Quite interesting, I can try with an Emacs 28. But I wonder if you found the
same discrepancy with Emacs v27.2 I did?
In the meantime I’ve searched down many rabbit holes and zeroed in on tramp’s
file handlers. I am using:
(require 'tramp-sh)
(defun my/test ()
(interactive)
(progn
(tramp-cleanup-all-connections)
(elp-instrument-list
'(file-remote-p
tramp-file-name-handler
completion-file-name-table
tramp-handle-substitute-in-file-name
substitute-in-file-name
completion--sifn-requote
file-name-completion
file-name-all-completions
completion--file-name-table
file-exists-p
completing-read
read-file-name-internal
read-file-name-default
tramp-handle-file-name-completion
tramp-sh-handle-file-name-all-completions))
(setq unread-command-events
(mapcar 'identity "///ssh:datpi:/home/pi/no\tex\t/do\tma\tce\t\n"))
(call-interactively #'find-file)
(let ((elp-recycle-buffers-p nil))
(elp-results)
(elp-reset-all))))
;; first browse to remote directory (here "pi")
(progn
(with-current-buffer "pi"
(my/test))
(my/test))
For me this implicates file-remote-p and tramp-file-name-handler, which again
exhibit 10x differences between starting in a local or remote buffer. If this
is fixed in Emacs 28 that’s wonderful, but it may be several years before
everyone has migrated.
Results:
Starting in *scratch*:
Tramp-file-name-handler 607 19.486156000
0.0321023986
read-file-name-default 1 6.738599 6.738599
completing-read 1 6.643299 6.643299
substitute-in-file-name 396 5.9316180000
0.0149788333
file-remote-p 1569 5.9054120000
0.0037638062
tramp-handle-substitute-in-file-name 90 5.8898210000
0.0654424555
read-file-name-internal 17 1.420594
0.0835643529
completion--file-name-table 17 1.420297
0.0835468823
completion--sifn-requote 5 1.3667250000
0.2733450000
completion-file-name-table 22 0.767013
0.0348642272
file-name-completion 6 0.733239 0.1222065
tramp-handle-file-name-completion 6 0.7313359999
0.1218893333
file-name-all-completions 6 0.7093809999
0.1182301666
tramp-sh-handle-file-name-all-completions 6 0.7077049999
0.1179508333
file-exists-p 16 0.037264 0.002329
Starting in remote dired buffer “pi”:
tramp-file-name-handler 1429 3.1352620000
0.0021940251
read-file-name-default 1 0.953455 0.953455
completing-read 1 0.94828 0.94828
completion-file-name-table 22 0.7587730000
0.0344896818
file-name-completion 6 0.7366199999
0.1227699999
tramp-handle-file-name-completion 6 0.734705
0.1224508333
file-name-all-completions 6 0.7178499999
0.1196416666
tramp-sh-handle-file-name-all-completions 6 0.7160390000
0.1193398333
substitute-in-file-name 482 0.3317939999
0.0006883692
tramp-handle-substitute-in-file-name 167 0.2854299999
0.0017091616
file-remote-p 4088 0.1552450000
3.797...e-05
read-file-name-internal 17 0.08041 0.00473
completion--file-name-table 17 0.080114
0.0047125882
file-exists-p 101 0.0680669999
0.0006739306
completion--sifn-requote 5 0.019765 0.003953
> On Sep 19, 2021, at 12:55 PM, Michael Albinus <[email protected]> wrote:
>
> JD Smith <[email protected]> writes:
>
> Hi,
>
>> Thanks very much for this much improved testing code (which also saves
>> the tedium of slowly retyping!). I altered it slightly to include a
>> few more function, and target my local wired server:
>>
>> (progn
>> (require 'tramp-sh)
>> (tramp-cleanup-all-connections)
>> (elp-instrument-list '(file-exists-p completing-read
>> read-file-name-internal read-file-name-default
>> tramp-handle-file-name-completion
>> tramp-sh-handle-file-name-all-completions))
>> (setq unread-command-events (mapcar 'identity
>> "///ssh:datpi:/home/pi/no\tex\t/do\tma\tce\t\n"))
>> (call-interactively #'find-file)
>> (elp-results)
>> (elp-reset-all))
>
> I've applied the same, except the remote file name.
>
> My timing results (emacs -Q, v28.0.50):
>
> From *scratch*:
>
> read-file-name-default 1 9.064781548
> 9.064781548
> completing-read 1 9.057834631
> 9.057834631
> tramp-handle-file-name-completion 3 8.331385864
> 2.7771286213
> tramp-sh-handle-file-name-all-completions 4 7.290909892
> 1.822727473
> file-exists-p 45 1.7296832099
> 0.0384374046
> read-file-name-internal 13 0.4099991600
> 0.0315383969
>
> From remote dired:
>
> read-file-name-default 1 8.23891833
> 8.23891833
> completing-read 1 8.228419034
> 8.228419034
> tramp-handle-file-name-completion 3 7.4998795480
> 2.4999598493
> tramp-sh-handle-file-name-all-completions 4 6.463144493
> 1.6157861232
> file-exists-p 45 1.6901479139
> 0.0375588425
> read-file-name-internal 13 0.3952138859
> 0.0304010681
>
>> So really only very little difference inside the tramp completion
>> functions; they seem to be doing the same job independent of the
>> starting point. But the completing-read itself, which is where
>> find-file is spending its time, is taking 8x longer! And
>> read-file-name-internal is 35x slower. Do you find this as well?
>> I’ll dig a bit deeper and see if I can narrow it down further.
>
> With Emacs 28, the differences aren't so relevant. The absolute time is
> much higher than in your case, but I have taken "savannah" as test host
> which is "far from me".
>
> Do you have a chance to compile Emacs from git, and run the test?
>
> Best regards, Michael.
