Re: RFC: Dropping Breeze + Oxygen Qt4 builds after

2020-01-07 Thread A. Wilcox
On Jan 7, 2020, at 3:51 AM, David Edmundson  wrote:
> 
> We're starting to see a lot of preparation for Qt6 porting API which
> we know will be deprecated.
> 
> One place where this is problematic is Breeze + Oxygen QStyles where
> we currently provide Qt4 builds. I don't want this to hold us back and
> be a problem later. It's also a problem for some shadow cleanup
> happening now (D26475)
> 
> I propose that we completely drop the Qt4 builds from Breeze+Oxygen
> after Plasma 5.18.
> 
> We're seeing distros phase Qt4 out, even Debian - and if a distro does
> still want to package these, they will still have the option of
> building the Plasma 5.18 tarball for Qt4 even after Plasma moves to a
> new revision.
> 
> CC'ing packagers to give them a chance to shout.
> 
> David


FWIW, we haven't shipped Qt 4 in over two years. +1 from here.

--
A. Wilcox (Sent from my iPhone - not signed)
Project Lead, Adélie Linux
https://www.adelielinux.org/

D19865: Linux softraid: Define _GNU_SOURCE for pipe2

2019-03-18 Thread A. Wilcox
awilcox created this revision.
Herald added a project: Plasma.
Herald added a subscriber: plasma-devel.
awilcox requested review of this revision.

REVISION SUMMARY
  In a strict POSIX compilation environment, pipe2 is not defined.  Use the 
_GNU_SOURCE FTM to expose it.
  
  Fixes a failure to build ksysguard on Adélie Linux.

REPOSITORY
  R106 KSysguard

REVISION DETAIL
  https://phabricator.kde.org/D19865

AFFECTED FILES
  ksysguardd/Linux/softraid.c

To: awilcox
Cc: plasma-devel, jraleigh, GB_2, ragreen, Pitel, ZrenBot, lesliezhai, 
ali-mohamed, jensreuterberg, abetts, sebas, apol, mart


D4376: Replace long-deprecated getpass(3) call

2017-06-14 Thread A. Wilcox
awilcox abandoned this revision.
awilcox added a comment.


  Since 
https://phabricator.kde.org/R133:a9c44a10b6a60777bf8ca997d3618e66bbd44baa 
conv_legacy is no longer an option for kscreenlocker.  This patch is no longer 
necessary as of Plasma 5.10.

REPOSITORY
  R133 KScreenLocker

REVISION DETAIL
  https://phabricator.kde.org/D4376

To: awilcox
Cc: graesslin, plasma-devel, ZrenBot, spstarr, progwolff, lesliezhai, 
ali-mohamed, jensreuterberg, abetts, sebas, apol, mart, lukas


[Differential] [Commented On] D4376: Replace long-deprecated getpass(3) call

2017-02-01 Thread A. Wilcox
awilcox added a comment.


  I just wanted to add that after a convoluted process (ripping all the Wayland 
checks out of CMakeLists.txt, and only building the kcheckpass target, since 
the screen locker itself does not even build at all...), I have tested that 
this change *does* work correctly on FreeBSD 10.3-RELEASE-p11.

REPOSITORY
  R133 KScreenLocker

REVISION DETAIL
  https://phabricator.kde.org/D4376

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: awilcox
Cc: graesslin, plasma-devel, lesliezhai, ali-mohamed, jensreuterberg, abetts, 
sebas


[Differential] [Commented On] D4376: Replace long-deprecated getpass(3) call

2017-01-31 Thread A. Wilcox
awilcox added a comment.


  > Out of interest: how did you stumble on that code?
  
  Using the musl libc, getpass is not defined unless you enable _GNU_SOURCE:
  
[ 75%] Building C object kcheckpass/CMakeFiles/kcheckpass.dir/kcheckpass.c.o
cd 
/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5_build/kcheckpass
 && /usr/bin/gcc  -D_LARGEFILE64_SOURCE -D_XOPEN_SOURCE=600 
-I/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5_build/kcheckpass
 -I/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5/kcheckpass 
-I/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5_build   
-DQT_NO_DEBUG -DNDEBUG -O2 -mlong-double-64 -ggdb -mcpu=G3 
-fno-omit-frame-pointer  -std=iso9899:1990 -fno-common -Wall -Wextra 
-Wcast-align -Wchar-subscripts -Wformat-security -Wno-long-long -Wpointer-arith 
-Wundef -Wmissing-format-attribute -Wwrite-strings 
-Werror=implicit-function-declaration -std=gnu90 -fvisibility=hidden
-U_REENTRANT -o CMakeFiles/kcheckpass.dir/kcheckpass.c.o -c 
/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5/kcheckpass/kcheckpass.c

