Re: [ANNOUNCE] libinput 0.4.0
1. congratulations! 2. I cannot build weston anymore, it complained about the RPI backend compositor first, and now it's complaining about the udev_input_destroy, libinput-seat, and a bounce of other stuff I'm on an Intel laptop that built weston on pure wayland 'till yesterday, any idea what could cause this? Thanks On Tue, Jun 24, 2014 at 9:06 PM, Peter Hutterer peter.hutte...@who-t.net wrote: A new release of libinput, 0.4, is available. There are a number of API breaks in this version, every user of libinput will need some changes, sorry. * The logging system now works per-context instead of per library and dropped the userdata argument. * libinput_destroy() was changed to libinput_unref(), plus you now have libinput_ref() too * A couple of renames of constants for API consistency: LIBINPUT_KEYBOARD_KEY_STATE - LIBINPUT_KEY_STATE LIBINPUT_POINTER_VERTICAL_SCROLL - LIBINPUT_POINTER_SCROLL_VERTICAL LIBINPUT_POINTER_HORIZONTAL_SCROLL - LIBINPUT_POINTER_SCROLL_HORIZONTAL * Triggered by the logging changes, creation of udev seats is now different. Instead of libinput_udev_create_for_seat() use: libinput = libinput_udev_context_create(...) libinput_udev_assign_seat(libinput, seat name); Other changes include: * A couple of bugfixes, mainly in the touchpad code * Absolute event coordinates are now provided in mm, not device-specific coordinates. This doesn't change anything for callers using the respective _get_transformed() function. * ref/unref() is now available for the libinput context itself. The full commit shortlog is below: Hans de Goede (11): Add a timer subsystem evdev-mt-touchpad-buttons: Switch over to new timer subsystem evdev-mt-touchpad-tap: Switch over to new timer subsystem timer.h: Add #include libinput-util.h Add a log_msg_va function touchpad: tp_current_touch: Fix off by one error test: Add touchpad 2 finger scroll test touchpad: Avoid spurious motion event for scroll movement below threshold tp_post_twofinger_scroll: Move setting of t-is_pointer = false up a bit test: Prefix litest created device names with litest timer: Complain if a timer is set more than 5 seconds from now Jonas Ådahl (5): doc: Fix spelling event-gui: Don't drop fraction of pointer motion events Make ref count unref/ref() functions return resulting object pointer Make context reference counted test: Add context reference counting test Peter Hutterer (41): tools: add a script to push the doxygen output to freedesktop.org Drop empty FFI_CFLAGS test: fix compiler warning for is_debugger_attached test: fix compiler warnings for missing prototypes test: fix compiler warnings for missing field initializers test: silence compiler warning for C++ build test test: drop TEST_CFLAGS test: add --verbose flag to litests Mark the log function as attribute printf test: drop unused includes for libinput-util.h test: use check's strcmp API instead of strcmp touchpad: log the invalid event as bug, not just as info touchpad: Prefix tap-debugging message touchpad: always call into the the tap state machine test: add litest_assert_empty_queue helper function test: add a bunch of test for click behavior on touchpads test: add clickpad software button tests test: Add description for the T440 synaptics touchpad test: add a couple of top software button test Add two doxygen cross-references for the logging functions doc: ignore LIBINPUT_ATTRIBUTE_PRINTF in doxygen path: log two errors as application bugs test: fix a path test Document the top software button behavior tools: fix the publish-doc script evdev: keep the absinfo struct around instead of min/max evdev: force a resolution of 1 where no resolution is set Change absolute and touch events to use mm as default unit Add a function to get the size of a device Name-space the scroll event types tools: add a tool for GUI-based debugging Rename KEYBOARD_KEY_STATE to KEY_STATE test: move the interface declaration up udev: split libinput_udev context init into two functions Change the logging system to be per-context Add msleep() helper function test: add one more test for two-finger tapping Drop the deprecated LIBINPUT_POINTER_*_SCROLL enums Fix a few comments referring to udev_create_for_seat() Drop the deprecated libinput_udev_create_for_seat() configure.ac: libinput 0.4 Stephen Chandler Paul (1): touchpad: Fix typo in documentation git tag: 0.4.0 http://wayland.freedesktop.org/releases/libinput-0.4.0.tar.xz MD5: 60969c60e11792fab4ecaa7222f43697 libinput-0.4.0.tar.xz SHA1: f7638a1e7f1306976211c6b11fce19e68ad2ac8e
