Re: libgit2-dev and detection in subsurface/scripts/build.sh

[Cristian Ionescu-Idbohrn via subsurface]

On Wed, 14 Oct 2020, Dirk Hohndel via subsurface wrote:
> 
> I actually think that this isn't a bad idea. Except that it also 
> should have a clause that allows ldconfig to be found in the PATH, 
> right?

Sure, but are there distributions out there that place ldconfig in 
the non-root PATH?

> Would you mind submitting a PR for that, or would you prefer that I 
> do this based on your idea?

I think it's best you train your shell scripting skills a bit ;)


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: fatal error: QPrintDialog: No such file or directory

[Cristian Ionescu-Idbohrn via subsurface]

On Sun, 11 Oct 2020, Robert Helling via subsurface wrote:
> 
> Building in a fresh build directory without webkit option I get the 
> error from
> 
> #ifndef NO_PRINTING
> #include 
> #include 
> #include "desktop-widgets/printdialog.h"
> #endif
> 
> Somehow, NO_PRINTING seems not to be defined.

It appears Commit 48adaf52971965a0e1465ea834cbe930d3c65ec4 did the 
trick.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

libgit2-dev and detection in subsurface/scripts/build.sh

[Cristian Ionescu-Idbohrn via subsurface]

Debian unstable here.

I did not have libgit2-dev installed and noticed this in build.log:

./subsurface/scripts/build.sh: line 268: ldconfig: command not found
realpath: missing operand
Try 'realpath --help' for more information.

That's because ldconfig is installed in /sbin/ldconfig and a non-root 
user does not normally have access to /sbin.  Thus the detection will 
always fail.  So, some sort of detection would be needed.  Something 
like:

if [ -x /sbin/ldconfig ]; then
LDCONFIG=/sbin/ldconfig
elif [ -x /usr/sbin/ldconfig ]; then
LDCONFIG=/usr/sbin/ldconfig
else
echo "no acces to ldconfig" >&2
exit 1
fi

LIBGIT=$(realpath $($LDCONFIG -p | grep libgit2\\.so\\. |
cut -d\  -f4) | awk -Fo '{ print $NF }')

maybe?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: fatal error: QPrintDialog: No such file or directory

[Cristian Ionescu-Idbohrn via subsurface]

On Sun, 11 Oct 2020, Dirk Hohndel wrote:
> 
> I'll try again.

Just did.  Same error:

[ 56%] Building CXX object 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o
.../subsurface/desktop-widgets/mainwindow.cpp:59:10: fatal error: QPrintDialog: 
No such file or directory
   59 | #include 
  |  ^~
compilation terminated.
make[2]: *** 
[desktop-widgets/CMakeFiles/subsurface_interface.dir/build.make:477: 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o] Error 1
make[2]: *** Waiting for unfinished jobs

> For me, on both Linux and Mac, all I needed to do in the build 
> directory was to run 'cmake .' and then a make completed fine.

$ cmake .
CMake Deprecation Warning at CMakeLists.txt:11 (cmake_policy):
  The OLD behavior for policy CMP0071 will be removed from a future version
  of CMake.

  The cmake-policies(7) manual explains that the OLD behaviors of all
  policies are deprecated and that a policy should be set to OLD only under
  specific short-term circumstances.  Projects should be ported to the NEW
  behavior and not rely on setting a policy to OLD.


-- GCC optimization flags: -O2
-- building without usermanual
-- system name Linux
-- Configuring done
-- Generating done
-- Build files have been written to: .../subsurface/build

$ make
[...]
[ 49%] Building CXX object 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o
.../subsurface/desktop-widgets/mainwindow.cpp:59:10: fatal error: QPrintDialog: 
No such file or directory
   59 | #include 
  |  ^~
compilation terminated.

> I can also tell from the test builds that a fresh checkout works fine.

Not here :(

[ 56%] Building CXX object 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o
/tmp/dive/subsurface/desktop-widgets/mainwindow.cpp:59:10: fatal error: 
QPrintDialog: No such file or directory
   59 | #include 
  |  ^~
compilation terminated.
make[2]: *** 
[desktop-widgets/CMakeFiles/subsurface_interface.dir/build.make:477: 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o] Error 1

> so you actually do an 'rm -rf subsurface/build' and then run the
> build script and you still get that error?

Yes.

> That's... puzzling. And rather surprising.

Yes.

> But it's of course entirely possible that I messed something up when
> making this change.

As mentioned, after checking out commit b3e28eed5d08cbbbcc640875f9b0f2e5f0af3506
builds fine, but commit 7f896ee8d50152e85c6f40ff800c5ad8f9ba4124
breaks my build with the above error.  It appears this 
/usr/include/x86_64-linux-gnu/qt5/QtPrintSupport header directory is 
missing from the build.

> Which OS are you? Maybe I can reproduce this and then fix it.

Debian unstable or testing.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: *** SPAM *** (was: Re: fatal error: QPrintDialog: No such file or directory)

[Cristian Ionescu-Idbohrn via subsurface]

On Sat, 10 Oct 2020, Robert Helling via subsurface wrote:
> On 10. Oct 2020, at 18:46, Dirk Hohndel via subsurface 
> mailto:subsurface@subsurface-divelog.org>> 
> wrote:
> 
> This is indeed a result of the recent removal of Grantlee and then 
> not rerunning cmake.

Not sure I'm following.

> In theory that should happen automatically through Makefile 
> dependency, but my guess is something went wrong there for you ;-)

Yes.

> Can you remove your build directory and start again running 
> subsurface/scripts/build.sh from the directory 'above' the 
> subsurface directory?

That's exactly what I did and stated.

> All the tests passed, but they of course always build from scratch...

And I did that too.

> as I wrote in the discussion of your pull request, I had the same 
> error message when running the build script from a fresh checkout 
> without the —with-webkit option. It should be „no webkit“ => „no 
> printing“ => this file should not be included.

That should be clearly documented, else it will confuse people.
So, what do I need to do besides stating with a clean clone?

> Busy right now, But could look into this a bit later.

Alright.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: fatal error: QPrintDialog: No such file or directory

[Cristian Ionescu-Idbohrn via subsurface]

On Sat, 10 Oct 2020, Dirk Hohndel via subsurface wrote:
> > On Oct 10, 2020, at 1:55 AM, Cristian Ionescu-Idbohrn via subsurface 
> >  wrote:
> > 
> > Anyone else experiences this?  Does this have anything to do with the
> > recent effort to "remove Grantlee references"?  It would seem so, as 
> > the code builds on commit b3f5473b66b8210d1276f197876e0e2d7bbd6fa7.
> > 
> > The error message is:
> > 
> > .../subsurface/desktop-widgets/mainwindow.cpp:59:10: fatal error: 
> > QPrintDialog: No such file or directory
> >   59 | #include 
> >  |  ^~
> > compilation terminated.
> > make[2]: *** 
> > [desktop-widgets/CMakeFiles/subsurface_interface.dir/build.make:477: 
> > desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o] Error 
> > 1
> > 
> > Both QPrintDialog and qprintdialog.h files exist and belong to package
> > qtbase5-dev (version 5.14.2+dfsg-6, debian unstable):
> > 
> > $ ls -l 
> > /usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/{QPrintDialog,qprintdialog.h}
> > -rw-r--r-- 1 root root   26 Mar 27  2020 
> > /usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/QPrintDialog
> > -rw-r--r-- 1 root root 3589 Mar 27  2020 
> > /usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/qprintdialog.h
> > 
> > and it appears they have been there for a while, or moved there from
> > elsewhere.
> 
> 
> This is indeed a result of the recent removal of Grantlee and then 
> not rerunning cmake.
> In theory that should happen automatically through Makefile 
> dependency, but my guess is something went wrong there for you ;-)
> 
> Can you remove your build directory and start again running 
> subsurface/scripts/build.sh from the directory 'above' the 
> subsurface directory?
> All the tests passed, but they of course always build from 
> scratch...

Thing is I got the same result after cleaning up my build tree as 
after cloning a brand new one.  As said, backing to commit 
b3f5473b66b8210d1276f197876e0e2d7bbd6fa7 solved the problem.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

fatal error: QPrintDialog: No such file or directory

[Cristian Ionescu-Idbohrn via subsurface]

Anyone else experiences this?  Does this have anything to do with the
recent effort to "remove Grantlee references"?  It would seem so, as 
the code builds on commit b3f5473b66b8210d1276f197876e0e2d7bbd6fa7.

The error message is:

.../subsurface/desktop-widgets/mainwindow.cpp:59:10: fatal error: QPrintDialog: 
No such file or directory
   59 | #include 
  |  ^~
compilation terminated.
make[2]: *** 
[desktop-widgets/CMakeFiles/subsurface_interface.dir/build.make:477: 
desktop-widgets/CMakeFiles/subsurface_interface.dir/mainwindow.cpp.o] Error 1

Both QPrintDialog and qprintdialog.h files exist and belong to package
qtbase5-dev (version 5.14.2+dfsg-6, debian unstable):

$ ls -l 
/usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/{QPrintDialog,qprintdialog.h}
-rw-r--r-- 1 root root   26 Mar 27  2020 
/usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/QPrintDialog
-rw-r--r-- 1 root root 3589 Mar 27  2020 
/usr/include/x86_64-linux-gnu/qt5/QtPrintSupport/qprintdialog.h

and it appears they have been there for a while, or moved there from
elsewhere.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: libqtgeoservices_googlemaps.so: undefined symbol: _ZN7QGeoMap25fitViewportToGeoRectangleERK13QGeoRectangle)

[Cristian Ionescu-Idbohrn via subsurface]

On Thu, 1 Oct 2020, Dirk Hohndel wrote:
> > On Oct 1, 2020, at 6:18 AM, Cristian Ionescu-Idbohrn via subsurface 
> >  wrote:
> > 
> > On Thu, 1 Oct 2020, JB2Cool wrote:
> >> 
> >> Strange, i built this just a few days ago on lots of different 
> >> distros (Debian Buster (albeit stable) was one of them) and it built 
> >> just fine. Can you try cloning the repo to a fresh directory and 
> >> retrying a build and see if that is any better. Googling around some 
> >> have said that by ditching any previous folders and starting afresh 
> >> this worked ok.
> > 
> > As said...
> 
> One thing that often goes wrong after a distro upgrade with new Qt 
> version is that people rebuild Subsurface, but not the other libs. 
> Can you remove the install-root directory, as well as the build 
> directories under googllemaps and grantee Then run the build.sh 
> script again to ensure those libraries are rebuilt against the 
> current Qt version

No grantee here, but removing the install-root directory seems to have 
done the trick.  So, this was a runtime, not buildtime, problem.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: libqtgeoservices_googlemaps.so: undefined symbol: _ZN7QGeoMap25fitViewportToGeoRectangleERK13QGeoRectangle)

[Cristian Ionescu-Idbohrn via subsurface]

On Thu, 1 Oct 2020, JB2Cool wrote:
> 
> Strange, i built this just a few days ago on lots of different 
> distros (Debian Buster (albeit stable) was one of them) and it built 
> just fine. Can you try cloning the repo to a fresh directory and 
> retrying a build and see if that is any better. Googling around some 
> have said that by ditching any previous folders and starting afresh 
> this worked ok.

