Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

Ooops... I had completely forgotten about that. At least that demonstrate how 
constant I am in my way of thinking.
So, the only thing to do is to replace the assoc-string by this 
TeX-assoc-string, right ?
   Vincent.


> From: m...@gnu.org
> Date: Mon, 9 Nov 2015 22:48:05 +0100
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> To: vincent@hotmail.fr
> CC: auctex@gnu.org
>
> Hi Vincent,
>
> 2015-11-09 17:55 GMT+01:00 Vincent Belaïche :
>> Dear Mosé,
>>
>> No there was no strong reason for assoc-string rather than assoc, the
>> point was just that when you know that the key is a string assoc-string
>> is supposed to be faster.
>>
>> My opinion --- with no regard to this specific patch --- is that it
>> would better in the case of XEmacs to define assoc-string as an alias of
>> assoc if the function is unbound.
>
> Well, it turned out there was already an alias `TeX-assoc-string' you
> defined last year ;-)
>
> Bye,
> Mosè
  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Mosè Giordano]

Hi Vincent,

2015-11-11 15:32 GMT+01:00 Vincent Belaïche :
> Ooops... I had completely forgotten about that. At least that demonstrate how 
> constant I am in my way of thinking.
> So, the only thing to do is to replace the assoc-string by this 
> TeX-assoc-string, right ?

Right, I've already done that ;-)

Cheers,
Mosè

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Mosè Giordano]

Hi Vincent,

2015-11-09 17:55 GMT+01:00 Vincent Belaïche :
> Dear Mosé,
>
> No there was no strong reason for assoc-string rather than assoc, the
> point was just that when you know that the key is a string assoc-string
> is supposed to be faster.
>
> My opinion --- with no regard to this specific patch --- is that it
> would better in the case of XEmacs to define assoc-string as an alias of
> assoc if the function is unbound.

Well, it turned out there was already an alias `TeX-assoc-string' you
defined last year ;-)

Bye,
Mosè

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

Dear Mosé,

No there was no strong reason for assoc-string rather than assoc, the
point was just that when you know that the key is a string assoc-string
is supposed to be faster.

My opinion --- with no regard to this specific patch --- is that it
would better in the case of XEmacs to define assoc-string as an alias of
assoc if the function is unbound.

Furthermore, I must say that the way I wrote this patch is quite generic
because I have anticipated that there are other specials characters
(hashsign, backslash, etc...) that we may want some day to
de-escape. However currently the map contains only the comma character,
so using assoc[-string] to search a map of one element might be a little
over-engineered. The code could be simplified without use of a map if 
everybody is content with not anticipating any other escapable characters.

VBR,
    Vincent


> From: m...@gnu.org
> Date: Sun, 8 Nov 2015 01:06:26 +0100
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> To: vincent@hotmail.fr
> CC: auctex@gnu.org
>
> Hi Vincent,
>
> 2015-10-25 21:15 GMT+01:00 Vincent Belaïche :
>> Dear AUCTeX experts,
>>
>> I would like to submit to you the attached patch.
>>
>> [...]
>
> XEmacs doesn't have `assoc-string', is it possible to use `assoc' or
> there was a strong reason for using `assoc-string' instead?
>
> Bye,
> Mosè
  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Mosè Giordano]

Hi Vincent,

2015-10-25 21:15 GMT+01:00 Vincent Belaïche :
> Dear AUCTeX experts,
>
> I would like to submit to you the attached patch.
>
> [...]

XEmacs doesn't have `assoc-string', is it possible to use `assoc' or
there was a strong reason for using `assoc-string' instead?

Bye,
Mosè

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Vincent Belaïche  writes:

>> If there's something else I have to do, please tell me.
>
> Indeed, this revision which I had proposed seems to be useless as far
> as execution speed is concerned --- at least that was not why Stefan
> did the change, and nobody had complained.
>
> However it still have some added value in terms of warning the user in
> case that a duplicate node name is detected.

