Re: [Elementary-dev-community] -lm and building the terminal
Hi, I'm not quite sure, but from what I know in the C level math.h is just a matter of adding a linker flag -lm Having that said, if doesn't work through vala make a vapi call to the C math lib... Cheers On Thu, 2012-09-13 at 08:52 +0530, Voldyman wrote: If we implement Math.floor function in pantheon terminal source wouldn't that work? On 13-Sep-2012, at 3:29 AM, David Gomes da...@elementaryos.org wrote: On the terminal we now need Math.floor and Math.ceil for zooming in and out using Ctrl-+ and Ctrl--. This brought along an issue when building the termina: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status make[2]: *** [pantheon-terminal] Error 1 In order to fix this, Ricotz suggested using the -lm flag on the linker. I tried to do it and then it worked just fine. However, for Victored and Eshat on Ubuntu (Luna), that made it stop working. It seems that on Ubuntu the linker includes lm by default, and on Arch it doesn't, and reincluding it causes troubles on Ubuntu. On CMakeLists.txt I just added the following to make it compile: set (CMAKE_EXE_LINKER_FLAGS -lm) Either way, I want Pantheon Terminal to compile across all GNU/Linux distributions, so I need a way to solve this. === modified file 'CMakeLists.txt' --- CMakeLists.txt 2012-07-26 20:23:31 + +++ CMakeLists.txt 2012-09-12 19:18:34 + @@ -30,6 +30,7 @@ find_package(PkgConfig) pkg_check_modules(DEPS REQUIRED gthread-2.0 gtk+-3.0 granite vte-2.90 libnotify gdk-3.0) +set (DEPS_LADD ${DEPS_LADD} -lm) add_definitions(${DEPS_CFLAGS}) link_libraries(${DEPS_LIBRARIES}) Ricotz suggesting the avobe path to fix the issue. I applied that patch and my patched CMakeLists.txt is attached. However, with this patch I can't build it on Arch, but Victored managed to built it on Ubuntu. I get the same good old error: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status So I'd appreciate some help guys, since I'm fairly new to CMake. Oh, and this is the branch we're working on: https://code.launchpad.net/~elementary-dev-community/pantheon-terminal/zoom-in-out Thank you. CMakeLists.txt -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp -- Darcy Brás da Silva dardeve...@cidadecool.com signature.asc Description: This is a digitally signed message part -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] -lm and building the terminal
Humm, if the fix was passing m, i guess cmake passes l prefix for us... Nice! On Thu, 2012-09-13 at 11:23 +0100, David Gomes wrote: Well Darcy, that email was @voldyman ;) Regarding the general issue, I found a way to fix it, thanks to devidfil's Github link, I just added: link_libraries (m) Right before the add_subdirectory call and it's working for everybody now. On Thu, Sep 13, 2012 at 11:16 AM, Darcy Brás da Silva dardeve...@cidadecool.com wrote: I'm not saying to implement it ourselves. I'm saying to make a vapi to have access of it from C math.h library in Vala, which is a very different thing :) On Thu, 2012-09-13 at 09:50 +0100, David Gomes wrote: Implementing Math.floor and Math.ceil ourselves would work, but I'm trying to avoid that for obvious reasons :) On Thu, Sep 13, 2012 at 4:22 AM, Voldyman voldyman...@gmail.com wrote: If we implement Math.floor function in pantheon terminal source wouldn't that work? On 13-Sep-2012, at 3:29 AM, David Gomes da...@elementaryos.org wrote: On the terminal we now need Math.floor and Math.ceil for zooming in and out using Ctrl-+ and Ctrl--. This brought along an issue when building the termina: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status make[2]: *** [pantheon-terminal] Error 1 In order to fix this, Ricotz suggested using the -lm flag on the linker. I tried to do it and then it worked just fine. However, for Victored and Eshat on Ubuntu (Luna), that made it stop working. It seems that on Ubuntu the linker includes lm by default, and on Arch it doesn't, and reincluding it causes troubles on Ubuntu. On CMakeLists.txt I just added the following to make it compile: set (CMAKE_EXE_LINKER_FLAGS -lm) Either way, I want Pantheon Terminal to compile across all GNU/Linux distributions, so I need a way to solve this. === modified file 'CMakeLists.txt' --- CMakeLists.txt 2012-07-26 20:23:31 + +++ CMakeLists.txt 2012-09-12 19:18:34 + @@ -30,6 +30,7 @@ find_package(PkgConfig) pkg_check_modules(DEPS REQUIRED gthread-2.0 gtk +-3.0 granite vte-2.90 libnotify gdk-3.0) +set (DEPS_LADD ${DEPS_LADD} -lm) add_definitions(${DEPS_CFLAGS}) link_libraries(${DEPS_LIBRARIES}) Ricotz suggesting the avobe path to fix the issue. I applied that patch and my patched CMakeLists.txt is attached. However, with this patch I can't build it on Arch, but Victored managed to built it on Ubuntu. I get the same good old error: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status So I'd appreciate some help guys, since I'm fairly new to CMake. Oh, and this is the branch we're working on: https://code.launchpad.net/~elementary-dev-community/pantheon-terminal/zoom-in-out