As said...

> On Thu, 1 Oct 2020 at 09:07, Cristian Ionescu-Idbohrn via subsurface 
> mailto:subsurface@subsurface-divelog.org>> 
> wrote:
> Any ideas?  This is debian unstable, own build from fresh master:
...^^^

I have no clues :(

$ readelf -a googlemaps/build/libqtgeoservices_googlemaps.so | grep 
_ZN7QGeoMap25fit
0002f8c8  00620001 R_X86_64_64    
_ZN7QGeoMap25fitV[...] + 0
98:  0 NOTYPE  GLOBAL DEFAULT  UND _ZN7QGeoMap25fit[...]
   414:  0 NOTYPE  GLOBAL DEFAULT  UND _ZN7QGeoMap25fit[...]

$ readelf -a googlemaps/build/libqtgeoservices_googlemaps.so | grep NEEDED  
 
 0x0001 (NEEDED) Shared library: [libQt5Location.so.5]
 0x0001 (NEEDED) Shared library: [libQt5Gui.so.5]
 0x0001 (NEEDED) Shared library: 
[libQt5Positioning.so.5]
 0x0001 (NEEDED) Shared library: [libQt5Network.so.5]
 0x0001 (NEEDED) Shared library: [libQt5Core.so.5]
 0x0001 (NEEDED) Shared library: [libstdc++.so.6]
 0x0001 (NEEDED) Shared library: [libc.so.6]

$ dpkg -l libqt5location5 libqt5positioning5 qtlocation5-dev qtpositioning5-dev 
libqt5gui5 libqt5network5 libqt5core5a | cat
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version   Architecture Description
+++--=--==
ii  libqt5core5a:amd64   5.14.2+dfsg-6 amd64Qt 5 core module
ii  libqt5gui5:amd64 5.14.2+dfsg-6 amd64Qt 5 GUI module
ii  libqt5location5:amd645.14.2+dfsg-2 amd64Qt Location module
ii  libqt5network5:amd64 5.14.2+dfsg-6 amd64Qt 5 network module
ii  libqt5positioning5:amd64 5.14.2+dfsg-2 amd64Qt Positioning module
ii  qtlocation5-dev  5.14.2+dfsg-2 amd64Qt 5 Location 
development files
ii  qtpositioning5-dev:amd64 5.14.2+dfsg-2 amd64Qt 5 Positioning 
development files


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

libqtgeoservices_googlemaps.so: undefined symbol: _ZN7QGeoMap25fitViewportToGeoRectangleERK13QGeoRectangle

[Cristian Ionescu-Idbohrn via subsurface]

Any ideas?  This is debian unstable, own build from fresh master:

...
Got keys from plugin meta data ("here")
QFactoryLoader::QFactoryLoader() looking at 
"/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_osm.so"
Found metadata in lib 
/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_osm.so, 
metadata=
{
"IID": "org.qt-project.qt.geoservice.serviceproviderfactory/5.0",
"MetaData": {
"Experimental": false,
"Features": [
"OnlineMappingFeature",
"OnlineGeocodingFeature",
"ReverseGeocodingFeature",
"OnlineRoutingFeature",
"OnlinePlacesFeature"
],
"Keys": [
"osm"
],
"Provider": "osm",
"Version": 100
},
"archreq": 0,
"className": "QGeoServiceProviderFactoryOsm",
"debug": false,
"version": 331264
}


Got keys from plugin meta data ("osm")
QFactoryLoader::QFactoryLoader() checking directory path 
".../subsurface/build/geoservices" ...
QFactoryLoader::QFactoryLoader() looking at 
".../subsurface/build/geoservices/libqtgeoservices_googlemaps.so"
Found metadata in lib 
.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so,
 metadata=
{
"IID": "org.qt-project.qt.geoservice.serviceproviderfactory/5.0",
"MetaData": {
"Experimental": false,
"Features": [
"OnlineGeocodingFeature",
"ReverseGeocodingFeature",
"OnlineRoutingFeature",
"AlternativeRoutesFeature",
"OnlineMappingFeature",
"SearchSuggestionsFeature"
],
"Keys": [
"googlemaps"
],
"Provider": "googlemaps",
"Version": 100
},
"archreq": 0,
"className": "QGeoServiceProviderFactoryGooglemaps",
"debug": false,
"version": 331264
}


Got keys from plugin meta data ("googlemaps")
Cannot load library 
.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so:
 
(.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so:
 undefined symbol: _ZN7QGeoMap25fitViewportToGeoRectangleERK13QGeoRectangle)
QLibraryPrivate::loadPlugin failed on 
".../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so"
 : "Cannot load library 
.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so:
 
(.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so:
 undefined symbol: _ZN7QGeoMap25fitViewportToGeoRectangleERK13QGeoRectangle)"
5 ,  "loader()->instance(idx) failed to return an instance. Set the environment 
variable QT_DEBUG_PLUGINS to see more details."
qrc:/qml/MapWidget.qml:24: Error: Cannot assign [undefined] to 
QDeclarativeGeoMapType*
QFactoryLoader::QFactoryLoader() checking directory path 
"/usr/lib/x86_64-linux-gnu/qt5/plugins/accessiblebridge" ...
QFactoryLoader::QFactoryLoader() checking directory path 
".../subsurface/build/accessiblebridge" ...


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: Mares Smart Dive Computer + Bluelink pro

[Cristian Ionescu-Idbohrn]

On Wed, 26 Sep 2018, Linus Torvalds wrote:
> 
> I think it will only speed things up (and make the BLE code work), but
> maybe I'm being overly optimistic.

Just noticed this blog post:

Reverse Engineering BLE Devices
https://sergioalberti.gitlab.io//gsoc/2018/09/24/reveng.html

may or may not help locate the oddities.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

seems appdata/subsurface.appdata.xml in the repo is edited during build

[Cristian Ionescu-Idbohrn]

$ git status
...
modified:   appdata/subsurface.appdata.

$ git diff
...
-
+


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: subsurface builds, but dumps core at startup :(

[Cristian Ionescu-Idbohrn]

On Tue, 11 Sep 2018, Dirk Hohndel wrote:
> 
> The packaging of private Qt headers changes - I believe they simply 
> got dropped (which is not entirely unreasonable given that they are, 
> you know, private). As a result in a completely clean build we are 
> now downloading them directly and a clean build on Debian and Ubuntu 
> /should/ work - I did one on 18.04 a few days ago and that went 
> smoothly.

Alright.  "a completely clean build" seems to work, but I wouldn't 
call that sanity, should I?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: subsurface builds, but dumps core at startup :(

[Cristian Ionescu-Idbohrn]

On Tue, 11 Sep 2018, Linus Torvalds wrote:
> On Tue, Sep 11, 2018 at 7:32 AM Cristian Ionescu-Idbohrn
>  wrote:
> >
> > >
> > > did you switch the qt version but did not rebuild the googlemaps
> > > qtlocation plugin? it depends on internal api so a rebuild is needed...
> >
> > Of course.
> 
> I think we have some stale information in our build trees that does't
> have the right dependencies for qt changes.
> 
> I had a similar unexplained SIGSEGV that went away when building the
> whole tree from scratch.

I did so too (I think).  Still no joy :(


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: subsurface builds, but dumps core at startup :(

[Cristian Ionescu-Idbohrn]

On Tue, 11 Sep 2018, Martin Gysel wrote:
> Am 11.09.18 um 18:21 schrieb Cristian Ionescu-Idbohrn:
> > This is on debian unstable, subsurface master.  Seen this happening 
> > since a few days back.
> > 
> > The kernel reports:
> > 
> > Sep 11 17:36:33 host kernel: [11915775.481135] subsurface[24251]: segfault 
> > at 7fea4684ec60 ip 7fea4684ec60 sp 7ffd83e6ad38 error 15
> > Sep 11 17:36:33 host kernel: [11915775.481145] 
> > inlibqtgeoservices_googlemaps.so[7fea4684b000+8000]
> 
> did you switch the qt version but did not rebuild the googlemaps
> qtlocation plugin? it depends on internal api so a rebuild is needed...

Of course.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

subsurface builds, but dumps core at startup :(

[Cristian Ionescu-Idbohrn]

This is on debian unstable, subsurface master.  Seen this happening 
since a few days back.