Ah, right, that's indeed useful.

> So, I am not really opinionated whether this further change is needed,
> it is up to you...

I think the hash-table approach is a bit over-engineered.  I've
committed something simpler in

http://git.savannah.gnu.org/cgit/auctex.git/commit/?id=fb09c26aa539c1a3a48ccf78098ed536f48c4870

I've also removed the optional NODES parameter because apparently nobody
used it anyway.

Bye,
Tassilo

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Mosè Giordano]

Il 28/ott/2015 10:33, "Tassilo Horn"  ha scritto:
>
> Vincent Belaïche  writes:
>
> > I have not tested the change, but reading this:
> >
> >   (message "There are duplicate nodes:")
> >   (dolist (dup (nreverse dups))
> > (message "%s on line %d" (car dup) (cdr dup
> >
> > I gather that the "There are duplicate nodes:" message will be output
> > even when dups is nil. Furthermore if you have N duplicates you will
have
> > N+1 message and only the last one will remain in the minibuffer ---
> > meaning that the user has to go to the *Message* buffer. That is why I
> > was using the `warn' function. If you want to use `message' it might be
> > better to contatenate all the messages together, maybe something like
> > this (not tested):
> >
> > (and dups
> >(message "There are duplicate nodes: %s"
> >   (mapconcat (lambda (x) (format "%s on line %d"
> >(car x) (cdr x))) (nreverse dups) "\n\t")))
>
> Gosh, and there was also a void-variable error because I didn't test
> again after removing the nodes argument. ;-)
>
> Ok, now I've fixed it and use `display-warning' with CLASS set to
> 'AUCTeX so that users also see who issued the warning.  That's much
> better.  But I'm not entirely sure if very old emacsen have that: I
> checked GNU Emacs 22+ and XEmacs 21.5 but it would be great if somebody
> could check Emacs 21.4 and XEmacs 21.4.

Didn't test the whole patch, but I can confirm XEmacs 21.4 has
display-warning.

Bye,
Mosè
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Mosè Giordano  writes:

>> Ok, now I've fixed it and use `display-warning' with CLASS set to
>> 'AUCTeX so that users also see who issued the warning.  That's much
>> better.  But I'm not entirely sure if very old emacsen have that: I
>> checked GNU Emacs 22+ and XEmacs 21.5 but it would be great if
>> somebody could check Emacs 21.4 and XEmacs 21.4.
>
> Didn't test the whole patch, but I can confirm XEmacs 21.4 has
> display-warning.

Perfect, thank you.

Bye,
Tassilo

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

> From: t...@gnu.org
> To: vincent@hotmail.fr
> CC: auctex@gnu.org
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> Date: Wed, 28 Oct 2015 08:43:08 +0100
>
> Vincent Belaïche  writes:
>
>>> If there's something else I have to do, please tell me.
>>
>> Indeed, this revision which I had proposed seems to be useless as far
>> as execution speed is concerned --- at least that was not why Stefan
>> did the change, and nobody had complained.
>>
>> However it still have some added value in terms of warning the user in
>> case that a duplicate node name is detected.
>
> Ah, right, that's indeed useful.
>
>> So, I am not really opinionated whether this further change is needed,
>> it is up to you...
>
> I think the hash-table approach is a bit over-engineered. I've
> committed something simpler in
>
> http://git.savannah.gnu.org/cgit/auctex.git/commit/?id=fb09c26aa539c1a3a48ccf78098ed536f48c4870
>

I have not tested the change, but reading this:

  (message "There are duplicate nodes:")
  (dolist (dup (nreverse dups))
    (message "    %s on line %d" (car dup) (cdr dup

I gather that the "There are duplicate nodes:" message will be output
even when dups is nil. Furthermore if you have N duplicates you will have
N+1 message and only the last one will remain in the minibuffer ---
meaning that the user has to go to the *Message* buffer. That is why I
was using the `warn' function. If you want to use `message' it might be
better to contatenate all the messages together, maybe something like
this (not tested):