[Elementary-dev-community] Coding Style parentheses exception
Hey there guys, attached is the elementary OS coding style (We have this on a Google Doc, but the formatting is broken on it, so I'm rewriting it with Writer to then port to Google Docs again). I want to add an exception regarding the space character before opening parentheses, and I want to know if you all agree: Usually, we put a space character before opening parentheses: public string get_text () {} if (a == 5) return 4; for (i = 0; i maximum; i++) {} my_function_name (); Object my_instance = new Object (); However, if we have something like this: if ((e.state Gdk.ModifierType.MOD1_MASK) != 0) { Before the second opening parentheses on that line, I didn't put a space character because I think it looks quite bad. Do you guys agree that when we have 2 parentheses in a row we should not put a space character before the second one? if ( (e.state Gdk.ModifierType.MOD1_MASK) != 0) { That's how it would look like. Thank youg uys. elementaryos_coding_style.odt Description: application/vnd.oasis.opendocument.text -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Coding Style parentheses exception
I agree that we should have consistency but at this level just seems unproductive and a wast of time... I think the things that really make a difference in this situations are in cases like: if(condition){ } vs if(condition) { } or in functions: return_type func_name(param, param) { } vs return_type func_name(param, param { } vs return_type func_name(param, param) { } or cases like if(cond) { }else if { } vs if(cond){ } else { if(cond){ } } now getting to the _single_ space level just seems rather picky. Also it's not like all current projects managed to correctly apply the CGL ... There are things far more important that have not seen being address. Such as warnings and unreachable code. But apparently they don't seem to be a problem since they are IN the submitted code. :/ On Thu, 2012-09-13 at 11:07 -0500, Dane Henson wrote: I agree that the space should only be on the first opening parentheses, but not before the second. It looks cleaner to me. On Sep 13, 2012 8:32 AM, David Gomes da...@elementaryos.org wrote: Hey there guys, attached is the elementary OS coding style (We have this on a Google Doc, but the formatting is broken on it, so I'm rewriting it with Writer to then port to Google Docs again). I want to add an exception regarding the space character before opening parentheses, and I want to know if you all agree: Usually, we put a space character before opening parentheses: public string get_text () {} if (a == 5) return 4; for (i = 0; i maximum; i++) {} my_function_name (); Object my_instance = new Object (); However, if we have something like this: if ((e.state Gdk.ModifierType.MOD1_MASK) != 0) { Before the second opening parentheses on that line, I didn't put a space character because I think it looks quite bad. Do you guys agree that when we have 2 parentheses in a row we should not put a space character before the second one? if ( (e.state Gdk.ModifierType.MOD1_MASK) != 0) { That's how it would look like. Thank youg uys. -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp -- Darcy Brás da Silva dardeve...@cidadecool.com signature.asc Description: This is a digitally signed message part -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Coding Style parentheses exception
All those cases have been handled. It took less time for us to decide that this exception is a +1 than it took you to write that post, I don't think this is a waste of time. On Thu, Sep 13, 2012 at 5:40 PM, Darcy Brás da Silva dardeve...@cidadecool.com wrote: I agree that we should have consistency but at this level just seems unproductive and a wast of time... I think the things that really make a difference in this situations are in cases like: if(condition){ } vs if(condition) { } or in functions: return_type func_name(param, param) { } vs return_type func_name(param, param { } vs return_type func_name(param, param) { } or cases like if(cond) { }else if { } vs if(cond){ } else { if(cond){ } } now getting to the _single_ space level just seems rather picky. Also it's not like all current projects managed to correctly apply the CGL ... There are things far more important that have not seen being address. Such as warnings and unreachable code. But apparently they don't seem to be a problem since they are IN the submitted code. :/ On Thu, 2012-09-13 at 11:07 -0500, Dane Henson wrote: I agree that the space should only be on the first opening parentheses, but not before the second. It looks cleaner to me. On Sep 13, 2012 8:32 AM, David Gomes da...