Re: [ANNOUNCE] libinput 0.4.0
On Tue, Jun 24, 2014 at 11:11:41PM -0700, Andrea Giammarchi wrote: 1. congratulations! 2. I cannot build weston anymore, it complained about the RPI backend compositor first, and now it's complaining about the udev_input_destroy, libinput-seat, and a bounce of other stuff Patch for weston is here: http://lists.freedesktop.org/archives/wayland-devel/2014-June/015590.html Cheers, Peter I'm on an Intel laptop that built weston on pure wayland 'till yesterday, any idea what could cause this? Thanks On Tue, Jun 24, 2014 at 9:06 PM, Peter Hutterer peter.hutte...@who-t.net wrote: A new release of libinput, 0.4, is available. There are a number of API breaks in this version, every user of libinput will need some changes, sorry. * The logging system now works per-context instead of per library and dropped the userdata argument. * libinput_destroy() was changed to libinput_unref(), plus you now have libinput_ref() too * A couple of renames of constants for API consistency: LIBINPUT_KEYBOARD_KEY_STATE - LIBINPUT_KEY_STATE LIBINPUT_POINTER_VERTICAL_SCROLL - LIBINPUT_POINTER_SCROLL_VERTICAL LIBINPUT_POINTER_HORIZONTAL_SCROLL - LIBINPUT_POINTER_SCROLL_HORIZONTAL * Triggered by the logging changes, creation of udev seats is now different. Instead of libinput_udev_create_for_seat() use: libinput = libinput_udev_context_create(...) libinput_udev_assign_seat(libinput, seat name); Other changes include: * A couple of bugfixes, mainly in the touchpad code * Absolute event coordinates are now provided in mm, not device-specific coordinates. This doesn't change anything for callers using the respective _get_transformed() function. * ref/unref() is now available for the libinput context itself. The full commit shortlog is below: Hans de Goede (11): Add a timer subsystem evdev-mt-touchpad-buttons: Switch over to new timer subsystem evdev-mt-touchpad-tap: Switch over to new timer subsystem timer.h: Add #include libinput-util.h Add a log_msg_va function touchpad: tp_current_touch: Fix off by one error test: Add touchpad 2 finger scroll test touchpad: Avoid spurious motion event for scroll movement below threshold tp_post_twofinger_scroll: Move setting of t-is_pointer = false up a bit test: Prefix litest created device names with litest timer: Complain if a timer is set more than 5 seconds from now Jonas Ådahl (5): doc: Fix spelling event-gui: Don't drop fraction of pointer motion events Make ref count unref/ref() functions return resulting object pointer Make context reference counted test: Add context reference counting test Peter Hutterer (41): tools: add a script to push the doxygen output to freedesktop.org Drop empty FFI_CFLAGS test: fix compiler warning for is_debugger_attached test: fix compiler warnings for missing prototypes test: fix compiler warnings for missing field initializers test: silence compiler warning for C++ build test test: drop TEST_CFLAGS test: add --verbose flag to litests Mark the log function as attribute printf test: drop unused includes for libinput-util.h test: use check's strcmp API instead of strcmp touchpad: log the invalid event as bug, not just as info touchpad: Prefix tap-debugging message touchpad: always call into the the tap state machine test: add litest_assert_empty_queue helper function test: add a bunch of test for click behavior on touchpads test: add clickpad software button tests test: Add description for the T440 synaptics touchpad test: add a couple of top software button test Add two doxygen cross-references for the logging functions doc: ignore LIBINPUT_ATTRIBUTE_PRINTF in doxygen path: log two errors as application bugs test: fix a path test Document the top software button behavior tools: fix the publish-doc script evdev: keep the absinfo struct around instead of min/max evdev: force a resolution of 1 where no resolution is set Change absolute and touch events to use mm as default unit Add a function to get the size of a device Name-space the scroll event types tools: add a tool for GUI-based debugging Rename KEYBOARD_KEY_STATE to KEY_STATE test: move the interface declaration up udev: split libinput_udev context init into two functions Change the logging system to be per-context Add msleep() helper function test: add one more test for two-finger tapping Drop the deprecated LIBINPUT_POINTER_*_SCROLL enums Fix a few comments referring to udev_create_for_seat() Drop the deprecated libinput_udev_create_for_seat()
Re: [ANNOUNCE] libinput 0.4.0