/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5/kcheckpass/kcheckpass.c:
 In function ‘conv_legacy’:

/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5/kcheckpass/kcheckpass.c:105:10:
 error: implicit declaration of function ‘getpass’ 
[-Werror=implicit-function-declaration]
  p = getpass(prompt ? prompt : "Password: ");
  ^

/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5/kcheckpass/kcheckpass.c:105:8:
 warning: assignment makes pointer from integer without a cast 
[-Wint-conversion]
  p = getpass(prompt ? prompt : "Password: ");
^
cc1: some warnings being treated as errors
kcheckpass/CMakeFiles/kcheckpass.dir/build.make:62: recipe for target 
'kcheckpass/CMakeFiles/kcheckpass.dir/kcheckpass.c.o' failed
make[2]: *** [kcheckpass/CMakeFiles/kcheckpass.dir/kcheckpass.c.o] Error 1
make[2]: Leaving directory 
'/usr/src/kde-plasma/kscreenlocker-5.8.5/work/kscreenlocker-5.8.5_build'
CMakeFiles/Makefile2:2042: recipe for target 
'kcheckpass/CMakeFiles/kcheckpass.dir/all' failed
make[1]: *** [kcheckpass/CMakeFiles/kcheckpass.dir/all] Error 2
  
  
  
  > If you think there is a risk: better be pedantic in this case. On the other 
hand getdelim should return -1 in error case and then your method returns null. 
So in my book that's good enough error checking.
  
  My concern was on the off-chance that getdelim reads a partial password but 
receives EOF before \n so it returns -1.  But, looking at the standard, there 
is no safe way to read/write to the buffer if the return code is -1, so that is 
actually a moot point.

REPOSITORY
  R133 KScreenLocker

REVISION DETAIL
  https://phabricator.kde.org/D4376

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: awilcox
Cc: graesslin, plasma-devel, lesliezhai, ali-mohamed, jensreuterberg, abetts, 
sebas


[Differential] [Updated, 46 lines] D4376: Replace long-deprecated getpass(3) call

2017-01-31 Thread A. Wilcox
awilcox updated this revision to Diff 10782.
awilcox added a comment.


  I have fixed the style issue with no space between if and bracket.

REPOSITORY
  R133 KScreenLocker

CHANGES SINCE LAST UPDATE
  https://phabricator.kde.org/D4376?vs=10779=10782

REVISION DETAIL
  https://phabricator.kde.org/D4376

AFFECTED FILES
  kscreenlocker-5.8.5/kcheckpass/kcheckpass.c

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: awilcox
Cc: graesslin, plasma-devel, lesliezhai, ali-mohamed, jensreuterberg, abetts, 
sebas


[Differential] [Request, 46 lines] D4376: Replace long-deprecated getpass(3) call

2017-01-31 Thread A. Wilcox
awilcox created this revision.
awilcox set the repository for this revision to R133 KScreenLocker.
Restricted Application added a project: Plasma.
Restricted Application added a subscriber: plasma-devel.

REVISION SUMMARY
  This modernises the kcheckpass utility to stop using the getpass(3) call, 
which was deprecated in SUSv2, removed from POSIX.1-2001, and deprecated in 
glibc 2.19.  It replaces it with a fully standards compliant getdelim(3) call.  
This also avoids the need to strdup the password buffer and temporarily have 
two copies.
  
  I didn't know how pedantic to make it; you could possibly want to check for 
password != NULL and memset-nul it out in the case of getdelim failure, since 
it could have read in a partial password but then received EOF before \n.  I 
didn't think this case was very likely so I did not author such a check.

TEST PLAN
  - Tested against shadow backend and PAM backend.
  - Each backend was tested on x86_64, x86_32, and PowerPC.
  - Both glibc and musl libc were tested.  FreeBSD testing is pending.

REPOSITORY
  R133 KScreenLocker

REVISION DETAIL
  https://phabricator.kde.org/D4376

AFFECTED FILES
  kscreenlocker-5.8.5/kcheckpass/kcheckpass.c

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: awilcox
Cc: plasma-devel, lesliezhai, ali-mohamed, jensreuterberg, abetts, sebas