bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 368 bytes --]


24.3.50; cperl-mode + which-function: Inaccurate reporting (corner case)

See the attached screenshot.  Note the following

1. Cursor position in the top window. 

   In the bottom window, the cursor is well within the toplevel block of
   sub.

2. which-function field of the modeline.

Hint: Note the presence/absence of `=' (most likely from the =cut line above)


[-- Attachment #2: cperl-which-function-is-inaccurate.png --]
[-- Type: image/png, Size: 85681 bytes --]

[-- Attachment #3: Type: text/plain, Size: 298 bytes --]




In GNU Emacs 24.3.50.4 (i686-pc-linux-gnu, GTK+ Version 2.20.1)
 of 2013-11-09 on debian-6.05
Bzr revision: 115051 jan.h.d@swipnet.se-20131109111953-cysfi19meuidy96q
Windowing system distributor `The X.Org Foundation', version 11.0.10707000
System Description:	Debian GNU/Linux 6.0.5 (squeeze)

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> Hint: Note the presence/absence of `=' (most likely from the =cut line
> above)

=cut directives are supposed to be followed by at least one empty line.
Please see if you can reproduce this using correct pod syntax.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 254 bytes --]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> =cut directives are supposed to be followed by at least one empty
> line.

I expected this.

> Please see if you can reproduce this using correct pod syntax.

Yes, I can.  See for yourself.  Here I go ..


[-- Attachment #2: cperl-which-function.png --]
[-- Type: image/png, Size: 25350 bytes --]

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 226 bytes --]

Jambunathan K <kjambunathan@gmail.com> writes:

>> Please see if you can reproduce this using correct pod syntax.
>
> Yes, I can.  See for yourself.  Here I go ..

I attach the screenshot again.  (This one shows the cursor.)


[-- Attachment #2: cperl-which-function.png --]
[-- Type: image/png, Size: 25814 bytes --]

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> I attach the screenshot again.  (This one shows the cursor.)

Thanks, but after looking into this, I don't think it's a bug.
which-function-mode displays either the definition containing point or,
if there is none, the definition before and closest to point.  In
cperl's case, "definition" intentionally includes pod =head directives.
All you're seeing is the previous defintion's name being correctly
displayed in the mode line.

Perhaps a cperl variable to disable imenu indexing of POD directives is
what you want.  I might be able to implement something like that if it's
not too hard.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> Jambunathan K <kjambunathan@gmail.com> writes:
>
>> I attach the screenshot again.  (This one shows the cursor.)
>
> Thanks, but after looking into this, I don't think it's a bug.

perl-mode behaves differently and correctly.  

> In cperl's case, "definition" intentionally includes pod =head
> directives.  All you're seeing is the previous defintion's name being
> correctly displayed in the mode line.

Is it possible to show the sub name when within a function and show the
head1 or head2 string when within a perldoc string.  (I am new to perl
so I can talk only fuzzily about the language or it's features)

My gut feeling is that the way cperl mode treats POD directives also
limits it from doing outlining.

See the screenshot in this thread.

http://lists.gnu.org/archive/html/bug-gnu-emacs/2013-11/msg00355.html

I would really like to get outlining work when in cperl-mode.

> Perhaps a cperl variable to disable imenu indexing of POD directives is
> what you want.  I might be able to implement something like that if it's
> not too hard.

I think anything that doesn't show "= " would do for me.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

[-- Attachment #1: Type: text/plain, Size: 1465 bytes --]

Jambunathan K <kjambunathan@gmail.com> writes:

> perl-mode behaves differently and correctly.  

I think you've misunderstood.  cperl-mode does exactly what it's
supposed to do, which is more than what perl-mode tries to do.

> Is it possible to show the sub name when within a function and show
> the head1 or head2 string when within a perldoc string.  (I am new to
> perl so I can talk only fuzzily about the language or it's features)

It's not really feasible, though it has nothing to do with Perl per se.
Like I said before, the "function" name displayed by which-function-mode
is the function containing point or, if none, the nearest function
before point.  In cperl-mode pod =head directives count for this purpose
as "functions", and regular subroutines begin at the position following
their name.

> My gut feeling is that the way cperl mode treats POD directives also
> limits it from doing outlining.

After a brief skim of the cperl source, I had the opposite impression.

> I would really like to get outlining work when in cperl-mode.

I'll take a look at that later.  With luck, it'll be as easy as this
one.

> I think anything that doesn't show "= " would do for me.

Again, the "= " means that the "current function" (according to the
above definition) is a pod =head directive.

Try the attached patch.  Make sure you set cperl-imenu-index-pod to nil
before invoking cperl-mode.  If it does what you want, we can ask for it
to be installed.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: cperl-mode.patch --]
[-- Type: text/x-diff, Size: 5438 bytes --]

# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: nbtrap@nbtrap.com-20131110162047-2pqh1c9up17hfzkd
# target_branch: file:///home/nathan/opt/etc/bzr-repos/emacs/trunk/
# testament_sha1: fd4a0d8284f9bbda92ca8a31dc190cfa3b8adb78
# timestamp: 2013-11-10 11:41:32 -0500
# base_revision_id: jan.h.d@swipnet.se-20131110121803-bv908hul7t8ji6c8
# 
# Begin patch
=== modified file 'lisp/ChangeLog'
--- lisp/ChangeLog	2013-11-10 08:35:04 +0000
+++ lisp/ChangeLog	2013-11-10 16:20:47 +0000
@@ -1,3 +1,10 @@
+2013-11-10  Nathan Trapuzzano  <nbtrap@nbtrap.com>
+
+	* progmodes/cperl-mode.el (cperl-imenu-index-pod)
+	(cperl-imenu--create-perl-index): Define and implement
+	`cperl-imenu-index-pod' to optionally disable cperl's indexing of
+	POD directives for imenu.  (Bug#15848)
+
 2013-11-10  Michael Albinus  <michael.albinus@gmx.de>
 
 	* net/tramp.el (tramp-methods):

=== modified file 'lisp/progmodes/cperl-mode.el'
--- lisp/progmodes/cperl-mode.el	2013-11-04 03:44:23 +0000
+++ lisp/progmodes/cperl-mode.el	2013-11-10 16:15:36 +0000
@@ -511,6 +511,13 @@
   :type 'boolean
   :group 'cperl-help-system)
 
+(defcustom cperl-imenu-index-pod t
+  "*Whether POD `=head' directives should be indexed by imenu.
+Non-nil means `which-function-mode' won't count POD directives as
+top-level definitions."
+  :type 'boolean
+  :group 'cperl-help-system)
+
 (defcustom cperl-max-help-size 66
   "*Non-nil means shrink-wrapping of info-buffer allowed up to these percents."
   :type '(choice integer (const nil))
@@ -844,7 +851,8 @@
                 interactive sweep over the buffer);
         d) Has support for imenu, including:
                 1) Separate unordered list of \"interesting places\";
-                2) Separate TOC of POD sections;
+                2) Separate TOC of POD sections by default
+                   (cf. `cperl-imenu-index-pod');
                 3) Separate list of packages;
                 4) Hierarchical view of methods in (sub)packages;
                 5) and functions (by the full name - with package);
@@ -5551,7 +5559,7 @@
 	      (push index index-alist))
 	    (if meth (push index index-meth-alist))
 	    (push index index-unsorted-alist)))
-	 ((match-beginning 16)		; POD section
+	 ((and cperl-imenu-index-pod (match-beginning 16))	; POD section
 	  (setq name (buffer-substring (match-beginning 17) (match-end 17))
 		marker (make-marker))
 	  (set-marker marker (match-beginning 17))

# Begin bundle
IyBCYXphYXIgcmV2aXNpb24gYnVuZGxlIHY0CiMKQlpoOTFBWSZTWXtrJDAABhbfgAA02PP//38F
zID////wYAn9sU+alUAKAhqtgrKoADQKhJITQRpNPQGJkyRoJhqABoMgBJTQJpkaBNEmJqGQ9I0B
oAAAOYBMBMjACMTEwmEwQ0xNMCJlNGpoYmRoGJ6g0xAA0yaaAAOYBMBMjACMTEwmEwQ0xNMAqSRG
IAjQNSeNTFMNAU09R6h6mmT0T1BMuINEW/xgOa1cQhUAEQAaQg6P1BQQyhQDx0mSkUESuitzKMnM
BUcsMNX+BogwchDB5UJxbIJA/F79D5E8J968ywCAZWWl0kD5H4sf96NuoPop0yGYTMzDNs+wiuAS
gyq0vfq8pXQUE5RsljIoM9q4DojPcnbM4agbrbzKcxfhfXoRyIvtuHDFYzEB7vzlu7/f38SoeZDX
EGa0P4gKhZTUqHkFPdlhAPbRgMicePAopl/rlVjjbj18FVoW9PYJubsm9OiHRm4e0RCaaeEjIRTF
GkaCzYDjl49r+4Y4zU29Ltnfu3ZJf3ptcBpHuOCWfD0z79ZvkHJMjejTzI9tp+nCFzZ6D3FiZI2r
1xmLPInht60eZjlN/Fb2+/pbvXzzzuWARur6CN1ws/INb17/Xh0xZmZmZm6qNEE3dkuLsiz1VmJw
LFuyhF10gy6Nt2jKY493z1nAFXskMLzxhkboH5np75XK7LRQBO1owldSQT+gri6z4PxB0g95Bq2Z
mcYdjd6DRNF6DSLGuXSXrzpqQZVXcOj0l2OaSFhGkmvaIgCxwCgZ7KeImjzPtTGPTsGeBQQiYkO+
9MVbzoJbW77dMefpOBlC+cX9cjov1kcKhk4XPj1gYkX70g/MUC9gxWm/twBQnACwVtyXzfwLk48M
h3e4YzER6Xg/3OEeEOAXyOAReIKwsN8ezBERrHx6dE+Qn0hlPiqDGfidjUxInZ9kKnGwn5K+CfQd
BRaU09doYYUTjugAi2RIeJGAjPY1XpVTaxbHOEJpb6oMpxPqeiJXV5wDkdDeWVMdDhnckukJktCO
HAiWpCbOqtoGYKjHDpyrQE8kTjB0PKCht1jCwLPNoZOIRICKYcwNHWk+IwrNnnW7DGzwxz8gm4HO
Th4EsANCsy1TXLckSMqEsxyDwEQfxSfDDlPNli41TTRWbXxFogpbS/QjtVjaT88KWHPmVLZAqQFL
UIuMjK2OWrZd/BNrJuh60dxMeG7gika9NpG1UHMPL3b1oHdXSwaT79A36OAyn3+i/cttbEMRjHmo
+o1BGW8e4iM5NpNKkDHrPuyLkquvVA+5huM4kZ0J1LupMffANw0BZ5aXvlmvZU1N3EAVBUsu4vhC
7pGciL4CJQelY0xezb23wxHNqa2+KFwc9HGdTNsnHE7Zd0AnuDBDQrrXDbZw6bMwbD2yXl5plsSu
C9C/xdBpAuti/K2IN02IE9iXETkwtM6cP7Wie+HYJJJwb20sY3dqo9RDC+a06gyJ8w02L5Y2C03H
Oc1si+YFroZ5JeXkowKzDGJjnKMWj3+QujLicxaqb7Sj8qqqqqP5d3I+pY9/r8xGE1a5NMSqlfb9
P21crCsKwqrWMH3/6L44VD/GWJ0cCpaJv+dwwmNBhsqslBX3DghJVDIIvccduFHaaJ6ZSo5j/T05
cSmVmV4sWrVzb2hpO6Jl8/h2R2F7KPpMDlxXnCIk8UIR4d+pCPgInOIozfpDws5b7G2XbZyU3v8I
Xd835mdybnwxfHDOj6uR3BFGE6yZT2oYGdvUXPX8PBKF2Ojn8E1E55MBeqxMqMVGpzXjrnI3T048
unEoBjw69Tw59ACNwWK7K5eIOcUUJkyv10jJbj1nuxIEibZAki1Q9J3Prw9QS3NvJy52FYW186DD
CkFhAMSkaXPQCvUDXEwzHnJsJWEbdk2S69k1j9swc6Sfk1N2AXIKEtATaAUhwoV6KDpyVLepXCIh
XQXd1xgS0fOuUdE5nyTQagjfrsyyYlY4vvXFvlD2BtUDgGCXAOQsj2r2wMBsye5bqRF7yGeILIa+
ISZwVYbdMWehhU5hienyT1ulQWjy7VQy0acdW4y9We1xuvz3f8KHZOzjxc/FQ3k8bEukn8YNPbke
XHu8STVGBmNA7JobGsIpLICiHSadXDBzi02kDfRuJB5N0ApozSby0hIWzqakWWamqYsL+oeJLdA/
c8TtL467QcQOohYKQzsNyQxUg1hBkJk+OLg9KLa0upZvsl3gFZo/dZ+7Hfmfy6k/DJth5zVDqSXG
McyPdKR+eHWt+d5sPwSi2J3izvJfI7TMKnyiXSLF82Z/Z8l7t+fs8TiZHBGzqKUSjrm+Scq9wbb9
zoPVin/tBhdGuVHGe9Ht8OVHxTticp5qnKkqCvZQbA2FjzlEvs9GBPY+Lxcvpj3mRbWmku1Rouoh
uqM0emjA8mdsknWUK9BibU0ptM2vQWGL0IzX41fUWRUo41DUchH9nOrRFLKtFLKsUsq3jG7I1qmm
STfpOiXRedKM6OY76VmYVnjkSl00YKiLZrElRPLCOZp1fKGolyOoL+DJHV+KZZjbs338ZBVQx5lJ
/pUTvNdpWGjFRr5i6JUNtiavNrYd6OmszfvEazMlhHPr6essUTOPu1RM9EYhA2UYiNK/gLwVahjk
j0AsBVQZkIpzmSqJ0HAI/aM1t+CaccxGk8C8X+NCaYe7HRrM9HhgYaUZ7Es2r1V9qiyuMOzox/uj
XeZz6Tc5p10pUqt5cwCxlPjUqlUqnpRf56B+WxoSeuat/fscLpwUaJ5y8uPkOBdLFt5iMjdnjJR1
eBtHbPRWCk8Ec+ma3cqJuJyCKmmiZGlhd1kpJ1o7P95HdNT1hv6WPMRBe8LW3Q15RZUoCh1JgmQw
B+hlOfElPqjmJhhI7OpG5FwY8b4socIO9W6eaKo2FiHgj/xdyRThQkHtrJDA

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Stefan Monnier]

> Perhaps a cperl variable to disable imenu indexing of POD directives is
> what you want.  I might be able to implement something like that if it's
> not too hard.

If you do, please try and share the code between perl-mode and
cperl-mode (we could (require 'perl-mode) at the beginning of
cperl-mode for that).


        Stefan

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> I think you've misunderstood.  cperl-mode does exactly what it's
> supposed to do, which is more than what perl-mode tries to do.

Actually, perl-mode _does_ index pod directives for imenu.  My mistake.
The difference is where each mode considers one "function" to end and
another to begin.  Moreover, cperl-mode puts "= " in front in order to
distinguish the "function" as a pod directive.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> If you do, please try and share the code between perl-mode and
> cperl-mode (we could (require 'perl-mode) at the beginning of
> cperl-mode for that).

The two modes hook into imenu by different interfaces, so this is not
really feasible.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Leo Liu]

On 2013-11-11 02:04 +0800, Stefan Monnier wrote:
> If you do, please try and share the code between perl-mode and
> cperl-mode (we could (require 'perl-mode) at the beginning of
> cperl-mode for that).

Why not get merge perl-mode into cperl-mode so that we can focus on
improving cperl-mode? There are buglets here and there in cperl-mode.

Having two modes provide no gains for anybody. Split development effort
and an extra decision for users to make.

Leo

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Stefan Monnier]

> Why not get merge perl-mode into cperl-mode so that we can focus on
> improving cperl-mode? There are buglets here and there in cperl-mode.

I find cperl-mode's code completely impenetrable and riddled with
irrelevant version checks.  And I don't like its behavior very much.

I do think, the two code bases should get closer, which is why
I encouraged Nathan in this direction.


        Stefan

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

[-- Attachment #1: Type: text/plain, Size: 382 bytes --]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> Try the attached patch.  

Thanks.

> Make sure you set cperl-imenu-index-pod to nil before invoking
> cperl-mode.  

Done.  See the screenshot.  

> If it does what you want

Pay particular attention to the hollow rectangle which tells you where
the cursor is.  Mode line is reporting the previous defun and NOT the
current defun.


[-- Attachment #2: which-function-with-nathans-patch.png --]
[-- Type: image/png, Size: 30239 bytes --]

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> Mode line is reporting the previous defun and NOT the current defun.

I've already said that cperl-mode considers a "function" to begin at the
position immediately _following_ the subroutine's name.  Perhaps I
misunderstood you, but what that patch did was add an option for telling
cperl-mode not to count POD =head directives as "functions".

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Leo Liu <sdl.web@gmail.com> writes:

> Having two modes provide no gains for anybody. Split development
> effort and an extra decision for users to make.

AFAIK, in terms of features, cperl-mode offers everything that perl-mode
offers plus a lot more.  Any effort to merge the two would have the
effect of replacing the latter with the former.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> I find cperl-mode's code completely impenetrable and riddled with
> irrelevant version checks.  And I don't like its behavior very much.

Changing some of the defaults helped me warm up to cperl-mode.  By
default, it tries to do way too much, especially with regard to
font-lock.  That said, the regex beutifying is quite impressive.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> I've already said that cperl-mode considers a "function" to begin at
> the position immediately _following_ the subroutine's name.

I am not retracting my position.  The behaviour - with and without your
patch - is incorrect.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

>> I've already said that cperl-mode considers a "function" to begin at
>> the position immediately _following_ the subroutine's name.
>
> I am not retracting my position.  The behaviour - with and without your
> patch - is incorrect.

If that were true, I might fix it; but as it is, the determination of
where the text of one definition ends and another begins seems to admit
of some legitimate flexibility.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

> If that were true, I might fix it; but as it is, the determination of
> where the text of one definition ends and another begins seems to admit
> of some legitimate flexibility.

From my standpoint, the sub has ended when it's '}' is seen.  And when I
am right on the sub's name it better be sub name. 

Instead of bickering :-), we should ask for a third party vote or a
veto. 

(Thanks for the patch, I appreciate it.)

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Jambunathan K <kjambunathan@gmail.com> writes:

> Instead of bickering :-), we should ask for a third party vote or a
> veto. 

(Looks like I am having a bad today.  Sorry for multiple posts.)

Just move around in an .el file and you know what I am talking about.  I
am just asking for the standard behaviour. 

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> Instead of bickering :-), we should ask for a third party vote or a
> veto.

Tell you what: if you can convince the maintainers to accept such a
change, I'll write the patch.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> Jambunathan K <kjambunathan@gmail.com> writes:
>
>> Instead of bickering :-), we should ask for a third party vote or a
>> veto.
>
> Tell you what: if you can convince the maintainers to accept such a
> change, I'll write the patch.

I am convinced that maintainers are unlikely to have any other opinion.

Did you look at the behaviour of `which-function-mode' on Emacs Lisp
files.  You can also look at the behaviour in C source files.

In my specific case, the cursor is "right on" the function name.  So the
the displayed value should be that function name and not the previous
function name.

On the buffer zone between two defuns, which-function-mode MAY return
nil.  But this case is a "don't care" for me.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

[-- Attachment #1: Type: text/plain, Size: 342 bytes --]

Jambunathan K <kjambunathan@gmail.com> writes:

Try this patch.  It includes the patch for the variable I added earlier.

Also, don't try to apply this on top of the patch I sent for the outline
problem.  You'll get conflicts.  I'll wait till that patch gets
installed and then resolve the conflicts with this patch and send an
updated one.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: cperl-mode-imenu-patch.patch --]
[-- Type: text/x-diff, Size: 12159 bytes --]

# Bazaar merge directive format 2 (Bazaar 0.90)
# revision_id: nbtrap@nbtrap.com-20131112140320-if3f086us590vp5c
# target_branch: file:///home/nathan/opt/etc/bzr-repos/emacs/trunk/
# testament_sha1: 6e81494670e3d66dc1c7cb0abd95be7a9831a403
# timestamp: 2013-11-12 09:05:04 -0500
# base_revision_id: jan.h.d@swipnet.se-20131110121803-bv908hul7t8ji6c8
# 
# Begin patch
=== modified file 'lisp/ChangeLog'
--- lisp/ChangeLog	2013-11-10 08:35:04 +0000
+++ lisp/ChangeLog	2013-11-12 14:03:20 +0000
@@ -1,3 +1,17 @@
+2013-11-12  Nathan Trapuzzano  <nbtrap@nbtrap.com>
+
+	* progmodes/cperl-mode.el (cperl-imenu--function-name-regexp-perl)
+	(cperl-outline-level, cperl-imeny--create-perl-index): Index imenu
+	definitions with the positions at which they begin instead of
+	their end positions.  (Bug#15848)
+
+2013-11-10  Nathan Trapuzzano  <nbtrap@nbtrap.com>
+
+	* progmodes/cperl-mode.el (cperl-imenu-index-pod)
+	(cperl-imenu--create-perl-index): Define and implement
+	`cperl-imenu-index-pod' to optionally disable cperl's indexing of
+	POD directives for imenu.  (Bug#15848)
+
 2013-11-10  Michael Albinus  <michael.albinus@gmx.de>
 
 	* net/tramp.el (tramp-methods):

=== modified file 'lisp/progmodes/cperl-mode.el'
--- lisp/progmodes/cperl-mode.el	2013-11-04 03:44:23 +0000
+++ lisp/progmodes/cperl-mode.el	2013-11-12 13:26:52 +0000
@@ -511,6 +511,13 @@
   :type 'boolean
   :group 'cperl-help-system)
 
+(defcustom cperl-imenu-index-pod t
+  "*Whether POD `=head' directives should be indexed by imenu.
+Non-nil means `which-function-mode' won't count POD directives as
+top-level definitions."
+  :type 'boolean
+  :group 'cperl-help-system)
+
 (defcustom cperl-max-help-size 66
   "*Non-nil means shrink-wrapping of info-buffer allowed up to these percents."
   :type '(choice integer (const nil))
@@ -844,7 +851,8 @@
                 interactive sweep over the buffer);
         d) Has support for imenu, including:
                 1) Separate unordered list of \"interesting places\";
-                2) Separate TOC of POD sections;
+                2) Separate TOC of POD sections by default
+                   (cf. `cperl-imenu-index-pod');
                 3) Separate list of packages;
                 4) Hierarchical view of methods in (sub)packages;
                 5) and functions (by the full name - with package);
@@ -1439,7 +1447,7 @@
 
 ;;; Details of groups in this are used in `cperl-imenu--create-perl-index'
 ;;;  and `cperl-outline-level'.
-;;;; Was: 2=sub|package; now 2=package-group, 5=package-name 8=sub-name (+3)
+;;;; Was: 2=sub|package; now 2=package-group, 5=package-name 9=sub-name (+4)
 (defvar cperl-imenu--function-name-regexp-perl
   (concat
    "^\\("				; 1 = all
@@ -1448,12 +1456,13 @@
 	    cperl-white-and-comment-rex ; 4 = pre-package-name
 	       "\\([a-zA-Z_0-9:']+\\)\\)?\\)" ; 5 = package-name
        "\\|"
-          "[ \t]*sub"
-	  (cperl-after-sub-regexp 'named nil) ; 8=name 11=proto 14=attr-start
-	  cperl-maybe-white-and-comment-rex	; 15=pre-block
+          "\\([ \t]*sub"
+	  (cperl-after-sub-regexp 'named nil) ; 9=name 12=proto 15=attr-start
+	  cperl-maybe-white-and-comment-rex	; 16=pre-block
+          "\\)"
    "\\|"
-     "=head\\([1-4]\\)[ \t]+"		; 16=level
-     "\\([^\n]+\\)$"			; 17=text
+     "=head\\([1-4]\\)[ \t]+"		; 17=level
+     "\\([^\n]+\\)$"			; 18=text
    "\\)"))
 
 (defvar cperl-outline-regexp
@@ -5486,17 +5495,16 @@
       (while (re-search-forward
 	      (or regexp cperl-imenu--function-name-regexp-perl)
 	      nil t)
-	;; 2=package-group, 5=package-name 8=sub-name
+	;; 2=package-group, 5=package-name 9=sub-name
 	(cond
 	 ((and				; Skip some noise if building tags
 	   (match-beginning 5)		; package name
 	   ;;(eq (char-after (match-beginning 2)) ?p) ; package
-	   (not (save-match-data
-		  (looking-at "[ \t\n]*;")))) ; Plain text word 'package'
+	   (not (looking-at-p "[ \t\n]*;"))) ; Plain text word 'package'
 	  nil)
 	 ((and
 	   (or (match-beginning 2)
-	       (match-beginning 8))		; package or sub
+	       (match-beginning 9))		; package or sub
 	   ;; Skip if quoted (will not skip multi-line ''-strings :-():
 	   (null (get-text-property (match-beginning 1) 'syntax-table))
 	   (null (get-text-property (match-beginning 1) 'syntax-type))
@@ -5535,10 +5543,13 @@
 	  (if (and is-proto (not is-pack)) nil
 	    (or is-pack
 		(setq name
-		      (buffer-substring (match-beginning 8) (match-end 8)))
+		      (buffer-substring (match-beginning 9) (match-end 9)))
 		(set-text-properties 0 (length name) nil name))
 	    (setq marker (make-marker))
-	    (set-marker marker (match-end (if is-pack 2 8)))
+	    (set-marker marker (save-excursion
+                                 (goto-char (match-beginning (if is-pack 2 6)))
+                                 (skip-chars-forward " \t")
+                                 (point)))
 	    (cond (is-pack nil)
 		  ((string-match "[:']" name)
 		   (setq meth t))
@@ -5551,13 +5562,13 @@
 	      (push index index-alist))
 	    (if meth (push index index-meth-alist))
 	    (push index index-unsorted-alist)))
-	 ((match-beginning 16)		; POD section
-	  (setq name (buffer-substring (match-beginning 17) (match-end 17))
+	 ((and cperl-imenu-index-pod (match-beginning 17))	; POD section
+	  (setq name (buffer-substring (match-beginning 18) (match-end 18))
 		marker (make-marker))
-	  (set-marker marker (match-beginning 17))
+	  (set-marker marker (- (match-beginning 17) 5))
 	  (set-text-properties 0 (length name) nil name)
 	  (setq name (concat (make-string
-			      (* 3 (- (char-after (match-beginning 16)) ?1))
+			      (* 3 (- (char-after (match-beginning 17)) ?1))
 			      ?\ )
 			     name)
 		index (cons name marker))
@@ -5624,11 +5635,11 @@
 (defun cperl-outline-level ()
   (looking-at outline-regexp)
   (cond ((not (match-beginning 1)) 0)	; beginning-of-file
-;;;; 2=package-group, 5=package-name 8=sub-name 16=head-level
+;;;; 2=package-group, 5=package-name 9=sub-name 17=head-level
 	((match-beginning 2) 0)		; package
-	((match-beginning 8) 1)		; sub
-	((match-beginning 16)
-	 (- (char-after (match-beginning 16)) ?0)) ; headN ==> N
+	((match-beginning 9) 1)		; sub
+	((match-beginning 17)
+	 (- (char-after (match-beginning 17)) ?0)) ; headN ==> N
 	(t 5)))				; should not happen
 
 \f

# Begin bundle
IyBCYXphYXIgcmV2aXNpb24gYnVuZGxlIHY0CiMKQlpoOTFBWSZTWeeMVtkAD/D/gHk0gAL8////
/yXOj/////RgFD1b57t7R6N7ffN6x9WzU1kKA6+nVABUOfTc5vrbm93vbW+766PbdX1qWwkDQbbI
GglCEaammmk8U9GRT1T/SU2niU9T1D1Nqeo00eoeoAAaekBKTIAJo0UyRpPVM1PJM1NNAAGg0NAA
AAlTSbKNRPSMm1HqNNANBoAAAAAAAASFImptNRoxE9RnpTTZQAT1PED1HqjAgxMjE0YIqKTxPUJ6
p4noo9NPSYUyeQaIMg00AAAANBFITQTBBDICMTII1Mj0TQaYmgAAYhqLj2DSw47ZJmKEGAfaFBlE
XnJ1oWM/r6Tkw+PusSIVT+xusmCE+dml08MxEQi1VLxB1X4ZVxB6h8dfS/vv7zyR+4n+maspW38I
/eT4HyLKdCztD6B4ZLdHEw+eof1by7drOIMo8X1QrVwoiF2wzlGZAWtG+4dapZCiyEyvLYGLi3WL
qWGuyomAlltsXN1al1KITEs7UPuPFRfAsIYRkyhkZk5wzdaITyqzrQh8UZDnkFJEvvolIbg/DIYt
erI1REeKiUm1RrT1BGgDdAYJtsSbbG0xg2Jj4v9gBy4WZwJoJj2bqDSLYpvoTOs47jvhL5S4SzZ9
0IlADwGAm8uXqo766zo02m+cpe1pUlCvWZWU6HY1bglbKNswO12uOvWsVdIray22ZqTFlVRUBVSc
1ybErRS8Q4OHWnJjcZ8RpHUsc20Ew3H7J4kpvR87ZuQrDJ2Kaj21jnGmRO48pr1yuml5N1GlDkQi
VzV6BhCuZXK0/FhDX2yjY6A812pKcpVn5tWcqKArPbOIhbM1/a5LO7YtvOiIPj7iKs8EOkuSjDPE
y9AQuntJ3RshVM6lRudPDDu3bff4dfSxx4r10kmOsDcZAP6keqCqIUwXHs7vTUd0qnRjXtXT8Fuo
33eIgnfLm01VczNZZQSPVmVwII4gYZT5071HC1zoYavOOO8+ZkT+0+wlP1irY9h6PVUlY0n6JU26
pPUzzDpVm6bOTLk4Tl15krgkREgXS5GZQu4IysoZsyT2oRTPmIzI0bzUNOenA9ky3q9ckscHb8q8
4hd2CuBDpEPAQiAMFB2gadIVqqYhJCbgmMS5sqYxO8U2qXmRjV+g1PT+sr2iQFpvJFJuHuE9Amoz
ANym5DZiXWhFbiBPiZs9gkHeK2VOWZfcd6mEq4iGIhcE8NCa6zepCFMy3eljnuvqxWxX4LcZ0QQT
UnBjELOuBbvGI582Q27+aTR1ee6/ckkkkkkkkkkZmZmbDXuQKNAgalBOComCgAy4YgTCH3ziFBs6
90XYExPTeWGBXtpeK6TqpZvKhUtExkCDWlfskAl8GnEszElj4CVcMdBJ3Scl6SwSRXzztpdYK6PB
SQm4mXxGu01tuzOoTG+lZTIUsLHEDmGq3LWHiU5TcCSIiZWWym8hnDCGEA6M0S6ktxpMtraUeQkG
UbrtS/IsxEi4h6RDKyvo1iEDRXzEFE7gQ294+kTgAkPBGcgtD0iqIVKIVUUUMz8HnhCGOIhEmDG9
gEMxoyP5Uso6HUeAJ08/mJWXNU1MAUsBBniKzVHnjMSBXDWBMROXRsuxNt6o6tCgJYAbPf5WdnG3
nxfL58z2ahLyEOLGZmQSVBBUEBIEuASELN8U3IGRHMJjaNNBWtd4dErxMgn4faIdxDNK9YpxrIhc
ReOx9BoQ8Y4o1HjwQXsohs4Qx4zlAsRfiCFYKP0IQOXmBCyUOrJcZ9ZbnERKKSIIaBDRSvwmznwS
6VcTAQfUl+tjbDgco5t5HvQg51ZbuMYEOubiiwVrm0M1i+J9DKJVColt3bSweQr8xAIOzdQVxEgg
Kl17sIBNkKB1dkkuEuw1L7BMKHTu6bxM7jZZeIVUeMASs0IepiLi7qoDrysO2ETICRsXfPxHrc5G
OJnqfjKIiXRzt7G/bNTMx9OCNOIDQwITEOhDFWuBvo2AhF8HNFo0UCZAEtFFYCQ8Qu9ENUs5/ADB
qIICLFwogwuiAsjZBUSQwKwbBIEkhrwNBOoCeWBttnsGUxd1VtAOH7zwAOdzHHGHuceVRBptdFPn
egETg37PI08OwhqVszmORDJPuR3O5my01F3J0VXfhZksSHw7UFF4bFIZErsRq4mIT2uIcbOouVco
503S+bcQuRkIRnoSHZeQH1GERODbDzl2hYwtTWBERh1CLQf4YYsJv0SOI5qg1ZFMdQpJXCGhETeP
U0e0FHypJUIEZMxIQte3iIYl4zQzYUtZ9CZN4gsn889Ek5GNKw2mEMVGEHEDXwBawbBsGwbBsGwb
TMAjMxFdwI1O5kqRlnegZLMOmRr0PnxCNS9Wfa6bsX4gXgIOLg4xhwsVKXtVSNW9eCASPStI4GlF
FgsqCuMdJlh+iu9h7KjDWjZAhYhS2AJDgMMIXjWKbM/WjMzrdQozjZxOGTiVi70LkxFBNyEKzdCJ
wQw3UJWGsjWU5A42A2bZS8htne4CzdEoPPWgsrvM65FOZ8Nrfa0LbWGxGx84iJ0ZNTGomwTATzE8
BXflW0rLNs1U1mJtlWpzEATogohodBlEHQPLh8iEDiRUo2HqEpLHrSx9wg4IiE+t+PU0I2zbJeG8
dQnN+8cWj6BDUnAhOEZ3v5+djo8xBZbe8njHiBrJNsj+CBxbe0SSnAS4F2ELiEBPYgOqd5zO6UAc
zMgWjGJSzpoTRDbaalTsszJemrQ4MlZ5KsVNQzmMiwCOgeHheuB2YjOMJB23Bp6NxSM8RcOeT2qM
/abyhKG8BDsIcCImTZ+hszHx5zJom8kwfXwKcuBTiAiJgQtlPUOM8TXpXRaizIRxqNavVYkkrQQp
GPURDoQclyowxhys3S2bM8xlfYdreXPQP18kOAc4RijhCe+5Hij+LQA9LyNsQv37bt0cyEKeAg6N
R54DKT5yVJnPhHWJPKCHbZ6q5yyyz3vEOidJ7MrDkhven3sxJgWWka8LuKYCvFDcXCe9JFBIPKiI
nPfQtgOjsl41nHv3JDzEiY6kT0FpfuncZ0kYFlCfHGmlTf5CcRL0UgTAOarpv0x3amus5XSNZ6t+
BSpxEsE8PgUdm0SSuGkb7JzlMSlkJYJoCSXYJzCbpVEo2iSmiVRSglBWBLRIsBLlW6okCXH6jZ8G
VIM3oemojO6JFYK7QIdghgxmYgDAgzrNtttttttttBXsv/4y6XvZFWt+UTy6iwT3CWrAlsxa+JRH
rOxfKlCALCAkPl4D7ZykcEOx2OMim0ygjaPp8xFQuJqtpJu+0xaGbBUxM6NnH9rUcdB6deURUVCG
BEks/NQEWLJC9IT9coPVr6cMsBX0+MyYlbePnXC3DxobNRLa++Bfa0xT14K3nIxeDGD4Rx6Z3+Mi
bnySSlb/9UZtbHQyaCwp8vKCYA1UiwpQ+Vu+XLwMSf113adE1TrEb2s24ciGASHCq4L/N3cvMhBE
yiTG0D0shNJSUzIZAaC3QKI/CAMxz5omW99CIg4wKGv1iaH4RZDqyE8oEMtT+4MxJ/P4dTT8+Ke2
EN2WXCDHRyejnqVp0tpI8T0FZTsKI8RCbk/RooplEpuJxmTo46GRcXk744nlhdjjqJXkjBhPYWHF
DulZoE9JgEwh2hRHfLAk7ITQK0tF85wRFTnKS63qheTGksFIOLnmGBiSKM00sSI8cRnc8ixM9Lio
TMVgVEPdUhkkKQdQqIPGxfJQySKjTJjxvJjK0j6OC8VscVObGr09Dt/i9ggONhIHhOMxbhIgdNTQ
JAS5h21EZ8gkmW7AiHQEgbZtgUgEAsO6IxGNMDYXnKaVlMAaicYFLEGxgTxWA5CDMMOt1R340LtO
ibgIQ35bh15fo2F1MO0S47MzYX6onj45aivfjr43CbSZuEyNm1txET1pkiQPUxPEPBw8YuLlj+8P
Es4QTx1qXyrywg2SnxKM6GyvMIakEL3kMa1ZuwELxCkDkJhyE4ivu9FxJfPTsg9P1yJCZ6nts+Ss
WT625opiQSLMjzNJYiCZhDPvU2XGV/bbhJn1aHb/Okrh2muy2l7+fgKFPqCGz8i7Zs/Z52e8avDN
Jpnxe/I58DwiLCzYdZoGrxaZ4yuoGEIDDiFFievWKjLYfsDikcamsIHISJCWPISFW/zgqnYOPXaY
EjgYbReL16eFbFdRNKaqQrVdLkXUJZqFcRkErou1xxwznK1SQmRBS0gy2/2ZpJ6gC4TuE5BmfGOH
AOcAVtisGbaGt9xC/pmvmKunQEJRrc3pz0t/wJAJmm3s7Dm8w83rNaCSgTnFaBIVmA/f7G8jAgJH
Yb/CYnsp3+sRcikRDaBtgmJiCVDucSW5NEEtANxCXuQAW3iZHESkUsNlG7bgJjR5oRhNGckJt2vI
RPMRNozqZs6RPbBJwIfCobewREzCxDYwVhQeSNyD40MqO4EGCRMikZYQbCWNIK6gmgrBKC/XaCfI
UPZ7P9CahN4S0uWIhMJWYAiJFm2hMYnAWUDoICCRmDIJs+Qm4TYhOCCEKMKsiIyZ4AGSIn54ymly
m6q8LmSz6xIpmQm3gEOwF5QiRDxrO7qBczyj4xHXEfmHCIuN8SluiVqJYEkp1xOVgJpQrE1StscM
4m0T/orAkqGAnSUH2Alj6hXilEYSKgkQyZeRVmR2EnZziKpXCANjORISaAtqyT0GfDyE7RMhOgTQ
6SIEhgVhDE6ngJ0hC+GZL4wE4HTaYcxEHSJ8eS/dYJdNLhNrFK3iE4N8OzAE5RN0DOq1uyZkOwPG
QAnJEBBJPSwEVgI3hEEd5cqIIaGCPtMrbDvwFcS4x90PA9X4jDxE2CSE2gGOifR5xKu5TScAA7YB
mh4BoGCl6MoCQktgJQi47t4D1iQCfIJKYmxS4HYAbhLM8BJAl5vE9Alb6wxZAMEohAkwJZAnpGkW
m5wRESNfZSELWdMe4hkHEMg4hkHEMg4hkHEMg4hkHEMg4DJES9wPDITayCQREGir97CDz6CdbNBq
h1CXCaiecG2StQuhW0Hps9Yr8olCoFUzwIJEKDOCUKjCrHvwOMCQGenYJoK9dFOwULNQNgln4MBf
YJpcZhvs7lVIhUvTeQj9RCr5CZyAjFuxvIEz3DETVcJImRBMVkbs7fQe41Jj18LfISXH5ZSZm8tk
1+GFRGoSgTWNhRGTU59A9VecSYSArgCfUcL4AIIBdBOkSw2taqekv5QbfSJp4gUMxJ5bkeQnXyFd
gmgkhN7WCIbaE5SIZhORKJFZpaJClFXtEoKI/vBqfkQ5hJFduNvOa5LHAjK8QdonuEoCWWfHeFoE
wBxRPhhZiJbCneG4zE+ITYakYiLjhEMoIxEElDgKwY23eQjA6RjJRgaBM9sbMLXiEszCZDXITCJ1
dDicdCTRCBBiEgICCA5CUNpsATeUEtM38sMQxDDTabTabGxsvEYDo3ARdvo1ZjGklwm6fxqJqWGD
Jg7KDmTQVb93b7mEvzutdUcWDnJQFlCEpPzMxKNBLASfcpCh0QTEmPvDpE35ImYlxcDMPsjIiYmm
onAE4nkTh9eErZJJfISp2iRfgQRtPaQq8BXPd2CiM2hDpArkJnpiJqY1XMVgA7hMFfRifaQJ63cc
RM/VJ+AlB9gmHDie0dJFQ5tBOcTnB+pNrnmK26tgcUjEVoMwEmOSGiGWJoJcEw8JoobBsCsHKJpE
8YnSJgFcAl0L0EfR3CaibRKB0CQKHaJGnfUGR0SBOwTehy6GX6RJ/ODqpkJiJoDj/8XckU4UJDnj
FbZA

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Your patch looks good.  Thanks.

Before trying the patch, I had the "wrong" impression that headN lines
won't be displayed in 

    Perl->Tools->Imenu-POD headers menu entry.  

I see that Imenu still shows the perldoc headers.

So

    cperl-imenu-index-pod is a variable defined in `cperl-mode.el'.
    Its value is t

    Documentation:
    *Whether POD `=head' directives should be indexed by imenu.

the first line of the docstring is misleading.

----------------------------------------------------------------

Btw, if you are convinced that what this patch does is "the norm" rather
than "an exception", you may as well consider removing the variable.

----------------------------------------------------------------

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> Your patch looks good.  Thanks.
>
> Before trying the patch, I had the "wrong" impression that headN lines
> won't be displayed in 
>
>     Perl->Tools->Imenu-POD headers menu entry.  
>
> I see that Imenu still shows the perldoc headers.
>
> So
>
>     cperl-imenu-index-pod is a variable defined in `cperl-mode.el'.
>     Its value is t
>
>     Documentation:
>     *Whether POD `=head' directives should be indexed by imenu.
>
> the first line of the docstring is misleading.

Perhaps I misunderstand, but when I have that variable set to nil, the
POD imenu entries go away, which is what the docstring says.  Indeed,
which-function-mode uses the imenu index for determining which function
point is in.  I don't know how often which-function-mode calls the
indexing function, so make sure you have the variable set before
visiting the file.  Try again and let me know.

> ----------------------------------------------------------------
>
> Btw, if you are convinced that what this patch does is "the norm" rather
> than "an exception", you may as well consider removing the variable.
>
> ----------------------------------------------------------------

I do think it's better to say that the definitions begin earlier than
what cperl currently says (though it's not a "bug").  However, I think
indexing the POD entries by default is the better way for several
reasons:

1. cperl-mode already does it that way.
2. perl-mode does it.
3. There is no "pod-mode".  For editing pod files, perl-/cperl-mode is
   the best Emacs has.  Moreover, many people _do_ write plain pod
   files, in which case the entries should get indexed.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> Jambunathan K <kjambunathan@gmail.com> writes:
>
>>     Perl->Tools->Imenu-POD headers menu entry.  
>>     cperl-imenu-index-pod is a variable defined in `cperl-mode.el'.

> Perhaps I misunderstand, but when I have that variable set to nil, the
> POD imenu entries go away, which is what the docstring says. 

The behaviour is as documented.  I should have done something wrong
yesterday.

> However, I think indexing the POD entries by default is the better way
> for several reasons:
>
> 1. cperl-mode already does it that way.
> 2. perl-mode does it.
> 3. There is no "pod-mode".  For editing pod files, perl-/cperl-mode is
>    the best Emacs has.  Moreover, many people _do_ write plain pod
>    files, in which case the entries should get indexed.

Hmm.... Losing something to gain some thing else doesn't seem good to
me.

I cursorily looked at cperl-mode.  My gut feeling is that which-function
decides the box it is in by looking at the markers that is returned.  So
we may have to examine what value goes in to the marker.  

My recommendation would be to put the patch on hold.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

> Hmm.... Losing something to gain some thing else doesn't seem good to
> me.

I don't follow. What's being lost?

> I cursorily looked at cperl-mode.  My gut feeling is that which-function
> decides the box it is in by looking at the markers that is returned.  So
> we may have to examine what value goes in to the marker.  

Yes, which-function just looks at the markers in the imenu index.
Changing those markers' positions is exactly what the patch does.

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

Nathan Trapuzzano <nbtrap@nbtrap.com> writes:

> Jambunathan K <kjambunathan@gmail.com> writes:
>
>> Hmm.... Losing something to gain some thing else doesn't seem good to
>> me.
>
> I don't follow. What's being lost?

Indexing of =headNs

>> I cursorily looked at cperl-mode.  My gut feeling is that which-function
>> decides the box it is in by looking at the markers that is returned.  So
>> we may have to examine what value goes in to the marker.  
>
> Yes, which-function just looks at the markers in the imenu index.
> Changing those markers' positions is exactly what the patch does.

When POD headers are there, the Perl file is cut/segmented a certain
way.  When POD headers are not indexed, the Perl file is cut/segmented
differently.  Shouldn't we be cutting at the same spot with or without
perl headers?

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Nathan Trapuzzano]

Jambunathan K <kjambunathan@gmail.com> writes:

>> I don't follow. What's being lost?
>
> Indexing of =headNs

Optionally, and not by default.

> When POD headers are there, the Perl file is cut/segmented a certain
> way.  When POD headers are not indexed, the Perl file is cut/segmented
> differently.  Shouldn't we be cutting at the same spot with or without
> perl headers?

Can you give an example or another screenshot?  I'm not sure what you
mean about "segmenting".

bug#15848: 24.3.50; cperl-mode + which-function: Improve reporting (corner case)

[Jambunathan K]

OP here.  Closed.