[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tramp (2.1.16-pre); Wicked slowdown

From: David Abrahams
Subject: Re: tramp (2.1.16-pre); Wicked slowdown
Date: Thu, 16 Apr 2009 13:05:15 -0400

On Apr 16, 2009, at 8:14 AM, David Abrahams wrote:

On Apr 16, 2009, at 7:26 AM, Michael Albinus wrote:

David Abrahams <address@hidden> writes:

I've noticed TRAMP being /really/ slow on my mac compared to my Linux machine. I am using a more recent snapshot of TRAMP on my mac, so it's
not the same thing of course.  I'd visit / and it
would just sit there for a long time after finding the remote shell
prompt. Spent a few hours debugging this: I turned on "enter debugger on quit" and hit `C-g' a few times during this long process. What the backtraces had in common was the mapcar call in tramp-get-remote- path.
CVS blame seems to indicate that's a recent change.

Unfortunately, the enclosure containing the debug buffer is empty. Could
you, please, ruren the test, and send the debug buffer then?

I have enclosed a bunch of backtraces from where I hit `C-g' while tying to connect. I hope that's helpful. Unfortunately this appears to be a Mac-specific problem, and it's not because of the mapcar call I referred to, because that call doesn't appear in the version of Tramp I was using for bt4-bt6. Aquamacs (which is Emacs 22) with -q exhibits it in addition to my current Emacs 23 snapshot. However, I installed the emacs-snapshot package in an ubuntu VM running on that same hardware and I don't see it there.

Here's something that might be more useful. I used elp to profile tramp. elp.mac is the mac profile; elp.lin is the linux profile. They are accessing the same directory with a CVS TRAMP, but are strikingly different!

Attachment: elp.mac
Description: Binary data

Attachment: elp.lin
Description: Binary data

David Abrahams
BoostPro Computing

reply via email to

[Prev in Thread] Current Thread [Next in Thread]