1. Thanks! :-) 2. how is this git patches via email is working for you, folks? It happens to be in a terminal from time to time and curl -O part of a thread, clean up patch part, and apply it ain't usually the fastest way to try one ... how many things am I missing in this process? how do you apply patches suggested via ML or cherry pick them? Thanks and apologies if the question is way too n00b Best Regards On Tue, Jun 24, 2014 at 11:16 PM, Peter Hutterer peter.hutte...@who-t.net wrote: On Tue, Jun 24, 2014 at 11:11:41PM -0700, Andrea Giammarchi wrote: 1. congratulations! 2. I cannot build weston anymore, it complained about the RPI backend compositor first, and now it's complaining about the udev_input_destroy, libinput-seat, and a bounce of other stuff Patch for weston is here: http://lists.freedesktop.org/archives/wayland-devel/2014-June/015590.html Cheers, Peter I'm on an Intel laptop that built weston on pure wayland 'till yesterday, any idea what could cause this? Thanks On Tue, Jun 24, 2014 at 9:06 PM, Peter Hutterer peter.hutte...@who-t.net wrote: A new release of libinput, 0.4, is available. There are a number of API breaks in this version, every user of libinput will need some changes, sorry. * The logging system now works per-context instead of per library and dropped the userdata argument. * libinput_destroy() was changed to libinput_unref(), plus you now have libinput_ref() too * A couple of renames of constants for API consistency: LIBINPUT_KEYBOARD_KEY_STATE - LIBINPUT_KEY_STATE LIBINPUT_POINTER_VERTICAL_SCROLL - LIBINPUT_POINTER_SCROLL_VERTICAL LIBINPUT_POINTER_HORIZONTAL_SCROLL - LIBINPUT_POINTER_SCROLL_HORIZONTAL * Triggered by the logging changes, creation of udev seats is now different. Instead of libinput_udev_create_for_seat() use: libinput = libinput_udev_context_create(...) libinput_udev_assign_seat(libinput, seat name); Other changes include: * A couple of bugfixes, mainly in the touchpad code * Absolute event coordinates are now provided in mm, not device-specific coordinates. This doesn't change anything for callers using the respective _get_transformed() function. * ref/unref() is now available for the libinput context itself. The full commit shortlog is below: Hans de Goede (11): Add a timer subsystem evdev-mt-touchpad-buttons: Switch over to new timer subsystem evdev-mt-touchpad-tap: Switch over to new timer subsystem timer.h: Add #include libinput-util.h Add a log_msg_va function touchpad: tp_current_touch: Fix off by one error test: Add touchpad 2 finger scroll test touchpad: Avoid spurious motion event for scroll movement below threshold tp_post_twofinger_scroll: Move setting of t-is_pointer = false up a bit test: Prefix litest created device names with litest timer: Complain if a timer is set more than 5 seconds from now Jonas Ådahl (5): doc: Fix spelling event-gui: Don't drop fraction of pointer motion events Make ref count unref/ref() functions return resulting object pointer Make context reference counted test: Add context reference counting test Peter Hutterer (41): tools: add a script to push the doxygen output to freedesktop.org Drop empty FFI_CFLAGS test: fix compiler warning for is_debugger_attached test: fix compiler warnings for missing prototypes test: fix compiler warnings for missing field initializers test: silence compiler warning for C++ build test test: drop TEST_CFLAGS test: add --verbose flag to litests Mark the log function as attribute printf test: drop unused includes for libinput-util.h test: use check's strcmp API instead of strcmp touchpad: log the invalid event as bug, not just as info touchpad: Prefix tap-debugging message touchpad: always call into the the tap state machine test: add litest_assert_empty_queue helper function test: add a bunch of test for click behavior on touchpads test: add clickpad software button tests test: Add description for the T440 synaptics touchpad test: add a couple of top software button test Add two doxygen cross-references for the logging functions doc: ignore LIBINPUT_ATTRIBUTE_PRINTF in doxygen path: log two errors as application bugs test: fix a path test Document the top software button behavior tools: fix the publish-doc script evdev: keep the absinfo struct around instead of min/max evdev: force a resolution of 1 where no resolution is set Change absolute and touch events to use mm as default unit Add a function to get the size of
Re: [ANNOUNCE] libinput 0.4.0
