Improve the use of semantic #11511

bet4it · 2018-10-26T14:55:40Z

This pull request has the same effects as #9354, except that it works globally rather that only works in emacs-lisp-mode, and it's more clear what we have done and what we will lose.
Other language layers are well configured in spacemacs, so I don't think we need semantic to search system databases for them.

As what has been described in #7736 (comment), the setting in #7736 has been overrided by mode local variables and has no effects. So I revert it.

Resolve #1907.

bet4it · 2018-11-04T14:11:26Z

#7736 has disabled many good features semantic can provide, such as:

semantic can help us complete library functions, and it's handy especially for small projects which I don't want to enable large completion tools such as rtags or lsp. The most important thing is that it's a built-in module in emacs.
So I add company-semantic to company-backends.

semantic-idle-summary-mode is conflict with gtags so we should disable it when gtags is enabled.

d12frosted · 2018-11-07T06:41:03Z

layers/+lang/c-c++/packages.el

+  ;; Disable semantic-idle-summary when use gtags layer.
+  ;; Gtags provides more useful information but can not do it properly
+  ;; when this mode is enabled since the minibuffer is cleared all the time."
+  (defun spacemacs//disable-semantic-idle-summary-mode ()


I think this function can be extracted to semantic layer. Since it's being used in multiple layers.

Done. Is it proper to put it into semantic's packages.el?

layers/+lang/python/packages.el

layers/+lang/python/funcs.el

duianto · 2019-05-04T10:39:58Z

And can you add entries for the changes to the changelog.develop.

This reverts commit 9acfcf3.

Solved syl20bnr#10477

duianto · 2019-05-04T15:22:52Z

In the changelog:

Changed the default throttle for =semanticdb-find= routines. Always don't
search system databases to speed up semantic (thanks to bet4it)

I'm guessing that the first word in the second sentence: "Always" was a typo, since the commit just says:

Don't let semanticdb search system databases

I fixed it.

duianto · 2019-05-04T15:32:27Z

Thank you for contributing to Spacemacs!
The PR has been cherry-picked into the develop branch, you can safely delete your branch.

bet4it · 2019-05-04T15:52:00Z

If semantic layer and lsp layer is both enabled, and when semantic want to read the C header file to do completion, lsp will ask us how to deal with the header file ( Import project or not ). Not sure who should have a change to fix it.

duianto · 2019-05-04T15:58:26Z

@yyoncho might be able to help, s/he seems to be part of the lsp-mode team.

IMO better names are the lsp-mode server ids. And just a heads up - at some point, we(lsp-mode team) will provide and option to manage the server priorities using variables so ideally, this configuration should disappear.

source: #11913 (comment)

yyoncho · 2019-05-04T16:09:47Z

@bet4it this seems to be a semantic bug(it runs the major mode hooks for the files that it opens temporary but I need a callstack to confirm that, you may do M-x debug-on-entry and post the callstack). You may workaround the issue by blacklisting the folder.

yyoncho · 2019-05-04T17:38:36Z

@duianto feel free to CC me for lsp related stuff.

bet4it · 2019-05-05T05:05:58Z

@yyoncho This is the callstack:

  lsp--find-root-interactively(#s(lsp-session :folders ("/usr/include/lldb/" "/usr/include/bits/" "/usr/include/llvm/Support/" "/usr/include/c++/8.2.1/" ") :folders-blacklist nil :server-id->folders #<hash-table equal 0/65 0x242a86d> :folder->servers #<hash-table equal 1/65 0x242aa9d> :metadata #<hash-table equal 0/65 0x242aabd>))
  lsp--calculate-root(#s(lsp-session :folders ("/usr/include/lldb/" "/usr/include/bits/" "/usr/include/llvm/Support/" "/usr/include/c++/8.2.1/") :folders-blacklist nil :server-id->folders #<hash-table equal 0/65 0x242a86d> :folder->servers #<hash-table equal 1/65 0x242aa9d> :metadata #<hash-table equal 0/65 0x242aabd>) "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h")
  lsp--try-project-root-workspaces(nil nil)
  lsp()
  (lambda nil (if (memql (function company-lsp) company-backends) (with-no-warnings company-backends) (setq company-backends (cons (function company-lsp) company-backends))) (require (quote ccls)) (remhash (quote clangd) lsp-clients) (lsp))()
  run-hooks(change-major-mode-after-body-hook prog-mode-hook c-mode-common-hook c-mode-hook)
  apply(run-hooks (change-major-mode-after-body-hook prog-mode-hook c-mode-common-hook c-mode-hook))
  run-mode-hooks(c-mode-hook)
  c-mode()
  set-auto-mode-0(c-mode nil)
  set-auto-mode()
  normal-mode(t)
  after-find-file(nil nil)
  find-file-noselect-1(#<buffer stddef.h> "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" t nil "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" (2640408 66310))
  find-file-noselect("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" t nil nil)
  semantic-find-file-noselect("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" t)
  semanticdb-create-table-for-file-not-in-buffer("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h")
  semanticdb-file-table-object("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h")
  semanticdb-find-load-unloaded-default("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h")
  semanticdb-find-load-unloaded("/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h")
  semanticdb-find-table-for-include-default(("stddef.h" include (:system-flag t) (dependency-file "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" :filename "/usr/include/wchar.h") [1157 1176]) #<semanticdb-table semanticdb-table (6 tags), DIRTY>)
  semanticdb-find-table-for-include-c-mode(("stddef.h" include (:system-flag t) (dependency-file "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" :filename "/usr/include/wchar.h") [1157 1176]) #<semanticdb-table semanticdb-table (6 tags), DIRTY>)
  semanticdb-find-table-for-include(("stddef.h" include (:system-flag t) (dependency-file "/usr/lib/gcc/x86_64-pc-linux-gnu/8.2.1/include/stddef.h" :filename "/usr/include/wchar.h") [1157 1176]) #<semanticdb-table semanticdb-table (6 tags), DIRTY>)
  semanticdb-find-translate-path-includes--internal(nil)
  semanticdb-find-translate-path-includes-default(nil)
  semanticdb-find-translate-path-default(nil nil)
  semanticdb-find-translate-path(nil nil)
  semanticdb-find-tags-collector(#f(compiled-function (table tags) #<bytecode 0x1b1961d>) nil nil)
  semanticdb-find-tags-by-class(using)
  semantic-ctxt-scoped-types-c++-mode(nil)
  semantic-ctxt-scoped-types()
  semantic-analyze-scoped-types-default(112)
  semantic-analyze-scoped-types(112)
  semantic-calculate-scope(112)
  semantic-analyze-current-context-default(112)
  semantic-analyze-current-context()
  company-semantic(candidates "")
  company--multi-backend-adapter-candidates((company-semantic company-dabbrev-code company-gtags company-etags company-keywords) "" nil)
  company--multi-backend-adapter((company-semantic company-dabbrev-code company-gtags company-etags company-keywords) candidates "")
  apply(company--multi-backend-adapter (company-semantic company-dabbrev-code company-gtags company-etags company-keywords) (candidates ""))
  company-call-backend-raw(candidates "")
  company--fetch-candidates("")
  company-calculate-candidates("" nil)
  company--begin-new()
  company--perform()
  company-auto-begin()
  company-idle-begin(#<buffer semantic.cpp> #<window 3 on *Backtrace*> 67 112)
  apply(company-idle-begin (#<buffer semantic.cpp> #<window 3 on *Backtrace*> 67 112))
  timer-event-handler([t 23758 27816 814117 nil company-idle-begin (#<buffer semantic.cpp> #<window 3 on *Backtrace*> 67 112) nil 132000])

yyoncho · 2019-05-05T05:22:30Z

@bet4it yes, my guess was correct - semantic should not use find-file-noselect but find-file-literally since it runs the major mode hooks.

agzam · 2019-05-06T17:19:54Z

This PR makes me happy beyond limits. Thank you @bet4it !

captainnova · 2021-11-02T20:57:49Z

yes, my guess was correct - semantic should not use find-file-noselect but find-file-literally since it runs the major mode hooks.