(and dups
   (message "There are duplicate nodes: %s"
  (mapconcat (lambda (x) (format "%s on line %d"
   (car x) (cdr x))) (nreverse dups) "\n\t")))

> I've also removed the optional NODES parameter because apparently nobody
> used it anyway.
>

Yes, I had used this nodes argument in case that the code is evolved to
do partial rescan instead of full rescan...  But I agree that this is
useless for now, let us not make the code confusing. Anyway, if we are
to do more clever things it would be better to integrate the change with
reftex so that node references can also appear in the reftex *toc* buffer
and we also get all the multifile stuff.
Anyway the current code does not handle the @anchor{...} reference, just 
nodes --- better than nothing.


> Bye,
> Tassilo

Bye,
   Vincent.

PS: BTW, what is the use of calling `nreverse' over `nodes' at function return ?
  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Vincent Belaïche  writes:

> I have not tested the change, but reading this:
>
>   (message "There are duplicate nodes:")
>   (dolist (dup (nreverse dups))
>     (message "    %s on line %d" (car dup) (cdr dup
>
> I gather that the "There are duplicate nodes:" message will be output
> even when dups is nil. Furthermore if you have N duplicates you will have
> N+1 message and only the last one will remain in the minibuffer ---
> meaning that the user has to go to the *Message* buffer. That is why I
> was using the `warn' function. If you want to use `message' it might be
> better to contatenate all the messages together, maybe something like
> this (not tested):
>
> (and dups
>    (message "There are duplicate nodes: %s"
>   (mapconcat (lambda (x) (format "%s on line %d"
>    (car x) (cdr x))) (nreverse dups) "\n\t")))

Gosh, and there was also a void-variable error because I didn't test
again after removing the nodes argument. ;-)

Ok, now I've fixed it and use `display-warning' with CLASS set to
'AUCTeX so that users also see who issued the warning.  That's much
better.  But I'm not entirely sure if very old emacsen have that: I
checked GNU Emacs 22+ and XEmacs 21.5 but it would be great if somebody
could check Emacs 21.4 and XEmacs 21.4.

>> I've also removed the optional NODES parameter because apparently nobody
>> used it anyway.
>
> Yes, I had used this nodes argument in case that the code is evolved
> to do partial rescan instead of full rescan...  But I agree that this
> is useless for now, let us not make the code confusing. Anyway, if we
> are to do more clever things it would be better to integrate the
> change with reftex so that node references can also appear in the
> reftex *toc* buffer and we also get all the multifile stuff.  Anyway
> the current code does not handle the @anchor{...} reference, just
> nodes --- better than nothing.

Patches welcome!

> PS: BTW, what is the use of calling `nreverse' over `nodes' at
> function return ?

Just to have the list of nodes be in the definition order.  That's not
really required but more intuitive.

Bye,
Tassilo

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Stefan Monnier  writes:

Hi guys,

I'm somehow lost.  It seems that some mail, presumably Stefan ->
Vincent, started this discussion but that hasn't been delivered to me.
So well, yes, `pushnew' does the right thing, and in Vincent's patch to
tex-info.el, I replaced his `add-to-list' call with (pushnew ... :test
#'equal) anyhow.

If there's something else I have to do, please tell me.

>> Ok, I should have read the *Help* more carefully, it states
>>    (cl-pushnew X PLACE): insert X at the head of the list if not already 
>> there.
>
>> and I understood "there" as meaning "head of the list", instead of
>> just "list".

Indeed, that's not overly clear.

>> Of course the second sentence clarifies the point, but I overlooked
>> it.
>
>> Maybe it could be rephrased as:
>
>>    (cl-pushnew X PLACE): insert X at the head of the list PLACE unless X is
>> already an element of PLACE.
>
>> If you agree with the rephrasal, I can implement the change.
>
> Fine by me,

A bit too long given that the first sentence of a docstring should fit
into one line and still mention all parameters.  But why does the
docstring start with a synopsis?  I'd just delete that and say

Insert X as head of list PLACE unless PLACE already contains X.