On Tue, Jun 24, 2014 at 11:25:51PM -0700, Andrea Giammarchi wrote: 1. Thanks! :-) 2. how is this git patches via email is working for you, folks? It happens to be in a terminal from time to time and curl -O part of a thread, clean up patch part, and apply it ain't usually the fastest way to try one ... how many things am I missing in this process? how do you apply patches suggested via ML or cherry pick them? given that you originally replied to an email on wayland-devel, you should have the patch in your inbox anyway. Just pass it to git-am and you're good to go, provided it applies on top of your current branch (applies cleanly on 671a1442a78). This is a fairly generic git question btw, there's probably about 5000 tutorials on google for how to do it with every email client ever written :) Cheers, Peter Thanks and apologies if the question is way too n00b Best Regards On Tue, Jun 24, 2014 at 11:16 PM, Peter Hutterer peter.hutte...@who-t.net wrote: On Tue, Jun 24, 2014 at 11:11:41PM -0700, Andrea Giammarchi wrote: 1. congratulations! 2. I cannot build weston anymore, it complained about the RPI backend compositor first, and now it's complaining about the udev_input_destroy, libinput-seat, and a bounce of other stuff Patch for weston is here: http://lists.freedesktop.org/archives/wayland-devel/2014-June/015590.html Cheers, Peter I'm on an Intel laptop that built weston on pure wayland 'till yesterday, any idea what could cause this? Thanks On Tue, Jun 24, 2014 at 9:06 PM, Peter Hutterer peter.hutte...@who-t.net wrote: A new release of libinput, 0.4, is available. There are a number of API breaks in this version, every user of libinput will need some changes, sorry. * The logging system now works per-context instead of per library and dropped the userdata argument. * libinput_destroy() was changed to libinput_unref(), plus you now have libinput_ref() too * A couple of renames of constants for API consistency: LIBINPUT_KEYBOARD_KEY_STATE - LIBINPUT_KEY_STATE LIBINPUT_POINTER_VERTICAL_SCROLL - LIBINPUT_POINTER_SCROLL_VERTICAL LIBINPUT_POINTER_HORIZONTAL_SCROLL - LIBINPUT_POINTER_SCROLL_HORIZONTAL * Triggered by the logging changes, creation of udev seats is now different. Instead of libinput_udev_create_for_seat() use: libinput = libinput_udev_context_create(...) libinput_udev_assign_seat(libinput, seat name); Other changes include: * A couple of bugfixes, mainly in the touchpad code * Absolute event coordinates are now provided in mm, not device-specific coordinates. This doesn't change anything for callers using the respective _get_transformed() function. * ref/unref() is now available for the libinput context itself. The full commit shortlog is below: Hans de Goede (11): Add a timer subsystem evdev-mt-touchpad-buttons: Switch over to new timer subsystem evdev-mt-touchpad-tap: Switch over to new timer subsystem timer.h: Add #include libinput-util.h Add a log_msg_va function touchpad: tp_current_touch: Fix off by one error test: Add touchpad 2 finger scroll test touchpad: Avoid spurious motion event for scroll movement below threshold tp_post_twofinger_scroll: Move setting of t-is_pointer = false up a bit test: Prefix litest created device names with litest timer: Complain if a timer is set more than 5 seconds from now Jonas Ådahl (5): doc: Fix spelling event-gui: Don't drop fraction of pointer motion events Make ref count unref/ref() functions return resulting object pointer Make context reference counted test: Add context reference counting test Peter Hutterer (41): tools: add a script to push the doxygen output to freedesktop.org Drop empty FFI_CFLAGS test: fix compiler warning for is_debugger_attached test: fix compiler warnings for missing prototypes test: fix compiler warnings for missing field initializers test: silence compiler warning for C++ build test test: drop TEST_CFLAGS test: add --verbose flag to litests Mark the log function as attribute printf test: drop unused includes for libinput-util.h test: use check's strcmp API instead of strcmp touchpad: log the invalid event as bug, not just as info touchpad: Prefix tap-debugging message touchpad: always call into the the tap state machine test: add litest_assert_empty_queue helper function test: add a bunch of test for click behavior on touchpads test: add clickpad software button tests test: Add description for the T440 synaptics touchpad
Re: [ANNOUNCE] libinput 0.4.0
Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. Best Regards On Tue, Jun 24, 2014 at 11:33 PM, Peter Hutterer peter.hutte...@who-t.net wrote: On Tue, Jun 24, 2014 at 11:25:51PM -0700, Andrea Giammarchi wrote: 1. Thanks! :-) 2. how is this git patches via email is working for you, folks? It happens to be in a terminal from time to time and curl -O part of a thread, clean up patch part, and apply it ain't usually the fastest way to try one ... how many things am I missing in this process? how do you apply patches suggested via ML or cherry pick them? given that you originally replied to an email on wayland-devel, you should have the patch in your inbox anyway. Just pass it to git-am and you're good to go, provided it applies on top of your current branch (applies cleanly on 671a1442a78). This is a fairly generic git question btw, there's probably about 5000 tutorials on google for how to do it with every email client ever written :) Cheers, Peter Thanks and apologies if the question is way too n00b Best Regards On Tue, Jun 24, 2014 at 11:16 PM, Peter Hutterer peter.hutte...@who-t.net wrote: On Tue, Jun 24, 2014 at 11:11:41PM -0700, Andrea Giammarchi wrote: 1. congratulations! 2. I cannot build weston anymore, it complained about the RPI backend compositor first, and now it's complaining about the udev_input_destroy, libinput-seat, and a bounce of other stuff Patch for weston is here: http://lists.freedesktop.org/archives/wayland-devel/2014-June/015590.html Cheers, Peter I'm on an Intel laptop that built weston on pure wayland 'till yesterday, any idea what could cause this? Thanks On Tue, Jun 24, 2014 at 9:06 PM, Peter Hutterer peter.hutte...@who-t.net wrote: A new release of libinput, 0.4, is available. There are a number of API breaks in this version, every user of libinput will need some changes, sorry. * The logging system now works per-context instead of per library and dropped the userdata argument. * libinput_destroy() was changed to libinput_unref(), plus you now have libinput_ref() too * A couple of renames of constants for API consistency: LIBINPUT_KEYBOARD_KEY_STATE - LIBINPUT_KEY_STATE LIBINPUT_POINTER_VERTICAL_SCROLL - LIBINPUT_POINTER_SCROLL_VERTICAL LIBINPUT_POINTER_HORIZONTAL_SCROLL - LIBINPUT_POINTER_SCROLL_HORIZONTAL * Triggered by the logging changes, creation of udev seats is now different. Instead of libinput_udev_create_for_seat() use: libinput = libinput_udev_context_create(...) libinput_udev_assign_seat(libinput, seat name); Other changes include: * A couple of bugfixes, mainly in the touchpad code * Absolute event coordinates are now provided in mm, not device-specific coordinates. This doesn't change anything for callers using the respective _get_transformed() function. * ref/unref() is now available for the libinput context itself. The full commit shortlog is below: Hans de Goede (11): Add a timer subsystem evdev-mt-touchpad-buttons: Switch over to new timer subsystem evdev-mt-touchpad-tap: Switch over to new timer subsystem timer.h: Add #include libinput-util.h Add a log_msg_va function touchpad: tp_current_touch: Fix off by one error test: Add touchpad 2 finger scroll test touchpad: Avoid spurious motion event for scroll movement below threshold tp_post_twofinger_scroll: Move setting of t-is_pointer = false up a bit test: Prefix litest created device names with litest timer: Complain if a timer is set more than 5 seconds from now Jonas Ådahl (5): doc: Fix spelling event-gui: Don't drop fraction of pointer motion events Make ref count unref/ref() functions return resulting object pointer Make context reference counted test: Add context reference counting test Peter Hutterer (41): tools: add a script to push the doxygen output to freedesktop.org Drop empty FFI_CFLAGS test: fix compiler warning for is_debugger_attached test: fix compiler warnings for missing prototypes test: fix compiler warnings for missing field initializers test: silence compiler warning for C++ build test test: drop TEST_CFLAGS test: add --verbose flag to litests Mark the log function as attribute printf test: drop unused includes for libinput-util.h test: use check's
Re: [ANNOUNCE] libinput 0.4.0
Ah, something else I forgot to mention: git-am requires that the patches be in .mbox format, otherwise it won't work properly. I don't think there's a way to do this in gmail since the last time I checked, it can't save e-mails in that format. To actually take advantage of this sort of thing you'd need to use a client of some sort instead of the web interface. On Wed, 2014-06-25 at 02:59 -0400, Lyude wrote: On Tue, 2014-06-24 at 23:40 -0700, Andrea Giammarchi wrote: Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. If you're referring to e-mail attachments that's not what he's getting at. When you apply a patch off the mailing list you literally apply the e-mail itself, not an attachment. Hence why you'd use git-am instead