On Thursday, January 23, 2020 9:17 PM, Tudor Roman <tudurom@gmail.com> wrote:

> Hello,
>
> I am currently learning wlroots and I got stuck on this issue.
>
> I am learning by trying to recreate/reimplement tinywl in C++, studying
> each encountered function.
>
> For some reason, on my laptop (with intel graphics), the following error
> happens in one in two frames:
>
> 2020-01-22 13:07:22 - [types/wlr_output.c:434] Tried to commit when a
> frame is pending

I think the In-Reply-To issue won't be fixed by this patch. We need to
strip the header field comments, no idea how to do this with Python
(tried to search, haven't found anything).

On Wednesday, January 22, 2020 7:00 PM, Christopher Wellons <wellons@nullprogram.com> wrote:

> I'm seeing quite a lot of oddball email today! Here's a similar
> situation:
>
> https://lists.sr.ht/~skeeto/public-inbox/<87pnfb69f3.fsf%40bernat.ch>
> https://lists.sr.ht/~skeeto/public-inbox/<87a76f4fro.fsf%40bernat.ch>
>
> The Message-ID and In-Reply-To match up, but they're not threaded
> together. Looking closely at the headers, Vincent's In-Reply-To includes
> a parenthetical. In my personal email archives that only occurs 0.49% of
> the time (i.e. 49 messages in every 10,000), so it's pretty rare.
> Perhaps lists is treating this as part of the ID? (Maybe the RFC says it
> should?)

This is not ideal, dropping the sources list prevents us from directly
piping the manifest to builds.sr.ht.

Chose to manually trigger the build instead.

Pushed, thanks!

A wlr_keyboard_group is used to multiplex events from multiple
keyboards.
---
 include/input.h |  8 ++++++++
 src/cursor.c    |  7 +++++--
 src/input.c     | 37 ++++++++++++++++++++++++++++++++++++-
 src/keyboard.c  | 43 +++++++++++++++++++++++++++++++++++++++++++
 src/meson.build |  1 +
 5 files changed, 93 insertions(+), 3 deletions(-)
 create mode 100644 src/keyboard.c

diff --git a/include/input.h b/include/input.h
index a8f229de0a6c..ef3b6aa4939f 100644
--- a/include/input.h
[message trimmed]

On Friday, January 17, 2020 12:22 AM, Reto Brunner <reto@labrat.space> wrote:

> Can we turn off the echo banner?
> I'm not sure what purpose it serves and it spams the logs

Agreed, it's an annoying "feature". Pushed, thanks!

Any reason why we don't always want to use the friendly format? Maybe we can
rename this to "formatdate" and rename the current "formatdate" function to
"formatfulldate"?

Templates can add a title attribute with the full date.

> +	"formatlistingdate": func(t time.Time) string {
> +		tYear, tMonth, tDay := t.Date()
> +
> +		now := time.Now()
> +		curYear, curMonth, curDay := now.Date()
> +		curWeekday := time.Duration(now.Weekday())
> +		if curWeekday == 0 {
> +			curWeekday = 7

Pushed as a single commit:

commit 83a5132d150e0db18653b5c90b9aff003fb02c68
Author: Dejan Strbac <me@dejanstrbac.com>
Date:   Thu Jan 16 13:36:03 2020 +0100

    Add current path to the global context

Pushed, thanks!