@elementaryos.org wrote: Hey there guys, attached is the elementary OS coding style (We have this on a Google Doc, but the formatting is broken on it, so I'm rewriting it with Writer to then port to Google Docs again). I want to add an exception regarding the space character before opening parentheses, and I want to know if you all agree: Usually, we put a space character before opening parentheses: public string get_text () {} if (a == 5) return 4; for (i = 0; i maximum; i++) {} my_function_name (); Object my_instance = new Object (); However, if we have something like this: if ((e.state Gdk.ModifierType.MOD1_MASK) != 0) { Before the second opening parentheses on that line, I didn't put a space character because I think it looks quite bad. Do you guys agree that when we have 2 parentheses in a row we should not put a space character before the second one? if ( (e.state Gdk.ModifierType.MOD1_MASK) != 0) { That's how it would look like. Thank youg uys. -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp -- Darcy Brás da Silva dardeve...@cidadecool.com -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Coding Style parentheses exception
IMHO we should keep the style comprehensible and easy to remember. Maybe that way we'll manage to actually follow it. If you're willing to put a lot of time into developing the coding style, I can suggest reading the relevant Code Craft chapter first (I can lend a copy of the book if you don't have it handy). I can't +1 or -1 because I don't code in Vala, but in BASH and Python I use both ( ( and ((, depending on readability of the content. I wouldn't be able to keep all those little rules in mind, I guess. -- Sergey Shnatsel Davidoff OS architect @ elementary -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] -lm and building the terminal
« l » is only here to tell gcc it is a lib, the lib name is actually « m ». Just remember that link_libraries is deprecated, please use target_link_libraries :P (well, I don't think it will be removed from CMake before a lot of time, but it is useless to keep writing code with it.) Lucas On 13/09/2012 14:28, Darcy Brás da Silva wrote: Humm, if the fix was passing m, i guess cmake passes l prefix for us... Nice! On Thu, 2012-09-13 at 11:23 +0100, David Gomes wrote: Well Darcy, that email was @voldyman ;) Regarding the general issue, I found a way to fix it, thanks to devidfil's Github link, I just added: link_libraries (m) Right before the add_subdirectory call and it's working for everybody now. On Thu, Sep 13, 2012 at 11:16 AM, Darcy Brás da Silva dardeve...@cidadecool.com wrote: I'm not saying to implement it ourselves. I'm saying to make a vapi to have access of it from C math.h library in Vala, which is a very different thing :) On Thu, 2012-09-13 at 09:50 +0100, David Gomes wrote: Implementing Math.floor and Math.ceil ourselves would work, but I'm trying to avoid that for obvious reasons :) On Thu, Sep 13, 2012 at 4:22 AM, Voldyman voldyman...@gmail.com wrote: If we implement Math.floor function in pantheon terminal source wouldn't that work? On 13-Sep-2012, at 3:29 AM, David Gomes da...@elementaryos.org wrote: On the terminal we now need Math.floor and Math.ceil for zooming in and out using Ctrl-+ and Ctrl--. This brought along an issue when building the termina: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status make[2]: *** [pantheon-terminal] Error 1 In order to fix this, Ricotz suggested using the -lm flag on the linker. I tried to do it and then it worked just fine. However, for Victored and Eshat on Ubuntu (Luna), that made it stop working. It seems that on Ubuntu the linker includes lm by default, and on Arch it doesn't, and reincluding it causes troubles on Ubuntu. On CMakeLists.txt I just added the following to make it compile: set (CMAKE_EXE_LINKER_FLAGS -lm) Either way, I want Pantheon Terminal to compile across all GNU/Linux distributions, so I need a way to solve this. === modified file 'CMakeLists.txt' --- CMakeLists.txt 2012-07-26 20:23:31 + +++ CMakeLists.txt 2012-09-12 19:18:34 + @@ -30,6 +30,7 @@ find_package(PkgConfig) pkg_check_modules(DEPS REQUIRED gthread-2.0 gtk +-3.0 granite vte-2.90 libnotify gdk-3.0) +set (DEPS_LADD ${DEPS_LADD} -lm) add_definitions(${DEPS_CFLAGS}) link_libraries(${DEPS_LIBRARIES}) Ricotz suggesting the avobe path to fix the issue. I applied that patch and my patched CMakeLists.txt is attached. However, with this patch I can't build it on Arch, but Victored managed to built it on Ubuntu. I get the same good old error: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit
Re: [Elementary-dev-community] -lm and building the terminal
Thanks xapantu, will do! On Thu, Sep 13, 2012 at 8:25 PM, xapantu xapa...@gmail.com wrote: « l » is only here to tell gcc it is a lib, the lib name is actually « m ». Just remember that link_libraries is deprecated, please use target_link_libraries :P (well, I don't think it will be removed from CMake before a lot of time, but it is useless to keep writing code with it.) Lucas On 13/09/2012 14:28, Darcy Brás da Silva wrote: Humm, if the fix was passing m, i guess cmake passes l prefix for us... Nice! On Thu, 2012-09-13 at 11:23 +0100, David Gomes wrote: Well Darcy, that email was @voldyman ;) Regarding the general issue, I found a way to fix it, thanks to devidfil's Github link, I just added: link_libraries (m) Right before the add_subdirectory call and it's working for everybody now. On Thu, Sep 13, 2012 at 11:16 AM, Darcy Brás da Silvadardeve...@cidadecool.com dardeve...@cidadecool.com wrote: I'm not saying to implement it ourselves. I'm saying to make a vapi to have access of it from C math.h library in Vala, which is a very different thing :) On Thu, 2012-09-13 at 09:50 +0100, David Gomes wrote: Implementing Math.floor and Math.ceil ourselves would work, but I'm trying to avoid that for obvious reasons :) On Thu, Sep 13, 2012 at 4:22 AM, Voldyman voldyman...@gmail.com voldyman...@gmail.com wrote: If we implement Math.floor function in pantheon terminal source wouldn't that work? On 13-Sep-2012, at 3:29 AM, David Gomes da...@elementaryos.org da...@elementaryos.org wrote: On the terminal we now need Math.floor and Math.ceil for zooming in and out using Ctrl-+ and Ctrl--. This brought along an issue when building the termina: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker command line /usr/lib/libm.so.6: could not read symbols: Invalid operation collect2: error: ld returned 1 exit status make[2]: *** [pantheon-terminal] Error 1 In order to fix this, Ricotz suggested using the -lm flag on the linker. I tried to do it and then it worked just fine. However, for Victored and Eshat on Ubuntu (Luna), that made it stop working. It seems that on Ubuntu the linker includes lm by default, and on Arch it doesn't, and reincluding it causes troubles on Ubuntu. On CMakeLists.txt I just added the following to make it compile: set (CMAKE_EXE_LINKER_FLAGS -lm) Either way, I want Pantheon Terminal to compile across all GNU/Linux distributions, so I need a way to solve this. === modified file 'CMakeLists.txt' --- CMakeLists.txt 2012-07-26 20:23:31 + +++ CMakeLists.txt 2012-09-12 19:18:34 + @@ -30,6 +30,7 @@ find_package(PkgConfig) pkg_check_modules(DEPS REQUIRED gthread-2.0 gtk +-3.0 granite vte-2.90 libnotify gdk-3.0) +set (DEPS_LADD ${DEPS_LADD} -lm) add_definitions(${DEPS_CFLAGS}) link_libraries(${DEPS_LIBRARIES}) Ricotz suggesting the avobe path to fix the issue. I applied that patch and my patched CMakeLists.txt is attached. However, with this patch I can't build it on Arch, but Victored managed to built it on Ubuntu. I get the same good old error: /usr/bin/ld: CMakeFiles/pantheon-terminal.dir/src/TerminalWidget.c.o: undefined reference to symbol 'floor@@GLIBC_2.2.5' /usr/bin/ld: note: 'floor@@GLIBC_2.2.5' is defined in DSO /usr/lib/libm.so.6 so try adding it to the linker
[Elementary-dev-community] Luna Update
Hey Guys, Time for another update on the status of Luna! Let's check the milestone: https://launchpad.net/elementary/+milestone/luna-beta1 Contractor has 1 bug Files has 6 bugs Granite has 1 bug Maya has 3 bugs Noise has 6 bugs LibPantheon has 1 bug Pantheon greeter has 1 bug Pantheon plugs has 1 bug Slingshot has 1 bug Switchboard has 2 bugs That's a total of 23 bugs left until we can release beta. Between all of us, we have this under control! Let's see if we can be beta-ready before the end of the month ;) If you don't see your app on this list, you should double check your beta1 milestone and see if you need to make a release. Afterwards, Please stop development on your app and help another app hit their beta1 milestone. If we all help each other out and pool our resources, we can knock this milestone out in no time at all and start work on beta2. Thanks again everyone and happy hacking!-- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Luna Update
23 bugs to go, these are very good news Daniel :) Regarding bug #1013836 (Slingshot's only bug to luna-beta1), it seems to have been triaged and it has also bounced back from luna-beta2 to luna-beta1. Interesting case I think we should look up :) On Thu, Sep 13, 2012 at 10:00 PM, Daniel Fore dan...@elementaryos.orgwrote: Hey Guys, Time for another update on the status of Luna! Let's check the milestone: https://launchpad.net/elementary/+milestone/luna-beta1 Contractor has 1 bug Files has 6 bugs Granite has 1 bug Maya has 3 bugs Noise has 6 bugs LibPantheon has 1 bug Pantheon greeter has 1 bug Pantheon plugs has 1 bug Slingshot has 1 bug Switchboard has 2 bugs That's a total of 23 bugs left until we can release beta. Between all of us, we have this under control! Let's see if we can be beta-ready before the end of the month ;) If you don't see your app on this list, you should double check your beta1 milestone and see if you need to make a release. Afterwards, *Please stop development on your app* and help another app hit their beta1 milestone. If we all help each other out and pool our resources, we can knock this milestone out in no time at all and start work on beta2. Thanks again everyone and happy hacking! -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Coding Style parentheses exception
Not that I do a lot of coding, but I'm +1 on the (( exception. It looks much more readable imho. On Thu, Sep 13, 2012 at 10:45 AM, Sergey Shnatsel Davidoff ser...@elementaryos.org wrote: IMHO we should keep the style comprehensible and easy to remember. Maybe that way we'll manage to actually follow it. If you're willing to put a lot of time into developing the coding style, I can suggest reading the relevant Code Craft chapter first (I can lend a copy of the book if you don't have it handy). I can't +1 or -1 because I don't code in Vala, but in BASH and Python I use both ( ( and ((, depending on readability of the content. I wouldn't be able to keep all those little rules in mind, I guess. -- Sergey Shnatsel Davidoff OS architect @ elementary -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp
Re: [Elementary-dev-community] Geary
Sergey, I know this has been a while, but I've been fighting fires and fixing bugs and had to put this off. I've committed your patches and they should be available in our next daily build. http://redmine.yorba.org/issues/5696 http://redmine.yorba.org/issues/5697 Thanks! -- Jim On Wed, Aug 22, 2012 at 5:28 AM, Sergey Shnatsel Davidoff ser...@elementaryos.org wrote: 2012/8/22 Eric Gregory e...@yorba.org Apport's traces could be quite helpful. We do have a PPA for daily builds for Geary and Shotwell. Would that help? How do we enable it? The daily build PPA is here, if that helps: https://launchpad.net/~yorba/+archive/daily-builds/ First of all, you'll need to make debugging symbols for Geary available from the PPA in addition to the stripped binary. In Debian/Ubuntu they're shipped in a separate -dbg package (e.g. geary-dbg). I've attached a patch that adds the geary-dbg package and also bumps debhelper version requirement to 8, as instructed by lintian. (By the way, I wrote an automated script for adding -dbg packages when we were enabling Apport in elementary. You can find at https://code.launchpad.net/~elementary-os/elementaryos/packaging-scripts) Second, you will need two config files for Apport: /usr/share/apport/package-hooks/source_geary.py and /etc/apport/crashdb.conf.d/geary-crashdb.conf (attached; generator script available from https://code.launchpad.net/~elementary-os/elementaryos/apport-hooks branch). On a stable system such as 12.04 you'll also need the user to manually enable Apport crash reporting by undoing this change: http://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu/precise/apport/ubuntu/revision/1948#etc/apport/crashdb.conf On elementary OS dailiy builds or on Ubuntu 12.10 it's not needed. This will enable crash submission, but that's just a half of the hunt. Most users don't have debugging symbols installed. Even if they do have some installed, it's not the whole dependency tree of an app. And even getting the whole tree is not completely sufficient for nontrivial reasons which I won't list here. Because of this crash reports need post-processing, or so-called retracing. Apport's retracer collects all required debugging symbols and extracts human-readable data from the core dump submitted with the crash report. Apport also has a component called crash-digger that scrapes LP bug reports and feeds all unretraced ones to the retracer, so the process is completely automated. However, there are several gotchas here. Obviously, somebody has to run crash-digger. Moreover, crash-digger should be invoked often, because if a PPA update is issued before a crash is retraced, most likely it will be impossible to retrace the crash. Apport-retrace does not support cross-architecture retracing, so there should be two instances of crash-digger, either on two machines (for i386 and amd64), or on an amd64 machine, with one instance placed in an i386 chroot. Apport-retrace needs some configs too, see man apport-retrace and our config dir as a real-life example: https://code.launchpad.net/~elementary-os/elementaryos/apport-retrace-sandbox Apport-retrace and crash-digger should be run from an Ubuntu 12.10 machine or a bzr checkout, because elementary is (surprisingly) the first non-Canonical party to adopt Apport, and prior to our adoption it hardcoded some Ubuntu-specific values and didn't support retracing crashes for LP other than Ubuntu. Running it from bzr checkout is explained here: https://bugs.launchpad.net/apport/+bug/1003506/comments/15 Crash-digger doesn't have a manpage, but I've documented its typical invocation as used in elementary, you can see it here: https://docs.google.com/document/d/16bUZqrSudlVt7Z7gAafS15U-ZlmvhdSMpLEL1wcatGg/edit#heading=h.v15mmt4if3m3 We already run two (amd64 and i386) instances of retracer, so if you're OK with adding our bot to bug supervisors group of Geary project in Launchpad (or getting crash reports only from elementary OS installations), you can simply use our retracer and not mess with the setup. Though our retracer is just me running crash-digger on my home PCs from time to time. 2012/8/21 Daniel Fore dan...@elementaryos.org Report bugs: http://redmine.yorba.org/projects/geary/issues Before you ask, if you can't find the Report a bug button, it's OK, because there's none. To report a bug you'll have to create an account (the old-school yet-another username/password pair - there's no OpenID or OAuth support), confirm it via email (my email confirmation was considered spam by Gmail; if it seems to get stuck, look in spam), re-enter the username and password to log in, then the navigate to Issues tab and click New issue. The New issue button won't be there unless you're logged in. Thanks Christian for showing it to me, I'd have never found it myself. I wonder how users of Yorba apps manage to figure it out... 2012/8/21 Daniel Fore dan...@elementaryos.org Check out the wiki for more:
[Elementary-dev-community] Fwd: [Shotwell] Shotwell 0.13: Call for Translation
Hey elementary devs, Another Yorba app could use our help with translations :) -- Forwarded message -- From: Lucas Beeler lu...@yorba.org Date: Thu, Sep 13, 2012 at 4:25 PM Subject: [Shotwell] Shotwell 0.13: Call for Translation To: shotwell shotw...@lists.yorba.org Hi Shotwell Fans, The next major release of Shotwell, 0.13, is due out in the next few weeks. Shotwell 0.13 is mostly a bug-fix and stability release, but it does include several exciting new features: * The Shotwell Extras plugins now include support for publishing photos to Tumblr * Shotwell correctly handles dates before 1969 * Shotwell lets you remove sensitive information from photos before publishing them to social media sites * Shotwell regenerates thumbnails automatically, making migrating libraries between computers easier * The Picasa connector now uses OAuth authentication for enhanced security Of course, all of these new features have added new text strings to Shotwell. So if you'd like to use Shotwell 0.13 in your native language, we invite you to update the translation files for your language to take these new strings into account. We froze the addition of new strings today, and we're accepting translation submissions for the 0.13 release until Wednesday, September 19. You can submit an updated translation on our Transifex project page at http://www.transifex.net/projects/p/shotwell/. Thanks for your help. We hope you're as excited about the future of Shotwell as we are! Regards, Lucas Lucas Beeler Shotwell Project Lead Developer Yorba Foundation ___ Shotwell mailing list shotw...@lists.yorba.org http://lists.yorba.org/cgi-bin/mailman/listinfo/shotwell -- Best Regards, Daniel Foré elementaryos.org -- Mailing list: https://launchpad.net/~elementary-dev-community Post to : elementary-dev-community@lists.launchpad.net Unsubscribe : https://launchpad.net/~elementary-dev-community More help : https://help.launchpad.net/ListHelp