Re: [dev] tabbed: Use PWD xprop to set workdir for spawned clients. Any clear cons?

From: Hiltjo Posthuma <hiltjo_AT_codemadness.org>
Date: Fri, 2 Jul 2021 17:53:27 +0200

On Fri, Jul 02, 2021 at 01:16:41AM +0500, Nikita Zlobin wrote:
> I use tabbed for multitab terminal emulator. But I lacked one feature,
> presenting in at least some vte-based terminals: when new tab is
> opened, it inherits workdir from previous active tab.
>
> I tried to find, it same could be implemented with tabbed. I found
> following solution, which would work independently of what exact
> underlying terminal is used to be spawned. When spawning new tab, tabbed
> checks for certain X property, where workdir is written (obviously, by
> client itself). If found, then it uses this path as workdir for new
> client. This seems simplest possible way for me. Since shells
> (including dash, let alone bash) set PWD env var, containing current
> workdir, I decided to use this name for new X prop too.
>
> And to adapt unprepared terminals - appended snippet to ~/.bashrc:
>
> workdir_to_xprop() {
> [ -v WINDOWID ] && \
> xprop -id "${WINDOWID}" -f PWD 8s -set "PWD" "${PWD}"
> }
>
> cd() {
> builtin cd "$_AT_" || return
> [ "$OLDPWD" = "$PWD" ] || {
> workdir_to_xprop
> }
> }
>
> workdir_to_xprop
>
> (can't be sure, how correct it is, but usually it works, as even path
> change from midnight commander is counted in this way too).
>
> I tried to submit it to hackers ML for simple reason: this list is
> described to be for patches, directed to upstream. As for my patches
> (there were more patches, mostly with fixes) - I targeted them for
> upstream, so sending them there was logical. Too bad - sending patches
> in the way it's required was first ever time, so I wasted couple of
> attempts, while finally sent them in most correct way (not sure if I
> can do it again). Looks like, whoever checks this got tired from all of
> this.
>
> I hoped, in case if any deficiencies, get resonable arguments why. But
> I got no replies besides two:
> - one is argumentless -1 against this very feature: some claimed they
> don't use it, despite I have at least 1 example where it presents;
> - second is simply that if has to go to dev
> (well, finally I got reason to subscribe here, but not for this yet).
>
> https://lists.suckless.org/hackers/2006/17451.html
>
> Main question is in topic.
> I don't ask about pros, which are just obvious for me :)
> There could be more of them, but for now I'm too lost to even
> formulate something more specific. Did not expect such attitude, looks
> like as if you have to deserve to even be replied. Or could be that
> hackers only react if you proposal affects what's in their current
> coding focus (yet another wild guess).
>

Hi,

What "attitude"? Theres no need to be negative in the assumptions or be
passive-aggressive. In any case it won't help.

As one of the few suckless devs that is still active I still do this as a
hobby. Interacting with such negative people is low on my priority todo list
and makes it less fun.

Reading back the thread I see no issue. I even reviewed the patch and gave
feedback.

In the meantime a new (partial, "temporary") maintainer has been assigned to
tabbed (me), because the old maintainer did not have time anymore.

Like the other thread you posted, if you have ideas: implement them and feel
free to share them if you want do, but don't demand.

-- 
Kind regards,
Hiltjo
Received on Fri Jul 02 2021 - 17:53:27 CEST

This archive was generated by hypermail 2.3.0 : Fri Jul 02 2021 - 18:00:08 CEST