Re: EBrowse obsolete?

[Tim Cross]

Jim Porter <jporterbugs@gmail.com> writes:

> On 7/21/2022 10:39 AM, Stefan Monnier wrote:
>>> I don't use VS Code myself, but from talking with people who do, my
>>> understanding is that it supports LSP over SSH. (VS Code is a bit different
>>> in that it requires you to install a specific VS Code server on the remote
>>> host though, so maybe the closest analogy in Emacs would really be using
>>> emacsclient over TCP.)
>> Indeed, I don't think we can hope to get good performance with an
>> approach like that of Tramp which works hard to try and avoid requiring
>> installation of a "Emacs server" on the other end.
>> This said, it's probably possible to make it "good enough" in most
>> cases, tho it may require extra work.
>
> I think there are probably two separate (but somewhat related) issues:
>
> 1. LSP over Tramp is "slow". I'm sure there are things we can do to make 
> Tramp faster,
> especially with "well-behaved" remote hosts. For example, maybe Tramp can get 
> more info
> about a host upon connection and then use that to avoid some work later on.
>
> 2. LSP over Tramp is "unreliable". I think this is the more-pressing issue. 
> It's one thing
> if a remote request takes longer than expected, but another thing entirely if 
> it just
> fails to work every so often. When trying to use Eglot + Tramp on some larger 
> projects,
> I'd get hit with "Forbidden reentrant call of Tramp" every so often, which, 
> if memory
> serves, usually required a bit of poking at Eglot, or at least retrying the 
> operation I
> wanted to do. Maybe this has improved since the last time I tried this, but 
> (I think)
> fixing these errors would go a long way towards resolving people's 
> complaints. That's
> probably easier said than done though...

I use to use tramp a lot. However, these days, I only use it for ad hoc
editing.

What I found worked much better and took a lot less configuration was to
use FUSE to mount a local ssh filesystem of the remote repository and
then just edit it as if it was local. I would then just use an ssh
session inside emacs to execute/run things.

Having said that, since the growth in git/hg, the amount of remote
editing I need to do has dropped significantly. This combined with the
increased availability of various 'containers' means these days I'm much
more likely to spin up a local development environment. This also has
the huge advantage of allowing me to work off line or when moving around
where I'm connected to many different networks (a scenario where remote
access control can become a PITA).

I do have the advantage of working on a GNU Emacs system and as all of
the software I work on runs on GNU LInux, setting up a local development
environment is straight-forward (i.e. have access to databases, web
servers, compilers, etc).