of git-apply. git-am takes a saved copy of the e-mail (in mbox format), extracts the inline patch for you, and applies it to the tree. It's somewhat of a different process then just applying a plain old patches. It's a very different system then most projects nowadays use, since most projects will just do pull requests and that sort of stuff :). So seeing misunderstandings like this isn't at all unusual or something to be disheartened about. We're not the only project to do it though, you'll see that the linux kernel requires you to submit patches in this manner too. It's mostly done because with organized projects like this, having the ability to easily review a patch inline is necessary to discuss changes in the protocol, point out typos and errors in patches, etc. There are alternatives for this, for example Github has the ability to allow you to comment inline on commits in pull requests, but a lot of people here just prefer to use their e-mail clients and whatever text editor they fancy. Hope I clarified things a bit! Cheers, Lyude signature.asc Description: This is a digitally signed message part ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
On Wed, Jun 25, 2014 at 9:02 AM, Lyude thatsly...@gmail.com wrote: Ah, something else I forgot to mention: git-am requires that the patches be in .mbox format, otherwise it won't work properly. I don't think there's a way to do this in gmail since the last time I checked, it can't save e-mails in that format. To actually take advantage of this sort of thing you'd need to use a client of some sort instead of the web interface. I *think* you can get a mbox-compatible format that can be used with git am from Gmail by clicking on the Show Original option in the More menu that is shown as a triangle in the upper right corner of mail message. Using a mail client that let's you save (multiple) mails quickly to a .mbox-formatted file (like mutt for example) is preferable though. Cheers, Silvan ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
On Wed, Jun 25, 2014 at 02:06:32PM +1000, Peter Hutterer wrote: A new release of libinput, 0.4, is available. [...] git tag: 0.4.0 http://wayland.freedesktop.org/releases/libinput-0.4.0.tar.xz MD5: 60969c60e11792fab4ecaa7222f43697 libinput-0.4.0.tar.xz SHA1: f7638a1e7f1306976211c6b11fce19e68ad2ac8e libinput-0.4.0.tar.xz SHA256: 4a6f2990f65384df012d45e5f58732fae06abb53318bf23e99c818c36184a7d1 libinput-0.4.0.tar.xz PGP: http://wayland.freedesktop.org/releases/libinput-0.4.0.tar.xz.sig well, things never quite work the way they should the first time round, and our new release script is no exception :) The correct path for the tarballs is: http://www.freedesktop.org/software/libinput/libinput-0.4.0.tar.xz PGP: http://www.freedesktop.org/software/libinput/libinput-0.4.0.tar.xz.sig I'll be removing the ones from the originally posted link soon. Cheers, Peter pgp0vVQTxqCHg.pgp Description: PGP signature ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
RE: [ANNOUNCE] libinput 0.4.0
From: wayland-devel [mailto:wayland-devel-boun...@lists.freedesktop.org] On Behalf Of Andrea Giammarchi Sent: Wednesday, June 25, 2014 12:11 PM To: Peter Hutterer Cc: Jonas Ådahl; wayland-devel@lists.freedesktop.org Subject: Re: [ANNOUNCE] libinput 0.4.0 Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. You can refer the below link to download the emails from gmail in the mbox format: http://www.pcworld.com/article/2068494/google-finally-lets-you-download-your-gmail-and-calendar-data-heres-how.html If you are using Thunderbird as email client, then it saves all the mails in the mbox format in the destination location. So you can apply patches from there directly. Thank-you, Hebbar ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
On Wed, Jun 25, 2014 at 2:59 AM, Lyude thatsly...@gmail.com wrote: On Tue, 2014-06-24 at 23:40 -0700, Andrea Giammarchi wrote: Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. If you're referring to e-mail attachments that's not what he's getting at. When you apply a patch off the mailing list you literally apply the e-mail itself, not an attachment. Hence why you'd use git-am instead of git-apply. git-am takes a saved copy of the e-mail (in mbox format), extracts the inline patch for you, and applies it to the tree. It's somewhat of a different process then just applying a plain old patches. It's a very different system then most projects nowadays use, since most projects will just do pull requests and that sort of stuff :). So seeing misunderstandings like this isn't