The kernel reports:

Sep 11 17:36:33 host kernel: [11915775.481135] subsurface[24251]: segfault at 
7fea4684ec60 ip 7fea4684ec60 sp 7ffd83e6ad38 error 15
Sep 11 17:36:33 host kernel: [11915775.481145] 
inlibqtgeoservices_googlemaps.so[7fea4684b000+8000]

This is what happens:

$ ./subsurface/build/subsurface -vvv
QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-user'
QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-user'
Subsurface v4.8.1-371-g619a52635a7f,
built with libdivecomputer v0.7.0-devel-Subsurface-NG 
(a9c582e26fe823bc7981ba5ee0510ad9d6f86017)
built with Qt Version 5.11.1, runtime from Qt Version 5.11.1
built with libgit2 0.26.0
"validateGL(): created OpenGLContext."
"validateGL(): obtained QOpenGLFunctions."
"validateGL(): detected OpenGL version 4.6."
Plugins Directory:  QDir( "./subsurface/build" , nameFilters = { "*" },  
QDir::SortFlags( Name | IgnoreCase ) , QDir::Filters( 
Dirs|Files|Drives|AllEntries ) )

$ gdb ./subsurface/build/subsurface core
GNU gdb (Debian 8.1-4+b1) 8.1
Copyright (C) 2018 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
.
Find the GDB manual and other documentation resources online at:
.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from ./subsurface/build/subsurface...done.
[New LWP 24251]
[New LWP 24253]
[New LWP 24256]
[New LWP 24254]
[New LWP 24252]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `./subsurface/build/subsurface -vvv'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x7fea4684ec60 in qt_meta_data_QGeoTiledMapGooglemaps ()
   from 
./install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so
[Current thread is 1 (Thread 0x7fea5c7b6a40 (LWP 24251))]
(gdb) bt full
#0  0x7fea4684ec60 in qt_meta_data_QGeoTiledMapGooglemaps ()
   from 
./install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so
No symbol table info available.
#1  0x7fea670575b4 in 
QDeclarativeGeoMapCopyrightNotice::setMapSource(QDeclarativeGeoMap*) () from 
/usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#2  0x7fea6703e1f3 in QDeclarativeGeoMap::mappingManagerInitialized() ()
   from /usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#3  0x7fea6703edc1 in QDeclarativeGeoMap::pluginReady() ()
   from /usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#4  0x7fea6703efc6 in 
QDeclarativeGeoMap::setPlugin(QDeclarativeGeoServiceProvider*) () from 
/usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#5  0x7fea670a29d1 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#6  0x7fea670a3673 in QDeclarativeGeoMap::qt_metacall(QMetaObject::Call, 
int, void**) () from /usr/lib/x86_64-linux-gnu/libQt5Location.so.5
No symbol table info available.
#7  0x7fea65ccdb27 in QQmlVMEMetaObject::metaCall(QObject*, 
QMetaObject::Call, int, void**) () from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#8  0x7fea65ce4092 in ?? () from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#9  0x7fea65ce132a in QQmlPropertyPrivate::write(QObject*, QQmlPropertyData 
const&, QVariant const&, QQmlContextData*, QFlags)
() from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#10 0x7fea65c97f7d in 
QV4::QObjectWrapper::setProperty(QV4::ExecutionEngine*, QObject*, 
QQmlPropertyData*, QV4::Value const&) ()
   from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#11 0x7fea65c9893e in 
QV4::QObjectWrapper::setQmlProperty(QV4::ExecutionEngine*, QQmlContextData*, 
QObject*, QV4::String*, QV4::QObjectWrapper::RevisionMode, QV4::Value const&) 
() from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#12 0x7fea65c98a34 in QV4::QObjectWrapper::put(QV4::Managed*, QV4::String*, 
QV4::Value const&) () from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info available.
#13 0x7fea65cab9dc in 
QV4::Runtime::method_storeProperty(QV4::ExecutionEngine*, QV4::Value const&, 
int, QV4::Value const&) ()
   from /usr/lib/x86_64-linux-gnu/libQt5Qml.so.5
No symbol table info 

Re: recent build; symbol lookup error in libqtgeoservices_googlemaps.so

[Cristian Ionescu-Idbohrn]

On Fri, 11 May 2018, Dirk Hohndel wrote:
> > On May 11, 2018, at 5:07 AM, Lubomir I. Ivanov  wrote:
> > 
> > CMake probably struggles with the detection of such changes.
> > safest bet is to clean the build dir manually.
> 
> We don't use cmake rules to trigger re-builds of dependencies.

FWIW, there may be ways to do that.  Looking at what's going on in 
scripts/build.sh:

1. keep the clone:

git clone --branch v$QT_VERSION git://code.qt.io/qt/qtlocation.git 
--depth=1 ./qtlocation_git

   between builds, instead of wiping it out

2. detect the current QT_VERSION with:

$ qmake -query QT_VERSION

3. compare it with what `git describe' returns in latest clone, to 
   detect changes and trigger rebuild of relevant parts

> The idea is that "unless you know what you are doing, please
> start from a clean build using the build script".

That's an unattractive idea to me.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH] Simplier way to get QT_VERSION

[Cristian Ionescu-Idbohrn]

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbo...@axis.com>
---
 scripts/build.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/build.sh b/scripts/build.sh
index f0f3f2f61..c8ca3bd0a 100755
--- a/scripts/build.sh
+++ b/scripts/build.sh
@@ -152,7 +152,7 @@ fi
 # QtPositioning headers aren't bundled. Download them if necessary.
 if [ $PLATFORM = Linux ] ; then
QT_HEADERS_PATH=`$QMAKE -query QT_INSTALL_HEADERS`
-   QT_VERSION=`$QMAKE -v | grep "Qt" | cut -d" " -f4`
+   QT_VERSION=`$QMAKE -query QT_VERSION`
 
if [ ! -d "$QT_HEADERS_PATH/QtLocation/$QT_VERSION/QtLocation/private" 
] &&
[ ! -d $INSTALL_ROOT/include/QtLocation/private ] ; then
-- 
2.17.0
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: recent build; symbol lookup error in libqtgeoservices_googlemaps.so

[Cristian Ionescu-Idbohrn]

On Fri, 11 May 2018, Lubomir I. Ivanov wrote:
> On 11 May 2018 at 17:30, Dirk Hohndel  wrote:
> >
> > We don't use cmake rules to trigger re-builds of dependencies.
> > The idea is that "unless you know what you are doing, please
> > start from a clean build using the build script".
> 
> i think the script to download the private headers for QtLocation
> doesn't handle sudden Qt version changes in regard to the already
> present objects code in the build folder.
> that aside, it's possible to get other build problems after bumping Qt
> so the clean build is really the way to go.

Please, tell me what a "clean build" is.
If anything fails, "wipe out the build dir"?  Which one?  There's at 
least:

./googlemaps/
./install-root/
./subsurface/build/

seen from the:

./subsurface/scripts/build.sh

perspective.  Is that it?  Are there more or less?
Is that documented and I just missed it?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: recent build; symbol lookup error in libqtgeoservices_googlemaps.so

[Cristian Ionescu-Idbohrn]

On Fri, 11 May 2018, Lubomir I. Ivanov wrote:
> 
> Christian, did you try deleting the build folder (including the 
> gmaps plugin) completely if you recently updated to a new Qt before 
> rebuilding?

Yes, that did the trick, but shouldn't that be detected and rebuilt?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

recent build; symbol lookup error in libqtgeoservices_googlemaps.so

[Cristian Ionescu-Idbohrn]

Any ideas?

$ ./subsurface/build/subsurface -vvv
Subsurface v4.7.8-124-g2c29cc96a34a,
built with libdivecomputer v0.7.0-devel-Subsurface-NG 
(e97a47cca55973199715df0f818b4955e60d3a31)
built with Qt Version 5.10.1, runtime from Qt Version 5.10.1
built with libgit2 0.26.0
"validateGL(): created OpenGLContext."
"validateGL(): obtained QOpenGLFunctions."
"validateGL(): detected OpenGL version 3.0."
Plugins Directory:  QDir( "...E/subsurface/build" , 
nameFilters = { "*" },  QDir::SortFlags( Name | IgnoreCase ) , 
QDir::Filters( Dirs|Files|Drives|AllEntries ) )
./subsurface/build/subsurface: symbol lookup error: 
.../install-root/usr/lib/x86_64-linux-gnu/qt5/plugins/geoservices/libqtgeoservices_googlemaps.so:
 
undefined symbol: 
_ZN11QGeoMapTypeC1ENS_8MapStyleERK7QStringS3_bbi10QByteArray


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

modified: libdivecomputer (new commits)

[Cristian Ionescu-Idbohrn]

This is what I did, yesterday:

$ git fetch
remote: Counting objects: 143, done.
remote: Compressing objects: 100% (18/18), done.
remote: Total 143 (delta 103), reused 118 (delta 103), pack-reused 22
Receiving objects: 100% (143/143), 49.36 KiB | 16.45 MiB/s, done.
Resolving deltas: 100% (103/103), completed with 34 local objects.
From git://github.com/Subsurface-divelog/subsurface
   47bf5cf5e..04626b089  master-> origin/master
 + 4bc50a2b9...22bfeb999 coverity_scan -> origin/coverity_scan  (forced update)
Fetching submodule libdivecomputer
From git://github.com/Subsurface-divelog/libdc
   c354ae0..70ad30b  Subsurface-branch -> origin/Subsurface-branch
 + c2d9108...b48a4a5 coverityCleanup   -> origin/coverityCleanup  (forced 
update)

$ git pull
Updating 47bf5cf5e..04626b089
Fast-forward
 CHANGELOG.md   |   1 +
 core/dive.h|   4 +-
 core/divelist.c|   8 +-
 core/divelist.h|   2 +-
 core/divesite.c|   4 +-
 core/divesitehelpers.cpp   |   2 -
 core/libdivecomputer.c |  12 +-
 core/load-git.c|   2 +-
 core/parse-xml.c   |   9 +-
 core/parse.c   |   2 +-
 core/parse.h   |   5 +-
 core/pref.h|   1 -
 core/save-git.c|   9 +-
 core/subsurface-qt/SettingsObjectWrapper.cpp   |  17 ---
 core/subsurface-qt/SettingsObjectWrapper.h |   4 -
 core/subsurfacestartup.c   |   1 -
 core/uemis-downloader.c|   2 -
 desktop-widgets/mainwindow.cpp | 131 +
 desktop-widgets/mainwindow.h   |   5 +-
 desktop-widgets/mainwindow.ui  |   9 +-
 .../preferences/preferences_network.cpp|   2 -
 desktop-widgets/preferences/preferences_network.ui |   7 --
 desktop-widgets/subsurfacewebservices.cpp  | 124 +--
 libdivecomputer|   2 +-
 mobile-widgets/qmlmanager.cpp  |   1 -
 qt-models/divelistmodel.cpp|   1 +
 tests/testgitstorage.cpp   |   1 -
 tests/testpreferences.cpp  |   4 -
 28 files changed, 134 insertions(+), 238 deletions(-)

and ended up with:

$ git status
On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
  (use "git add ..." to update what will be committed)
  (use "git checkout -- ..." to discard changes in working 
directory)

modified:   libdivecomputer (new commits)

no changes added to commit (use "git add" and/or "git commit -a")

which doesn't look wright to me.  Looking deeper into it, I see this:

$ git diff | cat
diff --git a/libdivecomputer b/libdivecomputer
index 70ad30bed..8ae735a4d 16
--- a/libdivecomputer
+++ b/libdivecomputer
@@ -1 +1 @@
-Subproject commit 70ad30bed30ce1b0b10eced0ffa7a115fed51d80
+Subproject commit 8ae735a4d70307ebe2a42d315697f02ce71dbe88

How am I (a git novice) expected to react?