Hi @yyoncho, thanks for finding that. Has it already been reported to the core emacs developers, since semantic is now in builtin? Unfortunately semantic is still causing problems in emacs 27.2.

corngood · 2022-04-13T01:47:03Z

semantic-find-file-noselect doesn't seem to have been changed in emacs. I'm having a similar problem with undo-tree now:

Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!bits!memoryfwd.h.~undo-tree~
Parsing new (LL)...done
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!new.~undo-tree~
Parsing alloc_traits.h (LL)...done
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!ext!alloc_traits.h.~undo-tree~
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!debug!assertions.h.~undo-tree~
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!bits!exception_defines.h.~undo-tree~
Parsing ptr_traits.h (LL)...done
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!bits!ptr_traits.h.~undo-tree~
Parsing type_traits (LL)...done
Wrote /home/david/.undo-tree/.!nix!store!2dv93bbc06c7zg866qid73j3r36zz3jx-gcc-10.3.0!include!c++!10.3.0!type_traits.~undo-tree~

It's also very slow parsing these files. I'm not sure how much that has to do with undo-tree, vs other hooks/modes, but it brought my attention to it.

I'll try to do some investigation soon, but if anyone has any progress/ideas to report, I'm interested.

semantic-find-file-noselect seems like a mess of hacks to blacklist things that don't cooperate, but I'm not sure why it was implemented that way.

bet4it force-pushed the semanticdb branch from 281bcf8 to ec90ca3 Compare October 28, 2018 09:49

bet4it changed the title ~~semantic: Don't let semanticdb search system databases~~ Improve the use of semantic Nov 4, 2018

d12frosted reviewed Nov 7, 2018

View reviewed changes

bet4it force-pushed the semanticdb branch 3 times, most recently from 894188a to b2965a2 Compare November 7, 2018 12:49

bet4it force-pushed the semanticdb branch from b2965a2 to f0077f5 Compare November 17, 2018 01:45

bet4it mentioned this pull request Dec 28, 2018

Add an option to Semantic layer to disable function description in minibuffer? #10477

Closed

bet4it force-pushed the semanticdb branch from f0077f5 to 24ea0ba Compare January 8, 2019 15:46

quicknir mentioned this pull request Apr 11, 2019

semantic in emacs lisp, still slow by default #12156

Closed

duianto reviewed May 4, 2019

View reviewed changes

layers/+lang/python/funcs.el Show resolved Hide resolved

duianto self-assigned this May 4, 2019

bet4it and others added 5 commits May 4, 2019 22:41

Revert "Make Semantic fast"

b7a17d0

This reverts commit 9acfcf3.

semantic: Don't let semanticdb search system databases

06b3ac2

semantic: enable company and disable idle summary when gtags is enabled

816ccb4

semantic: disable idle summary when lsp is enabled

71c9fc9

Solved syl20bnr#10477

Update changelog.develop

f50d867

bet4it force-pushed the semanticdb branch from 24ea0ba to f50d867 Compare May 4, 2019 15:14

duianto closed this May 4, 2019

This was referenced May 4, 2019

emacs-lisp: Speed up semantic #9354

Closed

When enable semantic in emacs-lisp buffer, the completion is very slow #1907

Closed

bet4it deleted the semanticdb branch May 4, 2019 15:43

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improve the use of semantic #11511

Improve the use of semantic #11511

bet4it commented Oct 26, 2018

bet4it commented Nov 4, 2018

d12frosted Nov 7, 2018

bet4it Nov 7, 2018

duianto commented May 4, 2019

duianto commented May 4, 2019 •

edited

Loading

duianto commented May 4, 2019

bet4it commented May 4, 2019

duianto commented May 4, 2019

yyoncho commented May 4, 2019

yyoncho commented May 4, 2019

bet4it commented May 5, 2019

yyoncho commented May 5, 2019

agzam commented May 6, 2019

captainnova commented Nov 2, 2021

corngood commented Apr 13, 2022

Improve the use of semantic #11511

Improve the use of semantic #11511

Conversation

bet4it commented Oct 26, 2018

bet4it commented Nov 4, 2018

d12frosted Nov 7, 2018

Choose a reason for hiding this comment

bet4it Nov 7, 2018

Choose a reason for hiding this comment

duianto commented May 4, 2019

duianto commented May 4, 2019 • edited Loading

duianto commented May 4, 2019

bet4it commented May 4, 2019

duianto commented May 4, 2019

yyoncho commented May 4, 2019

yyoncho commented May 4, 2019

bet4it commented May 5, 2019

yyoncho commented May 5, 2019

agzam commented May 6, 2019

captainnova commented Nov 2, 2021

corngood commented Apr 13, 2022

duianto commented May 4, 2019 •

edited

Loading