at all unusual or something to be disheartened about. We're not the only project to do it though, you'll see that the linux kernel requires you to submit patches in this manner too. It's mostly done because with organized projects like this, having the ability to easily review a patch inline is necessary to discuss changes in the protocol, point out typos and errors in patches, etc. There are alternatives for this, for example Github has the ability to allow you to comment inline on commits in pull requests, but a lot of people here just prefer to use their e-mail clients and whatever text editor they fancy. It's a really outdated system, especially considering Pipermail doesn't have an mbox link for download. My workaround is usually to hunt around for the archive on markmail.org or marc.info and look for the mbox download, but those don't seem to be tracking wayland-devel. So yeah, if you haven't subscribed to the list, hope you like copying and pasting patch files and cleaning them up to get them to apply! (Why do we subject ourselves to this terrible infrastructure?) Hope I clarified things a bit! Cheers, Lyude ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel -- Jasper ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
Wow, I couldn't expect all these detailed replies and yes, I haven't seen this way to deal with patches yet but thank you all for the help, I'll figure out following your instructions and now I know what to google too :-) Best Regards On Wed, Jun 25, 2014 at 5:56 AM, Jasper St. Pierre jstpie...@mecheye.net wrote: On Wed, Jun 25, 2014 at 2:59 AM, Lyude thatsly...@gmail.com wrote: On Tue, 2014-06-24 at 23:40 -0700, Andrea Giammarchi wrote: Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. If you're referring to e-mail attachments that's not what he's getting at. When you apply a patch off the mailing list you literally apply the e-mail itself, not an attachment. Hence why you'd use git-am instead of git-apply. git-am takes a saved copy of the e-mail (in mbox format), extracts the inline patch for you, and applies it to the tree. It's somewhat of a different process then just applying a plain old patches. It's a very different system then most projects nowadays use, since most projects will just do pull requests and that sort of stuff :). So seeing misunderstandings like this isn't at all unusual or something to be disheartened about. We're not the only project to do it though, you'll see that the linux kernel requires you to submit patches in this manner too. It's mostly done because with organized projects like this, having the ability to easily review a patch inline is necessary to discuss changes in the protocol, point out typos and errors in patches, etc. There are alternatives for this, for example Github has the ability to allow you to comment inline on commits in pull requests, but a lot of people here just prefer to use their e-mail clients and whatever text editor they fancy. It's a really outdated system, especially considering Pipermail doesn't have an mbox link for download. My workaround is usually to hunt around for the archive on markmail.org or marc.info and look for the mbox download, but those don't seem to be tracking wayland-devel. So yeah, if you haven't subscribed to the list, hope you like copying and pasting patch files and cleaning them up to get them to apply! (Why do we subject ourselves to this terrible infrastructure?) Hope I clarified things a bit! Cheers, Lyude ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel -- Jasper ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
Reading gmail using a local client works fine, I have done several git am calls using this, after doing save-as to the email. That said, it may be a nice idea to not update the master for weston at the same time the master for libinput is updated. On 06/25/2014 12:02 AM, Lyude wrote: Ah, something else I forgot to mention: git-am requires that the patches be in .mbox format, otherwise it won't work properly. I don't think there's a way to do this in gmail since the last time I checked, it can't save e-mails in that format. To actually take advantage of this sort of thing you'd need to use a client of some sort instead of the web interface. ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