Repeating the procedure today again, I pull more "stuff".  Still, the 
same ...

$ git diff | cat
diff --git a/libdivecomputer b/libdivecomputer
index 70ad30bed..8ae735a4d 16
--- a/libdivecomputer
+++ b/libdivecomputer
@@ -1 +1 @@
-Subproject commit 70ad30bed30ce1b0b10eced0ffa7a115fed51d80
+Subproject commit 8ae735a4d70307ebe2a42d315697f02ce71dbe88

What am I missing?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: 4.7.6

[Cristian Ionescu-Idbohrn]

On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> > On Dec 24, 2017, at 12:54 PM, Cristian Ionescu-Idbohrn 
> > <cristian.ionescu-idbo...@axis.com> wrote:
> > 
> > On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> >> 
> >> cd ~/src/subsurface
> >> git submodule update
> > 
> >> cd subsurface
> > [master=]> git submodule update
> > fatal: destination path 
> > '.../subsurface/libdivecomputer' already exists and 
> > is not an empty directory.
> 
> And there is your problem - I'm guessing that you have a version of 
> libdivecomputer checked out in subsurface/libdivecomputer

I guess subsurface/libdivecomputer came with:

commit 10237a19b7b65802a0ceb2f8e2192703ad2fcf97
Author: Dirk Hohndel <d...@hohndel.org>
Date:   Sun Dec 3 11:39:58 2017 -0800

libdc: add our libdivecomputer branch as git submodule

But that was previously cloned under ~/src, if I remember correctly.  
And what I missed was doing:

[master=]> git submodule update

at some point after pulling from 
git://github.com/Subsurface-divelog/subsurface

> move that directory elsewhere (or remove completely) and try again.

Yes, I ended up removing both ~/src/libdivecomputer and 
~/src/subsurface/libdivecomputer,

> cd ~/src/subsurface
[master=]> git pull
[master=]> git submodule update

and now it builds.

I imagine this must have happen for many other people who keep their 
cloned trees around for a long time.  If my interpretation is correct, 
would it be a good idea to mention something about it in the INSTALL 
file?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: 4.7.6

[Cristian Ionescu-Idbohrn]

On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> > On Dec 24, 2017, at 12:47 PM, Jan Mulder <jlmul...@xs4all.nl> wrote:
> > 
> > On 24-12-17 21:39, Cristian Ionescu-Idbohrn wrote:
> > 
> >> nothing added to commit but untracked files present (use "git add" to
> >> track)
> >> Still, no joy :(
> > 
> > The current head is:
> > 
> > git describe
> > v4.7.4-47-g8ae735a
> 
> We are both wrong. You are missing the v4.7.5 tag - but I gave the 
> commit count in Subsurface, not in libdivecomputer.
> Officially the latest is
> 
> v4.7.5-18-g8ae735a4d703
> 
> :-)
> 
> > Notice that you need to look at /subsurface/libdivecomputer 
> > (its a git submodule now).
> 
> Correct

[master=]> pwd
.../subsurface/libdivecomputer

[master=]> git describe 
v4.7.5-96-g36c8a7e4c


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: 4.7.6

[Cristian Ionescu-Idbohrn]

On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> 
> cd ~/src/subsurface
> git submodule update

> cd subsurface
[master=]> git submodule update
fatal: destination path 
'.../subsurface/libdivecomputer' already exists and 
is not an empty directory.
fatal: clone of 'git://github.com/Subsurface-divelog/libdc.git' into 
submodule path '.../subsurface/libdivecomputer' 
failed
Failed to clone 'libdivecomputer'. Retry scheduled
fatal: destination path 
'.../subsurface/libdivecomputer' already exists and 
is not an empty directory.
fatal: clone of 'git://github.com/Subsurface-divelog/libdc.git' into 
submodule path '.../subsurface/libdivecomputer' 
failed
Failed to clone 'libdivecomputer' a second time, aborting

Hmm...

> should get you to a working version. Assuming you are on something 
> resembling current master :-)

Doesn't seem so :(


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: 4.7.6

[Cristian Ionescu-Idbohrn]

On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> > On Dec 24, 2017, at 12:11 PM, Cristian Ionescu-Idbohrn 
> > <cristian.ionescu-idbo...@axis.com> wrote:
> > 
> > On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> >> 
> >> There are a couple of small issues with 4.7.5 (QtQuick version 
> >> causes issues on old Linux distros with Qt 5.5, the annoying warning 
> >> about the "m'" in the git storage) and of course we have a couple 
> >> cool new features like the picture zoom levels and semi-closed 
> >> rebreather support for the OSTC...
> >> 
> >> Is there anything I should wait for? Anything that we broke since 
> >> 4.7.5 and haven't fixed?
> > 
> > Didn't check what goes in there, but master fails to build for be, 
> > since a few days back:
> > 
> > [ 46%] Built target subsurface_corelib_autogen
> > [ 47%] Building C object 
> > core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o
> > .../subsurface/core/libdivecomputer.c: In function 
> > 'libdc_header_parser':
> > .../subsurface/core/libdivecomputer.c:738:8: error: 
> > 'DC_DIVEMODE_CCR' undeclared (first use in this function); did you 
> > mean 'DC_DIVEMODE_CC'?
> >   case DC_DIVEMODE_CCR:  /* Closed circuit rebreather*/
> >^~~
> >DC_DIVEMODE_CC
> > .../subsurface/core/libdivecomputer.c:738:8: note: 
> > each undeclared identifier is reported only once for each function it 
> > appears in
> > .../subsurface/core/libdivecomputer.c:741:8: error: 
> > 'DC_DIVEMODE_SCR' undeclared (first use in this function); did you 
> > mean 'DC_DIVEMODE_CCR'?
> >   case DC_DIVEMODE_SCR:  /* Semi-closed circuit rebreather */
> >^~~
> >DC_DIVEMODE_CCR
> > .../subsurface/core/libdivecomputer.c:730:3: warning: 
> > enumeration value 'DC_DIVEMODE_CC' not handled in switch [-Wswitch]
> >   switch(divemode) {
> >   ^~
> > core/CMakeFiles/subsurface_corelib.dir/build.make:374: recipe for 
> > target 'core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o' 
> > failed
> > make[2]: *** 
> > [core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o] Error 1
> > CMakeFiles/Makefile2:473: recipe for target 
> > 'core/CMakeFiles/subsurface_corelib.dir/all' failed
> > make[1]: *** [core/CMakeFiles/subsurface_corelib.dir/all] Error 2
> > Makefile:129: recipe for target 'all' failed
> > 
> > and I don't know what I'm missing :(
> 
> 
> You're not building against the current libdivecomputer.

I'd have expected build.sh would pull current libdivecomputer, which 
it seems to do:

[Subsurface-branch %=]> git fetch
[Subsurface-branch %=]> git pull 
Already up to date.
[Subsurface-branch %=]> git describe 
v4.6.4-131-gc0f025b
[Subsurface-branch %=]> git status 
On branch Subsurface-branch
Your branch is up to date with 'origin/Subsurface-branch'.

Untracked files:
  (use "git add ..." to include in what will be committed)

build/

nothing added to commit but untracked files present (use "git add" to 
track)

Still, no joy :(


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: 4.7.6

[Cristian Ionescu-Idbohrn]

On Sun, 24 Dec 2017, Dirk Hohndel wrote:
> 
> There are a couple of small issues with 4.7.5 (QtQuick version 
> causes issues on old Linux distros with Qt 5.5, the annoying warning 
> about the "m'" in the git storage) and of course we have a couple 
> cool new features like the picture zoom levels and semi-closed 
> rebreather support for the OSTC...
> 
> Is there anything I should wait for? Anything that we broke since 
> 4.7.5 and haven't fixed?

Didn't check what goes in there, but master fails to build for be, 
since a few days back:

[ 46%] Built target subsurface_corelib_autogen
[ 47%] Building C object 
core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o
.../subsurface/core/libdivecomputer.c: In function 
'libdc_header_parser':
.../subsurface/core/libdivecomputer.c:738:8: error: 
'DC_DIVEMODE_CCR' undeclared (first use in this function); did you 
mean 'DC_DIVEMODE_CC'?
   case DC_DIVEMODE_CCR:  /* Closed circuit rebreather*/
^~~
DC_DIVEMODE_CC
.../subsurface/core/libdivecomputer.c:738:8: note: 
each undeclared identifier is reported only once for each function it 
appears in
.../subsurface/core/libdivecomputer.c:741:8: error: 
'DC_DIVEMODE_SCR' undeclared (first use in this function); did you 
mean 'DC_DIVEMODE_CCR'?
   case DC_DIVEMODE_SCR:  /* Semi-closed circuit rebreather */
^~~
DC_DIVEMODE_CCR
.../subsurface/core/libdivecomputer.c:730:3: warning: 
enumeration value 'DC_DIVEMODE_CC' not handled in switch [-Wswitch]
   switch(divemode) {
   ^~
core/CMakeFiles/subsurface_corelib.dir/build.make:374: recipe for 
target 'core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o' 
failed
make[2]: *** 
[core/CMakeFiles/subsurface_corelib.dir/libdivecomputer.c.o] Error 1
CMakeFiles/Makefile2:473: recipe for target 
'core/CMakeFiles/subsurface_corelib.dir/all' failed
make[1]: *** [core/CMakeFiles/subsurface_corelib.dir/all] Error 2
Makefile:129: recipe for target 'all' failed

and I don't know what I'm missing :(


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: XML and pdf user-manual.text...

[Cristian Ionescu-Idbohrn]

On Sat, 4 Nov 2017, Miika Turkia wrote:
> 
> The problem seems to be the ordering of emphasis and literal tags (+
> and ' modifiers in the markdown). I just sent a pull request the
> re-arranges them to produce working XML on my system.

Sounds good.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: XML and pdf user-manual.text...

[Cristian Ionescu-Idbohrn]

On Sat, 21 Dec 2013, Willem Ferguson wrote:
> 
> One of the things I have been wanting to do is to get the xml and pdf 
> options working. Unless anyone else has a ready answer, I will look into 
> this during the next 2 days.

I'm still having the same/similar problem when trying to build the 
documentation :(  This is an old thread, started here:


http://lists.subsurface-divelog.org/pipermail/subsurface/2013-December/009399.html

This is what I do:

$ make -C Documentation user-manual.text
make: Entering directory '.../Documentation'
a2x -f text user-manual.txt
a2x: ERROR: "xmllint" --nonet --noout --valid 
".../Documentation/user-manual.xml" 
returned non-zero exit status 4
Makefile:27: recipe for target 'user-manual.text' failed

And the reason is:

$ xmllint --nonet --noout --valid ./user-manual.xml
./user-manual.xml:5792: element literal: validity error : Element 
emphasis is not declared in literal list of possible children
/emphasis>, RUNNING AND 
AUTH

   ^
./user-manual.xml:5792: element literal: validity error : Element 
emphasis is not declared in literal list of possible children
/emphasis>, RUNNING AND 
AUTH

   ^
./user-manual.xml:5792: element literal: validity error : Element 
emphasis is not declared in literal list of possible children
/emphasis>, RUNNING AND 
AUTH

   ^


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: segmentation fault

[Cristian Ionescu-Idbohrn]

On Sun, 1 Oct 2017, Dirk Hohndel wrote:
> 
> Can you send a patch for the INSTALL for to capture the right 
> packages that need to be installed?

I don't need to.  The packages are already listed in INSTALL:

101:libcrypto++-dev libssl-dev qml-module-qtpositioning 
qml-module-qtlocation
112:libcrypto++-dev libssl-dev qml-module-qtpositioning 
qml-module-qtlocation

for both debian and ubuntu.  In my case (debian unstable) those 
packages were removed when I did a major qt5 upgrade.

In any case, the point I'm trying to make is subsurface should _not_ 
segfault because of that.  That case should be properly error handled, 
shouldn't it?

The two libraries:

/usr/lib/x86_64-linux-gnu/qt5/qml/QtLocation/libdeclarative_location.so
/usr/lib/x86_64-linux-gnu/qt5/qml/QtPositioning/libdeclarative_positioning.so

are probably loaded at runtime (maybe with dlopen?).  The code that 
does that appears not to do error handling.

It also appears the cmake rules don't check either.  Shouldn't they?

> On October 1, 2017 2:00:45 AM PDT, Cristian Ionescu-Idbohrn 
> <cristian.ionescu-idbo...@axis.com> wrote:
> >Latest master (v4.6.4-871-g571918fe) build.
> >
> >qrc:/MapWidget.qml:4:1: module "QtPositioning" is not installed
> >qrc:/MapWidget.qml:3:1: module "QtLocation" is not installed
> >qrc:/MapWidget.qml:4:1: module "QtPositioning" is not installed
> >qrc:/MapWidget.qml:3:1: module "QtLocation" is not installed
> >
> >Installing packages qml-module-qtlocation and qml-module-qtpositioning 
> >solves that problem.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

segmentation fault

[Cristian Ionescu-Idbohrn]

Latest master (v4.6.4-871-g571918fe) build.

qrc:/MapWidget.qml:4:1: module "QtPositioning" is not installed
qrc:/MapWidget.qml:3:1: module "QtLocation" is not installed
qrc:/MapWidget.qml:4:1: module "QtPositioning" is not installed
qrc:/MapWidget.qml:3:1: module "QtLocation" is not installed

Installing packages qml-module-qtlocation and qml-module-qtpositioning 
solves that problem.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: cloud.subsurface-divelog.org auth error

[Cristian Ionescu-Idbohrn]

On Sun, 27 Aug 2017, Dirk Hohndel wrote:
> > On Aug 27, 2017, at 10:15 AM, Yury Akudovich  wrote:
> > I've rebuilt the library and Subsurface with https support and now it works!
> > Unfortunately I can't install your deb packages because of dependency
> > on libssl1.0.0, but in my Debian buster(testing) is only libssl1.1
> > available.
> 
> Hmm - I have a hard time matching Ubuntu and Debian versions. Even the 
> Artful repo isn't new enough for Buster?

Thing is, you can have both libssl1.0.2 _and_ libssl1.1 installed on 
debian buster/sid.  Your choice is to install libssl1.0-dev (instead 
of libssl-dev), should you choose to use libssl1.0-dev (headers, lib) 
for development.  In my case, this is what I currently have installed:

libssl1.0-dev:amd64 install
^
libssl1.0.0:amd64   install
libssl1.0.2:amd64   install
libssl1.1:amd64 install
libssl1.1:i386  install

on debian sid.  There are still quite a few packages that lack 
libssl1.1 support in the distro.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH 3/3] Simplify some tests

[Cristian Ionescu-Idbohrn]

Because `[ "$foo" != "" ] is equivalent to `[ "$foo" ]'
in all POSIX shells.

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbo...@axis.com>
---
 packaging/android/android-build-wrapper.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/packaging/android/android-build-wrapper.sh 
b/packaging/android/android-build-wrapper.sh
index cd96f965..0d1e6913 100644
--- a/packaging/android/android-build-wrapper.sh
+++ b/packaging/android/android-build-wrapper.sh
@@ -49,7 +49,7 @@ for i in git cmake autoconf libtool java ant wget unzip; do
MISSING="${MISSING}${i} "
fi
 done
-if [ "$MISSING" != "" ] ; then
+if [ "$MISSING" ] ; then
echo "The following packages are missing: $MISSING"
echo "Please install via your package manager."
exit 1
@@ -132,7 +132,7 @@ fi
 rm -f ./subsurface-mobile-build-arm/bin/QtApp-debug.apk
 rm -d ./subsurface-mobile-build-arm/AndroidManifest.xml
 rm -d ./subsurface-mobile-build-arm/bin/AndroidManifest.xml
-if [ "$USE_X" != "" ] ; then
+if [ "$USE_X" ] ; then
bash "$USE_X" subsurface/packaging/android/build.sh "$RELEASE" -buildnr 
"$BUILDNR" arm "$@"
 else
bash subsurface/packaging/android/build.sh "$RELEASE" -buildnr 
"$BUILDNR" arm "$@"
-- 
2.11.0

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH 0/3] Use `command' (shell builtin and POSIX) and cleanup stuff

[Cristian Ionescu-Idbohrn]

Cristian Ionescu-Idbohrn (3):
  Use `command' (shell builtin and POSIX) instead
  Remove useless quotes
  Simplify some tests

 packaging/android/android-build-wrapper.sh | 22 +++---
 1 file changed, 11 insertions(+), 11 deletions(-)

-- 
2.11.0

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH 1/3] Use `command' (shell builtin and POSIX) instead

[Cristian Ionescu-Idbohrn]

The `which' command is a fork and possible not standard in various
distributions, or builtin in certain (odd)? shells, like `zsh'.

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbo...@axis.com>
---
 packaging/android/android-build-wrapper.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/packaging/android/android-build-wrapper.sh 
b/packaging/android/android-build-wrapper.sh
index cb98fc9e..b111e293 100644
--- a/packaging/android/android-build-wrapper.sh
+++ b/packaging/android/android-build-wrapper.sh
@@ -40,7 +40,7 @@ fi
 # make sure we have the required commands installed
 MISSING=""
 for i in git cmake autoconf libtool java ant wget unzip; do
-   which $i > /dev/null 2>&1 ||
+   command -v $i >/dev/null ||
if [ "$i" = "libtool" ] ; then
MISSING="${MISSING}libtool-bin "
elif [ "$i" = "java" ] ; then
-- 
2.11.0

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH 2/3] Remove useless quotes

[Cristian Ionescu-Idbohrn]

Don't quote if you don't have to.  Spend those cpu cycles on doing
something more useful, instead.

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbo...@axis.com>
---
 packaging/android/android-build-wrapper.sh | 16 
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/packaging/android/android-build-wrapper.sh 
b/packaging/android/android-build-wrapper.sh
index b111e293..cd96f965 100644
--- a/packaging/android/android-build-wrapper.sh
+++ b/packaging/android/android-build-wrapper.sh
@@ -10,7 +10,7 @@
 
 exec 1> >(tee ./build.log) 2>&1
 
-USE_X=$(case "$-" in *x*) echo "-x" ;; esac)
+USE_X=$(case $- in *x*) echo "-x" ;; esac)
 
 # these are the current versions for Qt, Android SDK & NDK:
 
@@ -24,11 +24,11 @@ ANDROID_SDK=android-sdk-linux
 
 PLATFORM=$(uname)
 
-pushd "$(dirname "$0")/../../"
+pushd $(dirname "$0")/../../
 export SUBSURFACE_SOURCE=$PWD
 popd
 
-if [ "$PLATFORM" = "Linux" ] ; then
+if [ "$PLATFORM" = Linux ] ; then
QT_BINARIES=qt-opensource-linux-x64-android-${LATEST_QT}.run
NDK_BINARIES=${ANDROID_NDK}-linux-x86_64.zip
SDK_TOOLS=tools_${SDK_VERSION}-linux.zip
@@ -38,12 +38,12 @@ else
 fi
 
 # make sure we have the required commands installed
-MISSING=""
+MISSING=
 for i in git cmake autoconf libtool java ant wget unzip; do
command -v $i >/dev/null ||
-   if [ "$i" = "libtool" ] ; then
+   if [ $i = libtool ] ; then
MISSING="${MISSING}libtool-bin "
-   elif [ "$i" = "java" ] ; then
+   elif [ $i = java ] ; then
MISSING="${MISSING}openjdk-8-jdk "
else
MISSING="${MISSING}${i} "
@@ -112,7 +112,7 @@ popd
 
 
 # and now we need a monotonic build number...
-if [ ! -f "./buildnr.dat" ] ; then
+if [ ! -f ./buildnr.dat ] ; then
BUILDNR=0
 else
BUILDNR=$(cat ./buildnr.dat)
@@ -122,7 +122,7 @@ echo "${BUILDNR}" > ./buildnr.dat
 
 echo "Building Subsurface-mobile ${VERSION} for Android, build nr ${BUILDNR} 
as Subsurface-mobile-$VERSION-${NAME}arm.apk"
 
-if [ "$1" = "release" ] || [ "$1" = "Release" ] || [ "$1" = "debug" ] || [ 
"$1" = "Debug" ] ; then
+if [ "$1" = release ] || [ "$1" = Release ] || [ "$1" = debug ] || [ "$1" = 
Debug ] ; then
RELEASE=$1
shift
 else
-- 
2.11.0

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH 07/12] android-wrapper: Auto accept android sdk license

[Cristian Ionescu-Idbohrn]

On Mon, 6 Feb 2017, Anton Lundin wrote:
> On 06 February, 2017 - Cristian Ionescu-Idbohrn wrote:
> > On Mon, 6 Feb 2017, Anton Lundin wrote:
> > > 
> > > Signed-off-by: Anton Lundin <gla...@acc.umu.se>
> > > ---
> > >  packaging/android/android-build-wrapper.sh | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/packaging/android/android-build-wrapper.sh 
> > > b/packaging/android/android-build-wrapper.sh
> > > index c8dd1e1..6242809 100644
> > > --- a/packaging/android/android-build-wrapper.sh
> > > +++ b/packaging/android/android-build-wrapper.sh
> > > @@ -82,7 +82,7 @@ if [ ! -d $ANDROID_SDK ] ; then
> > >   mkdir $ANDROID_SDK
> > >   pushd $ANDROID_SDK
> > >   unzip -q ../$SDK_TOOLS
> > > - bash tools/android update sdk --no-ui -a -t 1,2,3,33
> > > + ( sleep 5 && while true ; do sleep 1; echo y; done ) | bash 
> > > tools/android update sdk --no-ui -a -t 1,2,3,33
> > 
> > Can it be so that the simplier:
> > 
> > yes y | bash tools/android update sdk --no-ui -a -t 1,2,3,33
> > 
> > is all you need?
> 
> Nope. sdk updater does read(0, dest, 1024) or something, so if yes have
> pushed too many y\n into the buffer, it will fail when doing
> strcmp(dest, "y\n") or something alike.
> 
> Thats why odd one here.

I guess you've tested `yes', that's why you know ;)


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH 03/12] android-wrapper: Add wget as prerequisite

[Cristian Ionescu-Idbohrn]

On Mon, 6 Feb 2017, Anton Lundin wrote:
> 
> Signed-off-by: Anton Lundin 
> ---
>  packaging/android/android-build-wrapper.sh | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/packaging/android/android-build-wrapper.sh 
> b/packaging/android/android-build-wrapper.sh
> index c1afcfb..03e4d5c 100644
> --- a/packaging/android/android-build-wrapper.sh
> +++ b/packaging/android/android-build-wrapper.sh
> @@ -35,7 +35,7 @@ fi
>  
>  # make sure we have the required commands installed
>  MISSING=""
> -for i in git cmake autoconf libtool java ant; do
> +for i in git cmake autoconf libtool java ant wget; do
>   which $i > /dev/null 2>&1 ||

Much better to use `command' (shell builtin and available in all POSIX 
complient shells), instead of `which' (fork and possible not available 
on all distributions):

command -v $i >/dev/null $$

>   if [ "$i" = "libtool" ] ; then
 ^  ^   ^   ^
No quotes needed.

>   MISSING="${MISSING}libtool-bin "


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: INSTALL recommends package libqt5declarative5...

[Cristian Ionescu-Idbohrn]

On Sun, 5 Feb 2017, Sebastian Kügler wrote:
> On zondag 5 februari 2017 15:32:47 CET Cristian Ionescu-Idbohrn wrote:
> > ... but I can't find anything by that name, neither in debian strech
> > (next stable) nor sid (unstable) :(  Does anyone know what that is?  A
> > fuzzy search brings this up:
> > 
> > libkdeclarative5 - declarative library for plasma
> > libqt4-declarative - Qt 4 Declarative module
> > libqt4-declarative-folderlistmodel - Qt 4 folderlistmodel QML plugin
> > libqt4-declarative-gestures - Qt 4 gestures QML plugin
> > libqt4-declarative-particles - Qt 4 particles QML plugin
> > libqt4-declarative-shaders - Qt 4 shaders QML plugin
> > libsmokeqtdeclarative4-3 - Qt Declarative SMOKE library
> 
> Probably libqt5quick5 and libqt5qml5. In Qt5 (as opposed to Qt 4), 
> QtDeclarative has been split up in QtQuick and QtQml.

So, for debian, libqt5declarative5 should be simply removed from the 
list.  The list already includes libqt5quick5, and libqt5quick5 has a 
dependency to libqt5qml5:

Depends: libc6 (>= 2.14),
 libqt5core5a (>= 5.7.0),
 libqt5gui5 (>= 5.7.0),
 libqt5network5 (>= 5.0.2),
 libqt5qml5 (>= 5.7.1),
 ^
 libstdc++6 (>= 5),
 qtbase-abi-5-7-1,
 qtdeclarative-abi-5-7-0


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH] Detect location of ldconfig before use

[Cristian Ionescu-Idbohrn]

On Sun, 5 Feb 2017, Robert Helling wrote:
> > On 05 Feb 2017, at 15:21, Cristian Ionescu-Idbohrn 
> > <cristian.ionescu-idbo...@axis.com> wrote:
> > 
> > I have an idea ;)
> > 
> > Why not error handle those the code that produces those "warnings" (as
> > you call them) on that particular architecture, instead of
> > ignoring/hiding the errors (exit status failure) for everyone?  That
> > kind of thing can bite you later in various ways, leading to bug
> > reports saying "I tried that but it doesn't work :/  The build script
> > exits with success, but...  Could you fix it for me?  I can't tell you
> > where to start looking :(".
> > 
> > 'errexit' was not invented to punish people, but to help coding robust
> > scripts.  In my experience, error handling is one of the most
> > neglected areas.
> 
> ok, took up the challenge. My memory was wrong. It wasn’t Qt. It was 
> the construct
> 
> echo $NAME | grep / > /dev/null 2>&1
> if [ $? -eq 1 ] ; then
> install_name_tool -id "$INSTALL_ROOT/lib/$NAME" 
> "$INSTALL_ROOT/lib/$NAME"
> fi
> 
> (this appears twice, you find it by searching for $?). Problem is, 
> when I run it, grep returns non-zero and this is then where the 
> script silently stops.

So, I read that as:

* if there is a '/' in $NAME, do nothing
* otherwise install $NAME

> Fixing this is beyond my bash abilities. If you want -e, please 
> rewrite these two occurrences so that the script does not silently 
> fail.

To even avoid the `grep' fork and the pipe, I would write that as:

case $NAME in
*/*)
:
;;
*)
install_name_tool -id "$INSTALL_ROOT/lib/$NAME" 
"$INSTALL_ROOT/lib/$NAME"
;;
esac

If you insist on using `grep':

echo $NAME | grep -q / ||
install_name_tool -id "$INSTALL_ROOT/lib/$NAME" 
"$INSTALL_ROOT/lib/$NAME"

and if it _must_ be done using an if/fi block:

if ! echo $NAME | grep -q /; then
install_name_tool -id "$INSTALL_ROOT/lib/$NAME" 
"$INSTALL_ROOT/lib/$NAME"
fi


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

INSTALL recommends package libqt5declarative5...

[Cristian Ionescu-Idbohrn]

... but I can't find anything by that name, neither in debian strech 
(next stable) nor sid (unstable) :(  Does anyone know what that is?  A 
fuzzy search brings this up:

libkdeclarative5 - declarative library for plasma
libqt4-declarative - Qt 4 Declarative module
libqt4-declarative-folderlistmodel - Qt 4 folderlistmodel QML plugin
libqt4-declarative-gestures - Qt 4 gestures QML plugin
libqt4-declarative-particles - Qt 4 particles QML plugin
libqt4-declarative-shaders - Qt 4 shaders QML plugin
libsmokeqtdeclarative4-3 - Qt Declarative SMOKE library


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH] Detect location of `ldconfig' before use

[Cristian Ionescu-Idbohrn]

On Sun, 5 Feb 2017, Robert C. Helling wrote:
> > Am 05.02.2017 um 12:53 schrieb Cristian Ionescu-Idbohrn 
> > <cristian.ionescu-idbo...@axis.com>:
> > 
> > That's the reason shell option '-e' (errexit) was introduced.
> 
> That option had been introduced before and then later been removed 
> again. The build on mac throws warnings about certain Qt things that 
> one should do (but which are not necessarily) and with -e that is 
> the end of the build.
> 
> So please don't. 

I have an idea ;)

Why not error handle those the code that produces those "warnings" (as 
you call them) on that particular architecture, instead of 
ignoring/hiding the errors (exit status failure) for everyone?  That 
kind of thing can bite you later in various ways, leading to bug 
reports saying "I tried that but it doesn't work :/  The build script 
exits with success, but...  Could you fix it for me?  I can't tell you 
where to start looking :(".

'errexit' was not invented to punish people, but to help coding robust 
scripts.  In my experience, error handling is one of the most 
neglected areas.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH] Detect location of `ldconfig' before use

[Cristian Ionescu-Idbohrn]

The unhandled error was:

./subsurface/scripts/build.sh: line 84: ldconfig: command not found

On my box, the path is /sbin/ldconfig, and /sbin is not in common users'
PATH.

The script just continued to execute after that error, which can result
in other odd bugs showing up later.  That's the reason shell option
'-e' (errexit) was introduced.

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbo...@axis.com>
---
 scripts/build.sh | 8 +++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/scripts/build.sh b/scripts/build.sh
index f60d7045..cf87f26b 100755
--- a/scripts/build.sh
+++ b/scripts/build.sh
@@ -16,6 +16,8 @@
 # of the current directory (except on Mac where the Subsurface.app ends up
 # in subsurface/build
 
+set -e
+
 # create a log file of the build
 exec 1> >(tee build.log) 2>&1
 
@@ -81,7 +83,11 @@ else
# check if we need to build libgit2 (and do so if necessary)
 
LIBGIT_ARGS=" -DLIBGIT2_DYNAMIC=ON "
-   LIBGIT=$(ldconfig -p | grep libgit2\\.so\\. | awk -F. '{ print $NF }')
+   _ldconfig=$(PATH=/usr/local/sbin:/usr/sbin:/sbin:$PATH command -v 
ldconfig) || {
+   echo "can't find ldconfig" >&2
+   exit 1
+   }
+   LIBGIT=$($_ldconfig -p | grep libgit2\\.so\\. | awk -F. '{ print $NF }')
 fi
 
 if [[ $PLATFORM = Darwin || "$LIBGIT" < "24" ]] ; then
-- 
2.11.0

___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: building master on a debian box

[Cristian Ionescu-Idbohrn]

On Sun, 19 Jun 2016, Miika Turkia wrote:
> On Sat, Jun 18, 2016 at 9:32 PM, Dirk Hohndel  wrote:
> > On Sat, Jun 18, 2016 at 02:19:50PM +0300, Miika Turkia wrote:
> >>
> >> The tests have been broken for a while, but Subsurface is already be
> >> compiled at that point. (set MAKE_TESTS to off on cmake configs will
> >> disable building of tests)
> >
> > Tests build fine here...
>
> Yeah, they work fine for me too. And now that I look at the original
> email, I suspect that installing libssl-dev fixed the building of
> tests for Cristian as well.

Seems to.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: building master on a debian box

[Cristian Ionescu-Idbohrn]

On Sat, 18 Jun 2016, Miika Turkia wrote:
>
> I just sent a patch that should take care of these in the INSTALL
> file. Thanks for letting us know what works on Debian.

Thanks.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

building master on a debian box

[Cristian Ionescu-Idbohrn]

Some observations:

CMake Error at /usr/lib/i386-linux-gnu/cmake/Qt5/Qt5Config.cmake:26 
(find_package):
  Could not find a package configuration file provided by "Qt5Positioning"
  with any of the following names:

Qt5PositioningConfig.cmake
qt5positioning-config.cmake

Qt5PositioningConfig.cmake is in package qtpositioning5-dev.
Similar stuff with Qt5BluetoothConfig.cmake (in package
qtconnectivity5-dev).

File INSTALL needs to be updated.

Though, still no joy :(

[ 69%] Linking CXX executable TestRenumber
/usr/bin/ld: cannot find -lssl
/usr/bin/ld: cannot find -lcrypto
/usr/bin/ld: cannot find -lcrypto
collect2: error: ld returned 1 exit status
tests/CMakeFiles/TestRenumber.dir/build.make:140: recipe for target 
'tests/TestRenumber' failed
make[2]: *** [tests/TestRenumber] Error 1
CMakeFiles/Makefile2:1233: recipe for target 
'tests/CMakeFiles/TestRenumber.dir/all' failed
make[1]: *** [tests/CMakeFiles/TestRenumber.dir/all] Error 2

[ 80%] Linking CXX static library libsubsurface_interface.a
[ 90%] Built target subsurface_interface
Makefile:127: recipe for target 'all' failed

Also, libgit2-dev and libcurl4-openssl-dev appear in the list of
recommended packages.  libgit2-dev depends on libcurl4-gnutls-dev and
installing libcurl4-openssl-dev will remove libgit2-dev.  OTOH,
libgit2 seems to be built locally for subsurface.  What's the deal?

Installing libssl-dev seems to be the key to success ;)


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: Compilation error with latest master

[Cristian Ionescu-Idbohrn]

On Fri, 30 Jan 2015, Cristian Ionescu-Idbohrn wrote:
 On Fri, 30 Jan 2015, Dirk Hohndel wrote:
  On Fri, Jan 30, 2015 at 06:21:45PM +, Pedro Neves wrote:
   Hi all:
  
   I'm trying to compile the latest subsurface master on my system
   (Debian/Sid) and I'm getting the following error.
   I've just recently installed my OS on this PC and I'm not sure
   if I'm missing any package:
  
  
   compiling qt-ui/completionmodels.cpp
   In file included from /usr/include/libxml2/libxml/parser.h:810:0,
from /usr/include/libxml2/libxml/globals.h:18,
from /usr/include/libxml2/libxml/threads.h:35,
from /usr/include/libxml2/libxml/xmlmemory.h:218,
from /usr/include/libxml2/libxml/tree.h:1306,
from dive.h:37,
from qt-gui.cpp:12:
   /usr/include/libxml2/libxml/encoding.h:31:26: fatal error: unicode/ucnv.h:
   No such file or directory
#include unicode/ucnv.h
 
  $ dpkg -S /usr/include/x86_64-linux-gnu/unicode/ucnv.h
  libicu-dev:amd64: /usr/include/x86_64-linux-gnu/unicode/ucnv.h
 
  I didn't install this explicitly, so this is implicitly pulled in
  by one of the dependencies of the packages that I DO explicitly
  require for a build on Debian / Ubuntu...

 And that may very well be a package dependency problem, as AFAICT:

   Package: libxml2-dev
   Depends: libxml2

 only.

See bug:

https://bugs.debian.org/776694

now resolved in version 2.9.2+dfsg1-3.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: Compilation error with latest master

[Cristian Ionescu-Idbohrn]

On Fri, 30 Jan 2015, Dirk Hohndel wrote:
 On Fri, Jan 30, 2015 at 06:21:45PM +, Pedro Neves wrote:
  Hi all:
 
  I'm trying to compile the latest subsurface master on my system (Debian/Sid)
  and I'm getting the following error.
  I've just recently installed my OS on this PC and I'm not sure if I'm
  missing any package:
 
 
  compiling qt-ui/completionmodels.cpp
  In file included from /usr/include/libxml2/libxml/parser.h:810:0,
   from /usr/include/libxml2/libxml/globals.h:18,
   from /usr/include/libxml2/libxml/threads.h:35,
   from /usr/include/libxml2/libxml/xmlmemory.h:218,
   from /usr/include/libxml2/libxml/tree.h:1306,
   from dive.h:37,
   from qt-gui.cpp:12:
  /usr/include/libxml2/libxml/encoding.h:31:26: fatal error: unicode/ucnv.h:
  No such file or directory
   #include unicode/ucnv.h

 $ dpkg -S /usr/include/x86_64-linux-gnu/unicode/ucnv.h
 libicu-dev:amd64: /usr/include/x86_64-linux-gnu/unicode/ucnv.h

 I didn't install this explicitly, so this is implicitly pulled in by one
 of the dependencies of the packages that I DO explicitly require for a
 build on Debian / Ubuntu...

And that may very well be a package dependency problem, as AFAICT:

Package: libxml2-dev
Depends: libxml2

only.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH] qmake: use a dedicated build script to generate ssrf-version.h

[Cristian Ionescu-Idbohrn]

On Fri, 30 Jan 2015, Lubomir I. Ivanov wrote:

I'll just nitpick on this one:

 diff --git a/scripts/write-version b/scripts/write-version
 new file mode 100644
 index 000..42505ae
 --- /dev/null
 +++ b/scripts/write-version
 @@ -0,0 +1,61 @@
 +#!/bin/sh

Try this same thing with adding:

set -eu
#set -x #xtrace

 +#
 +# arguments:
 +#$1 - target H file
 +#$2 - fallback version string
 +#$3 - os name {linux|darwin|win}
 +#
 +# !!! has no error checking ATM !!!
 +# should be started from where .git is!
 +#
 +
 +# set OS and TARGET

Any of these may fail in absence of

[ $# -eq 3 ] || {
echo ERROR: missing argument(s)   # alternative, usage message
exit 1
}

 +TARGET=$1
 +TARGET_TEMP=$TARGET.tmp
 +VERSION=$2
 +OS=$3
 +
 +# check if git is installed
 +GIT_ERROR=1

This is, AFAIK, not portable:

 +HAS_GIT=`git --version 2 NUL`
 ^^^
I'd write this:

 +if [ $HAS_GIT ]; then

as:

if HAS_GIT=$(git --version 2/dev/null)  [ $HAS_GIT ]; then

Backticks are considered deprecated.

 + # check if .git exists
 + if [ -d .git ]; then
^^
Quotes are useless in that context.

 + FULL_VER=`sh ./scripts/get-version linux`
 + GIT_ERROR=0
 + else
 + echo WARNING: not a git repository
 + fi
 +else
 + echo WARNING: git command not found
 +fi
 +
 +#if there was an error, fallback to .gitversion or use the dummy version
 +if [ $GIT_ERROR == 1 ]; then
 ^^ ^ ^
'==' is a bashism and quotes are useless.

 + if [ -f .gitversion ]; then
 ^^
Existence of that regular file and readabylity of the same is not
realy the same thing.

 + FULL_VER=`cat .gitversion`

read FULL_VER .gitversion

is more efficient.

 + else
 + FULL_VER=$VERSION
 + fi
 +fi
 +
 +# grab some strings from get-version

Both these:

 +CANONICAL_VER=`sh ./scripts/get-version full $FULL_VER`
 +OS_USABLE_VER=`sh ./scripts/get-version $OS $FULL_VER`

lack error handling, and deprecated backticks apply.

 +
 +# write to a temp file

Error handling (I know, this is really paranoic, but):

 +echo #define VERSION_STRING \$OS_USABLE_VER\  $TARGET_TEMP
 +echo #define GIT_VERSION_STRING \$FULL_VER\  $TARGET_TEMP
 +echo #define CANONICAL_VERSION_STRING \$CANONICAL_VER\  $TARGET_TEMP

that won't work if you happen to not have write permissions.

 +
 +# if the target file is missing create it
 +if [ ! -f $TARGET ]; then

or [ ! -w $TARGET ] ?

and more backticks:

 + CMD=`touch $TARGET`
 +fi
 +
 +# if the temp file and the target file differ, replace the target file with 
 the temp file
 +CMD=`diff -q $TARGET $TARGET_TEMP || cp $TARGET_TEMP $TARGET`
 +
 +# remove the temp file
 +CMD=`rm -f $TARGET_TEMP`


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: Compilation error with latest master

[Cristian Ionescu-Idbohrn]

On Fri, 30 Jan 2015, Pedro Neves wrote:

 I'm trying to compile the latest subsurface master on my system
 (Debian/Sid) and I'm getting the following error.
 I've just recently installed my OS on this PC and I'm not sure if
 I'm missing any package:


 compiling qt-ui/completionmodels.cpp
 In file included from /usr/include/libxml2/libxml/parser.h:810:0,
  from /usr/include/libxml2/libxml/globals.h:18,
  from /usr/include/libxml2/libxml/threads.h:35,
  from /usr/include/libxml2/libxml/xmlmemory.h:218,
  from /usr/include/libxml2/libxml/tree.h:1306,
  from dive.h:37,
  from qt-gui.cpp:12:
 /usr/include/libxml2/libxml/encoding.h:31:26: fatal error: unicode/ucnv.h: No 
 such file or directory
  #include unicode/ucnv.h
   ^
 compilation terminated.

That seems to be comming from package:

libicu-dev: /usr/include/x86_64-linux-gnu/unicode/ucnv.h

Yet another dependency?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: build broken as '-lusb-1.0' is missing...

[Cristian Ionescu-Idbohrn]

On Sat, 27 Dec 2014, Thiago Macieira wrote:
 On Saturday 27 December 2014 11:38:16 Cristian Ionescu-Idbohrn wrote:
  ...from the generated Makefile (on the master branch).
 
  If I manually add it:
 
  ... -ldivecomputer -lusb-1.0 -lxslt -lxml2 ...
 ^
  at a strategic point, on line 21:
 
  LIBS  = $(SUBLIBS) ...
 
  I get a ./subsurface executable.

 Indirect dependency due to static linking of libdivecomputer. We
 already have code to deal with that by parsing its .la file.

 How did you install libdivecomputer and where is the .la file?

I don't want to install libdivecomputer, nor marble.  I use `qmake'
arguments 'LIBS+=...' and 'INCLUDEPATH+=...' to point to marble and
libdivecomputer library locations.

Relative to subsurface repo code tree:

../libdivecomputer/src/.libs/libdivecomputer.a
../libdivecomputer/src/.libs/libdivecomputer.lai
../libdivecomputer/src/libdivecomputer.la
../libdivecomputer/src/.libs/libdivecomputer.la - ../libdivecomputer.la

Yes, the .la file refers to libusb:

# Libraries that this one depends upon.
dependency_libs=' -lusb-1.0'


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: build broken as '-lusb-1.0' is missing...

[Cristian Ionescu-Idbohrn]

On Sat, 27 Dec 2014, Thiago Macieira wrote:
 On Saturday 27 December 2014 14:14:00 Cristian Ionescu-Idbohrn wrote:
  On Sat, 27 Dec 2014, Thiago Macieira wrote:
   On Saturday 27 December 2014 11:38:16 Cristian Ionescu-Idbohrn wrote:
...from the generated Makefile (on the master branch).
   
If I manually add it:
... -ldivecomputer -lusb-1.0 -lxslt -lxml2 ...
   ^
   
at a strategic point, on line 21:
LIBS  = $(SUBLIBS) ...
   
I get a ./subsurface executable.
  
   Indirect dependency due to static linking of libdivecomputer. We
   already have code to deal with that by parsing its .la file.
  
   How did you install libdivecomputer and where is the .la file?
 
  I don't want to install libdivecomputer, nor marble.  I use `qmake'
  arguments 'LIBS+=...' and 'INCLUDEPATH+=...' to point to marble and
  libdivecomputer library locations.
 
  Relative to subsurface repo code tree:
 
  ../libdivecomputer/src/.libs/libdivecomputer.a
  ../libdivecomputer/src/.libs/libdivecomputer.lai
  ../libdivecomputer/src/libdivecomputer.la
  ../libdivecomputer/src/.libs/libdivecomputer.la - ../libdivecomputer.la
 
  Yes, the .la file refers to libusb:
 
  # Libraries that this one depends upon.
  dependency_libs=' -lusb-1.0'

 Add to your qmake command-line:

   LIBDCDEVEL=1

That works well, thanks.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

[PATCH] Add missing qt5-qmake package

[Cristian Ionescu-Idbohrn]

See 'Build-Depends:' in packaging/ubuntu/debian/control.

Signed-off-by: Cristian Ionescu-Idbohrn cristian.ionescu-idbo...@axis.com
---
 INSTALL | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/INSTALL b/INSTALL
index 2c88a28..37d4ec6 100644
--- a/INSTALL
+++ b/INSTALL
@@ -84,8 +84,9 @@ On Debian Jessie this seems to work
 sudo apt-get install git g++ make autoconf libtool cmake pkg-config \
libxml2-dev libxslt1-dev libzip-dev libsqlite3-dev \
libusb-1.0-0-dev libgit2-dev \
-   qt5-default qtchooser qttools5-dev-tools libqt5svg5-dev 
libqt5webkit5-dev \
-   libqt5qml5 libqt5quick5 libqt5declarative5 qtscript5-dev
+   qt5-default qt5-qmake qtchooser qttools5-dev-tools libqt5svg5-dev \
+   libqt5webkit5-dev libqt5qml5 libqt5quick5 libqt5declarative5 \
+   qtscript5-dev

 On Debian style distributions you can install the required development
 packages by running (once Qt version of subsurface is available in your
-- 
2.1.4
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: [PATCH] Add missing qt5-qmake package

[Cristian Ionescu-Idbohrn]

On Sat, 27 Dec 2014, Miika Turkia wrote:
  On 27 Dec 2014, at 19:04, Cristian Ionescu-Idbohrn 
  cristian.ionescu-idbo...@axis.com wrote:
 
 
  See 'Build-Depends:' in packaging/ubuntu/debian/control.
 
  Signed-off-by: Cristian Ionescu-Idbohrn cristian.ionescu-idbo...@axis.com
  ---
  INSTALL | 5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)
 
  diff --git a/INSTALL b/INSTALL
  index 2c88a28..37d4ec6 100644
  --- a/INSTALL
  +++ b/INSTALL
  @@ -84,8 +84,9 @@ On Debian Jessie this seems to work
  sudo apt-get install git g++ make autoconf libtool cmake pkg-config \
 libxml2-dev libxslt1-dev libzip-dev libsqlite3-dev \
 libusb-1.0-0-dev libgit2-dev \
  -qt5-default qtchooser qttools5-dev-tools libqt5svg5-dev 
  libqt5webkit5-dev \
  -libqt5qml5 libqt5quick5 libqt5declarative5 qtscript5-dev
  +qt5-default qt5-qmake qtchooser qttools5-dev-tools libqt5svg5-dev \
  +libqt5webkit5-dev libqt5qml5 libqt5quick5 libqt5declarative5 \
  +qtscript5-dev
 
  On Debian style distributions you can install the required development
  packages by running (once Qt version of subsurface is available in your

 isn't that dependency listed in qt5-default? Thus it is not
 explicitly needed in Subsurface. (At least this is the case in all
 the packages that we build with Launchpad.)

Debian package qt5-default shows it as:

Suggests: qt5-qmake, qtbase5-dev-tools

'Depends:' is:

Depends: qtbase5-dev, qtchooser

'Recommends:' or/and 'Suggests:' are not compulsory.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: ERROR: Package libgit2 not found

[Cristian Ionescu-Idbohrn]

On Mon, 10 Nov 2014, Cristian Ionescu-Idbohrn wrote:

 The problem is debian packaging.  There's already a bug filed for
 that, as I mentioned in an earlier post.  See
 https://bugs.debian.org/768482.

And the bug is now closed.  A new libgit2-dev with proper dependencies
was released to debian unstable:

Depends: libgit2-21 (= 0.21.2-1), libz-dev, libssl-dev

It will probably soon migrate to debian jessie (next stable).


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: ERROR: Package libgit2 not found

[Cristian Ionescu-Idbohrn]

On Mon, 10 Nov 2014, Dirk Hohndel wrote:
  On Nov 9, 2014, at 3:27 PM, Dirk Hohndel d...@hohndel.org wrote:
  Yes.  Thanks.  Unfortunately, there's still a problem when
  building subsurface from source :(
 
  It seems reasonable to at least document this on our web page.
  I'll try to do this today

 Done. Please let me know if this makes sense - it would be nice if
 people could add this to the translated pages as well

Look good.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

ERROR: Package libgit2 not found

[Cristian Ionescu-Idbohrn]

Anyone else seeing this (on recent debian unstable)?

,
| $ qmake
| Project ERROR: Package libgit2 not found
`

Still, I have the libgit2 packages installed:

,
| libgit2-21:amd64install
| libgit2-dev:amd64   install
`

I have the libs:

,
| /usr/lib/x86_64-linux-gnu/libgit2.a
| /usr/lib/x86_64-linux-gnu/libgit2.so
| /usr/lib/x86_64-linux-gnu/libgit2.so.0.21.2
| /usr/lib/x86_64-linux-gnu/libgit2.so.21
`

and I see header files installed in /usr/include/git2.

And:

,
| $ cat /usr/lib/x86_64-linux-gnu/pkgconfig/libgit2.pc
| libdir=/usr/lib/x86_64-linux-gnu
| includedir=/usr/include
|
| Name: libgit2
| Description: The git library, take 2
| Version: 0.21.2
| Requires.private:  openssl zlib libssh2
| Libs.private:  -lhttp_parser -lrt
| Libs: -L${libdir} -lgit2
| Cflags: -I${includedir}
`

What could cause the error?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: ERROR: Package libgit2 not found

[Cristian Ionescu-Idbohrn]

On Sun, 9 Nov 2014, Cristian Ionescu-Idbohrn wrote:

 Package libgit2-21 shows these dependencies:

   Depends: libc6 (= 2.15), libhttp-parser2.1 (= 2.1),
libssh2-1 (= 1.2.3), libssl1.0.0 (= 1.0.0),
zlib1g (= 1:1.1.4)

 while the corresponding -dev package lists:

   Depends: libgit2-21 (= 0.21.2-1), libz-dev, libssl-dev

 libssh2-1-dev is missing from the dependency list.

 Fortunately, some debian developers/maintainers should be watching
 this list and react.

 Still, until the problem is taken care of, the workaround is to
 install libssh2-1-dev.  The patch for subsurface-configure.pri should
 not be needed, but reveals the problem.

There's even a bug files against libgit2-dev:

https://bugs.debian.org/768482

Nevertheless:

,
| $ qmake
| Project ERROR: Package libgit2 not found
`

is still confusing.


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Re: ERROR: Package libgit2 not found

[Cristian Ionescu-Idbohrn]

On Sun, 9 Nov 2014, Dirk Hohndel wrote:
  On Nov 9, 2014, at 1:37 PM, Cristian Ionescu-Idbohrn 
  cristian.ionescu-idbo...@axis.com wrote:
 
  There's even a bug files against libgit2-dev:
 
  https://bugs.debian.org/768482
 
  Nevertheless:
 
  ,
  | $ qmake
  | Project ERROR: Package libgit2 not found
  `
 
  is still confusing.

 I agree. I’m not quite sure what to do about this. Do you have a
 suggestion for a patch that would fix the problem, Cristian?

I wish I had...  Oh well, at least we now know what the problem is and
how to work around it.  People with better knowledge of pkg-config
might have some idea.  /usr/lib/x86_64-linux-gnu/pkgconfig/libgit2.pc
looks like this on my box:

,
| libdir=/usr/lib/x86_64-linux-gnu
| includedir=/usr/include
|
| Name: libgit2
| Description: The git library, take 2
| Version: 0.21.2
| Requires.private:  openssl zlib libssh2
| Libs.private:  -lhttp_parser -lrt
| Libs: -L${libdir} -lgit2
| Cflags: -I${includedir}
`

What does that file show on other distributions?


Cheers,

-- 
Cristian
___
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface