienkov
Sent: Thursday, July 28, 2016 8:34:03 AM
To: development@qt-project.org
Subject: Re: [Development] [QLockFile] Whether it is reasonable to use 0644
So, as I understand, you agree that QLockFile should be fixed.. Who will
produce a patch? Thiago, maybe you can do it? :)
PS: Because I do not understand an idea with the umask... I simply would
replace 0644 with 0666 in sources... and nothing more..
28.07.2016 0:23, Thiago Macieira пишет:
O
> That's because we removed the bundled copy of libmng.
Then, please, adjust the documentation, because currently it confused users:
http://doc.qt.io/qt-5/qtimageformats-index.html
since MNG is not 'bundled' anymore.
> You need to install it yourself first.
Ohh... seems it is problem (especia
On Wed, Jul 27, 2016 at 11:00 PM, Giuseppe D'Angelo wrote:
>
>> http://testresults.qt.io/logs/qt/qtbase/ddfe2ba984be5d652fe500f6695040959db98fa4/OSXOSX_10_11x86_64OSXOSX_10_11x86_64Clangqtci-osx-10.11-x86_64DebugAndRelease_Release/85d6b000f945a84bc84a4f01f53ac65bc05cbe86/buildlog.txt.gz
>
> And ag
On quarta-feira, 27 de julho de 2016 23:00:36 PDT Giuseppe D'Angelo wrote:
> On Tue, Jul 26, 2016 at 11:24 PM, Giuseppe D'Angelo
wrote:
> > It just happened again:
> >
> > http://testresults.qt.io/logs/qt/qtbase/ddfe2ba984be5d652fe500f6695040959d
> > b98fa4/OSXOSX_10_11x86_64OSXOSX_10_11x86_64Cl
On quarta-feira, 27 de julho de 2016 22:15:45 PDT Kevin Kofler wrote:
> Thiago Macieira wrote:
> > If you don't have permission to delete it, then you're going to have a
> > problem to drop your own lock if you steal the lock by writing to the
> > file.
>
> So this means that the permission should
On Tue, Jul 26, 2016 at 11:24 PM, Giuseppe D'Angelo wrote:
>
>
> It just happened again:
>
> http://testresults.qt.io/logs/qt/qtbase/ddfe2ba984be5d652fe500f6695040959db98fa4/OSXOSX_10_11x86_64OSXOSX_10_11x86_64Clangqtci-osx-10.11-x86_64DebugAndRelease_Release/85d6b000f945a84bc84a4f01f53ac65bc05cbe
Thiago Macieira wrote:
> If you don't have permission to delete it, then you're going to have a
> problem to drop your own lock if you steal the lock by writing to the
> file.
So this means that the permission should be 0644 to prevent that from
happening. Thus, I don't understand how you get to:
On quarta-feira, 27 de julho de 2016 11:47:03 PDT BogDan Vatra wrote:
> On miercuri, 27 iulie 2016 10:35:01 EEST André Somers wrote:
> > Op 27/07/2016 om 09:53 schreef BogDan Vatra:
> > Would that not break existing QML applications that now use enums backed
> > by scoped enums in C++?
>
> It will
On quarta-feira, 27 de julho de 2016 09:53:32 PDT Jędrzej Nowacki wrote:
> No, the machine gets destroyed after test failure. Normally gdb session is
> attached to crashing process, that was not the case for some odd reason.
> Can it be that QTestLib is crashing too?
There have been tools to col
On quarta-feira, 27 de julho de 2016 15:06:29 PDT Denis Shienkov wrote:
> Ahh... seems, the config.tests for MNG fails..
That's because we removed the bundled copy of libmng. You need to install it
yourself first.
We did that because our copy was very old and, at Qt 5.6 times, we thought
that l
On quarta-feira, 27 de julho de 2016 11:37:51 PDT Oswald Buddenhagen wrote:
> also, isn't this actually mostly about the permissions of the directory
> containing the locks? lock grabbing should be mostly about deleting and
> re-creating the file.
That's also a good point. If you have permission t
On quarta-feira, 27 de julho de 2016 10:31:35 PDT Denis Shienkov wrote:
> > and let the umask determine which bits to exclude.
>
> what do you mean here? why we need to use umask, and why we need to
> exclude bits?
Because we don't have a choice. The umask bits apply when you create a fie.
--
On Wednesday 27 of July 2016 13:58:44 Dominik Holland wrote:
> Hi,
>
> Am 07/27/2016 um 01:47 PM schrieb Jędrzej Nowacki:
> > Hi,
> >
> > Good news everyone! Last month we managed to deploy a new automatic
> >
> > provisioning system to Coin. What does it mean to you? Now you can modify
> > ex
Ahh... seems, the config.tests for MNG fails..
2016-07-27 15:03 GMT+03:00 Denis Shienkov :
> Hmm.. Now I have cloned the "imageformats" source module, open its *.pro
> file from QtC, and re-build it...
>
> BUT: I too do not see the output qmng.dll target, though, I see this
> 'mng' plugin in the
Hmm.. Now I have cloned the "imageformats" source module, open its *.pro
file from QtC, and re-build it...
BUT: I too do not see the output qmng.dll target, though, I see this 'mng'
plugin in the source's tree...
Seems it did not compiles.. o_O
2016-07-27 14:47 GMT+03:00 Denis Shienkov :
> Hi al
Hi,
Am 07/27/2016 um 01:47 PM schrieb Jędrzej Nowacki:
> Hi,
>
> Good news everyone! Last month we managed to deploy a new automatic
> provisioning system to Coin. What does it mean to you? Now you can modify
> existing VM templates used by Coin. Installing additional packages should be
> tr
Hi,
Good news everyone! Last month we managed to deploy a new automatic
provisioning system to Coin. What does it mean to you? Now you can modify
existing VM templates used by Coin. Installing additional packages should be
trivial and more or less safe from now.
I. The goal:
- Coin shoul
Hi all.
I want to use an MNG animation, using official Qt 5.7 from MSVC2015
installer..
BUT: I do not see in 5.7\msvc2015\plugins\imageformats\ apropriate qmng.dll
(or something like) plugin, so leads to fails with the QML's AnimatedSprite
item, with error:
: QML Sprite: Error decoding: file:///
On Wed, Jul 27, 2016 at 12:01:20AM -0700, Thiago Macieira wrote:
> On quarta-feira, 27 de julho de 2016 09:49:14 PDT Denis Shienkov wrote:
> > Why just do not use 0666 permissions for this?
>
> We should use 0666 and let the umask determine which bits to exclude.
>
how exactly is this response su
On Wed, Jul 27, 2016 at 09:49:14AM +0300, Denis Shienkov wrote:
> Currently this problem belongs to QSerialPort which internally uses
> QLockFile to indication of employment of the /dev/ttyXYZ resources.
>
well, that's kinda the problem. this approach is a hack to work around
inadequate locking on
I've seen that patch, and it is definitely a good starter. I was able to
change the screen and run two screens simultaneously using Raspbian's
Dispmanx example (one screen was connected through the PI's DSI port the
other one through HDMI) quite easily, so I think it does worth poking
around with.
On Wed, Jul 27, 2016 at 10:47 AM, BogDan Vatra wrote:
> Of course if we can find a better solution to have them both is even better !
I guess you'll need a pragma to enable this feature as opt-in. And/or
a pragma to warn about the other usage so that in 5.10/5.11 you can
break that code.
--
Giu
On miercuri, 27 iulie 2016 10:35:01 EEST André Somers wrote:
> Op 27/07/2016 om 09:53 schreef BogDan Vatra:
> >> That said, it would be a behavior change not to allow anymore the
> >> previous
> >> (unscoped) syntax to work with QML
> >
> > The unscoped enums should work as they are woring now, ju
Op 27/07/2016 om 09:53 schreef BogDan Vatra:
That said, it would be a behavior change not to allow anymore the previous
(unscoped) syntax to work with QML
The unscoped enums should work as they are woring now, just the scoped ones
should be scoped in QML as well.
Based on your patch, I'll t
On Wednesday 27 of July 2016 08:28:08 Friedemann Kleint wrote:
> Hi,
>
> to summarize for Windows:
>
> - Crashes have been observed over time in various tests, mostly in
> QtCore (QDir,QFile, QProcess see fex
> https://bugreports.qt.io/browse/QTBUG-47370
>
> - It only affects 64bit (!) (correct
On Tuesday 26 of July 2016 15:30:48 Thiago Macieira wrote:
> Em terça-feira, 26 de julho de 2016, às 23:24:21 PDT, Giuseppe D'Angelo
>
> escreveu:
> > On Mon, Jul 25, 2016 at 5:51 PM, Thiago Macieira
> >
> > wrote:
> > > Let's pay attention of what crashes and where. But we'll need to
> > > repr
On marți, 26 iulie 2016 20:08:52 EEST Olivier Goffart wrote:
> On Dienstag, 26. Juli 2016 08:33:31 CEST Thiago Macieira wrote:
> > Em terça-feira, 26 de julho de 2016, às 17:50:38 PDT, BogDan Vatra
escreveu:
> > > Hi,
> > >
> > > Is there any reason why C++11 scoped enums are not scoped in QML ?
FWIW this is slightly related to https://bugreports.qt.io/browse/QTBUG-53570
On Wed, Jul 27, 2016 at 9:31 AM, Denis Shienkov
wrote:
>> and let the umask determine which bits to exclude.
>
> what do you mean here? why we need to use umask, and why we need to exclude
> bits?
>
>
> 27.07.2016 10:01,
> and let the umask determine which bits to exclude.
what do you mean here? why we need to use umask, and why we need to
exclude bits?
27.07.2016 10:01, Thiago Macieira пишет:
On quarta-feira, 27 de julho de 2016 09:49:14 PDT Denis Shienkov wrote:
Why just do not use 0666 permissions for th
On quarta-feira, 27 de julho de 2016 09:49:14 PDT Denis Shienkov wrote:
> Why just do not use 0666 permissions for this?
We should use 0666 and let the umask determine which bits to exclude.
--
Thiago Macieira - thiago.macieira (AT) intel.com
Software Architect - Intel Open Source Technology C
32 matches
Mail list logo