On Wed, Jun 25, 2014 at 08:56:55AM -0400, Jasper St. Pierre wrote: On Wed, Jun 25, 2014 at 2:59 AM, Lyude thatsly...@gmail.com wrote: On Tue, 2014-06-24 at 23:40 -0700, Andrea Giammarchi wrote: Yeah Peter, I know how to apply a patch and no, I don't have any of them in my inbox, that's why I've asked what was wrong in my process ... right now I'm on Gmail in chrome and wayland-devel channel and zero attachment whatsoever. If you're referring to e-mail attachments that's not what he's getting at. When you apply a patch off the mailing list you literally apply the e-mail itself, not an attachment. Hence why you'd use git-am instead of git-apply. git-am takes a saved copy of the e-mail (in mbox format), extracts the inline patch for you, and applies it to the tree. It's somewhat of a different process then just applying a plain old patches. It's a very different system then most projects nowadays use, since most projects will just do pull requests and that sort of stuff :). So seeing misunderstandings like this isn't at all unusual or something to be disheartened about. We're not the only project to do it though, you'll see that the linux kernel requires you to submit patches in this manner too. It's mostly done because with organized projects like this, having the ability to easily review a patch inline is necessary to discuss changes in the protocol, point out typos and errors in patches, etc. There are alternatives for this, for example Github has the ability to allow you to comment inline on commits in pull requests, but a lot of people here just prefer to use their e-mail clients and whatever text editor they fancy. It's a really outdated system, especially considering Pipermail doesn't have an mbox link for download. My workaround is usually to hunt around for the archive on markmail.org or marc.info and look for the mbox download, but those don't seem to be tracking wayland-devel. So yeah, if you haven't subscribed to the list, hope you like copying and pasting patch files and cleaning them up to get them to apply! (Why do we subject ourselves to this terrible infrastructure?) setting up patchwork should be an option, that provides at least the raw patch format and it's already setup for a couple of other fdo projects. Cheers, Peter ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
On Wed, Jun 25, 2014 at 12:07:07PM -0700, Bill Spitzak wrote: Reading gmail using a local client works fine, I have done several git am calls using this, after doing save-as to the email. That said, it may be a nice idea to not update the master for weston at the same time the master for libinput is updated. I'm assuming you mean nice idea to update because not updating doesn't make a lot of sense to me. The libinput announcement had the weston patch supplied on the list and Jonas pushed it to weston less than 2h later. Which is not perfect but IMO still not that bad. especially given that the libinput announcement had a number of warnings about API breaks in it... Cheers, Peter On 06/25/2014 12:02 AM, Lyude wrote: Ah, something else I forgot to mention: git-am requires that the patches be in .mbox format, otherwise it won't work properly. I don't think there's a way to do this in gmail since the last time I checked, it can't save e-mails in that format. To actually take advantage of this sort of thing you'd need to use a client of some sort instead of the web interface. ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
Re: [ANNOUNCE] libinput 0.4.0
Sorry I did not realize that it had actually been updated. Updating within 2 hours certainly seems ok to me. On 06/25/2014 03:01 PM, Peter Hutterer wrote: On Wed, Jun 25, 2014 at 12:07:07PM -0700, Bill Spitzak wrote: Reading gmail using a local client works fine, I have done several git am calls using this, after doing save-as to the email. That said, it may be a nice idea to not update the master for weston at the same time the master for libinput is updated. I'm assuming you mean nice idea to update because not updating doesn't make a lot of sense to me. The libinput announcement had the weston patch supplied on the list and Jonas pushed it to weston less than 2h later. Which is not perfect but IMO still not that bad. especially given that the libinput announcement had a number of warnings about API breaks in it... Cheers, Peter On 06/25/2014 12:02 AM, Lyude wrote: Ah, something else I forgot to mention: git-am requires that the patches be in .mbox format, otherwise it won't work properly. I don't think there's a way to do this in gmail since the last time I checked, it can't save e-mails in that format. To actually take advantage of this sort of thing you'd need to use a client of some sort instead of the web interface. ___ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel
