CEDET

Hi,

Short term, you can disable semantic-mru-bookmark-mode which overloads 
push-mark.

Better would be to try and debug semantic--tag-similar-types-p in 
semantic-c.el to see where the nil is coming from.  It looks like the 
type of the first blub is nil, so a patch like this might be needed:

=== modified file 'semantic/bovine/semantic-c.el'
*** semantic/bovine/semantic-c.el	2010-10-21 01:42:27 +0000
--- semantic/bovine/semantic-c.el	2011-02-12 00:27:37 +0000
***************
*** 1888,1896 ****
   	    (mapcar
   	    (lambda (tag)
   	      (let ((type (semantic-tag-type tag)))
! 		(unless (stringp type)
! 		  (setq type (semantic-tag-name type)))
! 		(setq type (semantic-analyze-split-name type))
   		(when (stringp type)
   		  (setq type (list type)))
   		type))
--- 1888,1898 ----
   	    (mapcar
   	    (lambda (tag)
   	      (let ((type (semantic-tag-type tag)))
! 		(cond ((semantic-tag-p type)
! 		       (setq type (semantic-tag-name type)))
! 		      ((stringp type)
! 		       (setq type (semantic-analyze-split-name type)))
! 		      )
   		(when (stringp type)
   		  (setq type (list type)))
   		type))


If that or something like it works, it would be great if someone could 
check in a fix.

Thanks
Eric


On 02/10/2011 11:03 AM, Thomas Karolski wrote:
> Hello there,
>
> I get "Wrong type argument: stringp, nil" as a message whenever I try to
> set a mark, or try to yank something into the body of a c++ function
> definition which has similar parts in a python function. I can reproduce
> it like this:
>
> Create a new buffer called blub.cc and enter the following:
> ---------------------------------------
> template<class Y>
> void blub(Y::Complex rho) {
>      rho = 2;
> }
> --------------------------------------
>
> Afterwards, create another buffer called blub.py and enter the following:
> --------------------------------------
> def blub(rho):
>      rho = 1
> --------------------------------------
>
> Let semantic parse both buffers' function definitions (simply stay
> inside the buffer and let semantic-idle parse the contents). Now try to
> yank something into the c++ function body.
> At this point, the error I get is:
> --------------------------------------
> Debugger entered--Lisp error: (wrong-type-argument stringp nil)
>    string-match(":" nil 0)
>    split-string(nil ":")
>    semantic-analyze-split-name-c-mode(nil)
>    semantic-analyze-split-name(nil)
>    #[(tag) <Author: binary code omitted>
>    mapcar(#[(tag) <Author: binary code omitted>
>    semantic--tag-similar-types-p-c-mode(("blub" function (:arguments
> (("rho" variable nil (reparse-symbol function_parameters) #<overlay from
> 10 to 13 in blub.py>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> function (:template (("Y" type (:type "class") nil nil)) :type "void"
> :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> blub.cc>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
>    semantic--tag-similar-types-p(("blub" function (:arguments (("rho"
> variable nil (reparse-symbol function_parameters) #<overlay from 10 to
> 13 in blub.py>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> function (:template (("Y" type (:type "class") nil nil)) :type "void"
> :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> blub.cc>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
>    semantic-tag-similar-p-default(("blub" function (:arguments (("rho"
> variable nil (reparse-symbol function_parameters) #<overlay from 10 to
> 13 in blub.py>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> function (:template (("Y" type (:type "class") nil nil)) :type "void"
> :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> blub.cc>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>) nil)
>    semantic-tag-similar-p(("blub" function (:arguments (("rho" variable
> nil (reparse-symbol function_parameters) #<overlay from 10 to 13 in
> blub.py>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> function (:template (("Y" type (:type "class") nil nil)) :type "void"
> :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> blub.cc>))) (unlink-copy-hook
> (semantic--tag-unlink-copy-secondary-overlays) link-hook
> (semantic--tag-link-secondary-overlays) unlink-hook
> (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
>    #[(sbr point &optional reason) <Author: binary code omitted>
>    push-mark(nil nil t)
>    push-mark-command(nil)
>    set-mark-command(nil)
>    (if (not (wholenump (prefix-numeric-value arg))) (icicle-goto-marker)
> (setq this-command (quote set-mark-command)) (set-mark-command arg))
>    icicle-goto-marker-or-set-mark-command(nil)
>    call-interactively(icicle-goto-marker-or-set-mark-command nil nil)
> --------------------------------------
>
> I'm using svn emacs and bzr cedet.
>
> My guess is, semantic tries to parse the argument for the c++ function
> from the python definition - and makes assumptions on the syntax that
> the definition should have. Since the syntax is different (its python
> code after all), the semantic analyzer throws errors.
> Anyone else who's got this problem?
>
> Kind regards
> Thomas
>
>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
>
>
>
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic

Hello Eric,

many thanks, your patch fixes the problem. No idea where/how to submit a
patch though.

Kind regards,
Thomas

2011/2/12 Eric M. Ludlam <eric@...>

> Hi,
>
> Short term, you can disable semantic-mru-bookmark-mode which overloads
> push-mark.
>
> Better would be to try and debug semantic--tag-similar-types-p in
> semantic-c.el to see where the nil is coming from.  It looks like the
> type of the first blub is nil, so a patch like this might be needed:
>
> === modified file 'semantic/bovine/semantic-c.el'
> *** semantic/bovine/semantic-c.el       2010-10-21 01:42:27 +0000
> --- semantic/bovine/semantic-c.el       2011-02-12 00:27:37 +0000
> ***************
> *** 1888,1896 ****
>            (mapcar
>            (lambda (tag)
>              (let ((type (semantic-tag-type tag)))
> !               (unless (stringp type)
> !                 (setq type (semantic-tag-name type)))
> !               (setq type (semantic-analyze-split-name type))
>                (when (stringp type)
>                  (setq type (list type)))
>                type))
> --- 1888,1898 ----
>            (mapcar
>            (lambda (tag)
>              (let ((type (semantic-tag-type tag)))
> !               (cond ((semantic-tag-p type)
> !                      (setq type (semantic-tag-name type)))
> !                     ((stringp type)
> !                      (setq type (semantic-analyze-split-name type)))
> !                     )
>                (when (stringp type)
>                  (setq type (list type)))
>                type))
>
>
> If that or something like it works, it would be great if someone could
> check in a fix.
>
> Thanks
> Eric
>
>
> On 02/10/2011 11:03 AM, Thomas Karolski wrote:
> > Hello there,
> >
> > I get "Wrong type argument: stringp, nil" as a message whenever I try to
> > set a mark, or try to yank something into the body of a c++ function
> > definition which has similar parts in a python function. I can reproduce
> > it like this:
> >
> > Create a new buffer called blub.cc and enter the following:
> > ---------------------------------------
> > template<class Y>
> > void blub(Y::Complex rho) {
> >      rho = 2;
> > }
> > --------------------------------------
> >
> > Afterwards, create another buffer called blub.py and enter the following:
> > --------------------------------------
> > def blub(rho):
> >      rho = 1
> > --------------------------------------
> >
> > Let semantic parse both buffers' function definitions (simply stay
> > inside the buffer and let semantic-idle parse the contents). Now try to
> > yank something into the c++ function body.
> > At this point, the error I get is:
> > --------------------------------------
> > Debugger entered--Lisp error: (wrong-type-argument stringp nil)
> >    string-match(":" nil 0)
> >    split-string(nil ":")
> >    semantic-analyze-split-name-c-mode(nil)
> >    semantic-analyze-split-name(nil)
> >    #[(tag) <Author: binary code omitted>
> >    mapcar(#[(tag) <Author: binary code omitted>
> >    semantic--tag-similar-types-p-c-mode(("blub" function (:arguments
> > (("rho" variable nil (reparse-symbol function_parameters) #<overlay from
> > 10 to 13 in blub.py>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> > from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> > function (:template (("Y" type (:type "class") nil nil)) :type "void"
> > :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> > nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> > blub.cc>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> > from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
> >    semantic--tag-similar-types-p(("blub" function (:arguments (("rho"
> > variable nil (reparse-symbol function_parameters) #<overlay from 10 to
> > 13 in blub.py>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> > from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> > function (:template (("Y" type (:type "class") nil nil)) :type "void"
> > :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> > nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> > blub.cc>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> > from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
> >    semantic-tag-similar-p-default(("blub" function (:arguments (("rho"
> > variable nil (reparse-symbol function_parameters) #<overlay from 10 to
> > 13 in blub.py>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> > from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> > function (:template (("Y" type (:type "class") nil nil)) :type "void"
> > :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> > nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> > blub.cc>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> > from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>) nil)
> >    semantic-tag-similar-p(("blub" function (:arguments (("rho" variable
> > nil (reparse-symbol function_parameters) #<overlay from 10 to 13 in
> > blub.py>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 16 in blub.py> #<overlay from 1 to 2 in blub.py> #<overlay
> > from 15 to 28 in blub.py>)) #<overlay from 1 to 28 in blub.py>) ("blub"
> > function (:template (("Y" type (:type "class") nil nil)) :type "void"
> > :arguments (("rho" variable (:type ("Y::Complex" type (:type "class")
> > nil nil)) (reparse-symbol arg-sub-list) #<overlay from 29 to 44 in
> > blub.cc>))) (unlink-copy-hook
> > (semantic--tag-unlink-copy-secondary-overlays) link-hook
> > (semantic--tag-link-secondary-overlays) unlink-hook
> > (semantic--tag-unlink-secondary-overlays) secondary-overlays (#<overlay
> > from 1 to 19 in blub.cc> #<overlay from 1 to 2 in blub.cc> #<overlay
> > from 18 to 61 in blub.cc>)) #<overlay from 1 to 61 in blub.cc>))
> >    #[(sbr point &optional reason) <Author: binary code omitted>
> >    push-mark(nil nil t)
> >    push-mark-command(nil)
> >    set-mark-command(nil)
> >    (if (not (wholenump (prefix-numeric-value arg))) (icicle-goto-marker)
> > (setq this-command (quote set-mark-command)) (set-mark-command arg))
> >    icicle-goto-marker-or-set-mark-command(nil)
> >    call-interactively(icicle-goto-marker-or-set-mark-command nil nil)
> > --------------------------------------
> >
> > I'm using svn emacs and bzr cedet.
> >
> > My guess is, semantic tries to parse the argument for the c++ function
> > from the python definition - and makes assumptions on the syntax that
> > the definition should have. Since the syntax is different (its python
> > code after all), the semantic analyzer throws errors.
> > Anyone else who's got this problem?
> >
> > Kind regards
> > Thomas
> >
> >
> >
> >
> ------------------------------------------------------------------------------
> > The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> > Pinpoint memory and threading errors before they happen.
> > Find and fix more than 250 security defects in the development cycle.
> > Locate bottlenecks in serial and parallel code that limit performance.
> > http://p.sf.net/sfu/intel-dev2devfeb
> >
> >
> >
> > _______________________________________________
> > cedet-semantic mailing list
> > cedet-semantic@...
> > https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>

Eric M. Ludlam writes:
> Hi,
>
> Short term, you can disable semantic-mru-bookmark-mode which overloads 
> push-mark.
>
> Better would be to try and debug semantic--tag-similar-types-p in 
> semantic-c.el to see where the nil is coming from.  It looks like the 
> type of the first blub is nil, so a patch like this might be needed:

[...]

Actually, Thomas' assumption that Semantic tries to compare two tags
with the same name but from two different major-modes was
spot-on. What's happening here is that the c-mode override from
semantic--tag-similar-types-p is dealing with a python tag. Hence I
think it would be better to make sure that tags are from the same
major-mode mode before calling semantic-tag-similar-p. I think I can
push a fix tomorrow.

-David

David Engster writes:
> Actually, Thomas' assumption that Semantic tries to compare two tags
> with the same name but from two different major-modes was
> spot-on. What's happening here is that the c-mode override from
> semantic--tag-similar-types-p is dealing with a python tag. Hence I
> think it would be better to make sure that tags are from the same
> major-mode mode before calling semantic-tag-similar-p. I think I can
> push a fix tomorrow.

I've now pushed this patch:

--- semantic/semantic-mru-bookmark.el   2010-04-09 01:54:41 +0000
+++ semantic/semantic-mru-bookmark.el   2011-02-15 22:10:53 +0000
@@ -195,8 +195,10 @@
         (idx 0))
     (when tag
       (while (and (not (ring-empty-p ring)) (< idx (semantic-ring-size ring)))
-       (if (semantic-tag-similar-p (oref (ring-ref ring idx) tag)
-                                   tag)
+       (if (and (eq (semantic-tag-mode tag)
+                    (semantic-tag-mode (oref (ring-ref ring idx) tag)))
+                (semantic-tag-similar-p (oref (ring-ref ring idx) tag)
+                                        tag))
            (ring-remove ring idx))
        (setq idx (1+ idx)))
       ;; Create a new mark


This does the trick for me; at least I don't get an error
anymore. semantic-mru-bookmark would also have to be fixed to deal with
two equally-named tags from different modes. I guess one the tags would
have to be renamed in some meaningful way - maybe simply "foo (C)" and
"foo (Python)". But that's something for another day...

-David