bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Dove Young]

[-- Attachment #1: Type: text/plain, Size: 2090 bytes --]

The original code reads like this


(defun cperl-write-tags (&optional file erase recurse dir inbuffer noxs
topdir)
  ;; If INBUFFER, do not select buffer, and do not save
  ;; If ERASE is `ignore', do not erase, and do not try to delete old info.
  (require 'etags)
  (if file nil
    (setq file (if dir default-directory (buffer-file-name)))
    (if (and (not dir) (buffer-modified-p)) (error "Save buffer first!")))
  (or topdir
      (setq topdir default-directory))
  (let ((tags-file-name "TAGS")
    (case-fold-search (eq system-type 'emx))
    xs rel tm)
... ... ... ...

This line (let ((tags-file-name "TAGS") hard coded the tags-file-name to
"TAGS".  This way is very bad. It prevents any users to customise their own
tags files. For example in my environment I have to customise tags file
names.

I use pde-mode (Perl Development Environment) in my daily work. it leverage
cperl-mode but would generate tags files base on Perl syntax. It is quite
different than what etags would do. This kind of tags file cannot be used in
any other features, such as speedbar, etc. So the reasonable way is to store
this kind of tags into separate files. But it cannot because of the hard
code in cperl-mode.el .


So would I suggest to re-factory cperl-mode.el and change "TAGS" to a
variable? Like following:

(defvar cperl-tags-file-name "TAGS" "TAGS file name"

(defun cperl-write-tags (&optional file erase recurse dir inbuffer noxs
topdir)
  ;; If INBUFFER, do not select buffer, and do not save
  ;; If ERASE is `ignore', do not erase, and do not try to delete old info.
  (require 'etags)
  (if file nil
    (setq file (if dir default-directory (buffer-file-name)))
    (if (and (not dir) (buffer-modified-p)) (error "Save buffer first!")))
  (or topdir
      (setq topdir default-directory))
  (let ((tags-file-name cperl-tags-file-name)
    (case-fold-search (eq system-type 'emx))
    xs rel tm)
... ... ... ...

In this way, I can customise it by set cperl-tags-file-name in other value
to avoid to break any other features:

(setq cperl-tags-file-name "PDE-TAGS")

-- 
M-x Thinks

[-- Attachment #2: Type: text/html, Size: 2402 bytes --]

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Lars Magne Ingebrigtsen]

Dove Young <dove.young@gmail.com> writes:

> So would I suggest to re-factory cperl-mode.el and change "TAGS" to a
> variable? Like following:
>
> (defvar cperl-tags-file-name "TAGS" "TAGS file name"

That sounds like a reasonable request to me, but Emacs is in a feature
freeze at the moment, and I'm not sure adding stuff like this is
"allowed" during the freeze.

Stefan, Chong?

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Stefan Monnier]

>> So would I suggest to re-factory cperl-mode.el and change "TAGS" to a
>> variable? Like following:
>> (defvar cperl-tags-file-name "TAGS" "TAGS file name"

> That sounds like a reasonable request to me, but Emacs is in a feature
> freeze at the moment, and I'm not sure adding stuff like this is
> "allowed" during the freeze.

I'd rather that such a change be installed in Ilya's upstream
version first.  We already have plenty of merge problems because of
patches he rejected.  BTW, the recent change to use define-derived-mode
introduced a problem where the cperl-mode-hook gets called twice, so
I think we had better revert it.


        Stefan

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Lars Magne Ingebrigtsen]

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

> I'd rather that such a change be installed in Ilya's upstream
> version first.  We already have plenty of merge problems because of
> patches he rejected.

Right.  Would forwarding the report to Ilya and closing the bug report
in debbugs be the right thing to do?

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Glenn Morris]

Lars Magne Ingebrigtsen wrote:

> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>
>> I'd rather that such a change be installed in Ilya's upstream
>> version first.  We already have plenty of merge problems because of
>> patches he rejected.

This always makes me wonder: what is the point of having a file in Emacs
where modifying it is not allowed (or at least, is strongly
discouraged)? Why not move it to GNU ELPA? It's not like there isn't
another perl mode in Emacs, which is actually used by default.
	
> Right. Would forwarding the report to Ilya and closing the bug report
> in debbugs be the right thing to do?

Debbugs has the ability to mark a report as forwarded, see
http://debbugs.gnu.org/Developer.html#forward

Such reports are not usually closed.

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Stefan Monnier]

>>> I'd rather that such a change be installed in Ilya's upstream
>>> version first.  We already have plenty of merge problems because of
>>> patches he rejected.
> This always makes me wonder: what is the point of having a file in Emacs
> where modifying it is not allowed (or at least, is strongly
> discouraged)? Why not move it to GNU ELPA? It's not like there isn't
> another perl mode in Emacs, which is actually used by default.

Good point.  But AFAIK it does have a fairly large number of users.
Chong, Ilya, what do you guys think?


        Stefan

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Lars Magne Ingebrigtsen]

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

> Good point.  But AFAIK it does have a fairly large number of users.
> Chong, Ilya, what do you guys think?

cperl-mode has lots and lots of users, I think.  This would break their
Emacs 24 work flow significantly, in my opinion.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#8802: cperl-write-tags function in cperl-mode.el hard coded tags-file-name

[Stefan Kangas]

tags 8802 + fixed
close 8802 28.1
thanks

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

>>> So would I suggest to re-factory cperl-mode.el and change "TAGS" to a
>>> variable? Like following:
>>> (defvar cperl-tags-file-name "TAGS" "TAGS file name"
>
>> That sounds like a reasonable request to me, but Emacs is in a feature
>> freeze at the moment, and I'm not sure adding stuff like this is
>> "allowed" during the freeze.
>
> I'd rather that such a change be installed in Ilya's upstream
> version first.  We already have plenty of merge problems because of
> patches he rejected.

So given that the above merge problems are no longer a concern, I've
now added such a variable on master.

I'm therefore also closing this bug report.