Bye,
Tassilo


___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Joost Kremers  writes:

> On Di, Okt 27 2015, Tassilo Horn  wrote:
>> A bit too long given that the first sentence of a docstring should fit
>> into one line and still mention all parameters.
>
> Just a nit pick, but AFAIK the convention is that the first line
> should summarise what the function does, but there is no requirement
> that it also mentions all variables that a function takes.

Right, but in general that goes hand in hand with the "main" parameters,
e.g., with `pushnew' X and PLACE are surely relevant for such a summary
whereas the additional keyword arguments it takes are not.

Bye,
Tassilo


___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Joost Kremers]

On Di, Okt 27 2015, Tassilo Horn  wrote:
> A bit too long given that the first sentence of a docstring should fit
> into one line and still mention all parameters.

Just a nit pick, but AFAIK the convention is that the first line should
summarise what the function does, but there is no requirement that it
also mentions all variables that a function takes. (Well, that's to say,
I don't always do this in my Elisp files and flycheck never complains
about it... ;-)


-- 
Joost Kremers
Life has its moments

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[David Kastrup]

Joost Kremers  writes:

> On Di, Okt 27 2015, Tassilo Horn  wrote:
>> A bit too long given that the first sentence of a docstring should fit
>> into one line and still mention all parameters.
>
> Just a nit pick, but AFAIK the convention is that the first line should
> summarise what the function does, but there is no requirement that it
> also mentions all variables that a function takes. (Well, that's to say,
> I don't always do this in my Elisp files and flycheck never complains
> about it... ;-)

The entire DOC string should mention all parameters in order.  The "in
order" criterion does not make a whole lot of sense to me really since
the order of parameters is often determined by technical considerations
rather than how one would like to arrange a DOC string grammatically,
but there you are.

-- 
David Kastrup

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

> Hi guys,
> 

Hello Tassilo,

> I'm somehow lost.  It seems that some mail, presumably Stefan ->
> Vincent, started this discussion but that hasn't been delivered to me.
> So well, yes, `pushnew' does the right thing, and in Vincent's patch
> to tex-info.el, I replaced his `add-to-list' call with (pushnew
> ... :test #'equal) anyhow.

Another discussion has started concerning the docstring of cl-pushnew.


Just to come back to the initial point concerning revision of function
Texinfo-make-node-list, what has happened is that I was not aware of the
rationale behind replacing add-to-list by pushnew, Stefan has clarified
that add-to-list is not good for variables bound by a let statement and
that this was the reason.

I was speculating that there was an execution speed issue, and on top of
that I had misunderstood that pushnew checks duplication against the
whole list. That is why I had proposed a revision.

> 
> If there's something else I have to do, please tell me.

Indeed, this revision which I had proposed seems to be useless as far as
execution speed is concerned --- at least that was not why Stefan did
the change, and nobody had complained.

However it still have some added value in terms of warning the user in
case that a duplicate node name is detected.

So, I am not really opinionated whether this further change is needed,
it is up to you...

Please note that I have just made some evaluation of the speed
difference for 1225 nodes in a file that contrains only @node
statements, on my machine the current implementation takes 0.04s, while
the proposed revision 0.01s, although it is 4 times as fast, I cannot
say that speed on its own is an issue worth the change. All the more
that the difference would certainly be smaller in a buffer where nodes
are not empty.

VBR,
    Vincent Belaïche  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Tassilo Horn]

Vincent Belaïche  writes:

Hi Vincent,

> I would like to submit to you the attached patch.

Very cool, thanks a lot.

> + [snip]
> +
>  2014-09-13  Vincent Belaïche  

As you see, the patch was against a pretty old version and didn't apply
anymore cleanly.  I think I was able to merge it correctly but in the
future it would be great if you could base your patches on the Git
version.

Please validate that I didn't screw up things in

http://git.savannah.gnu.org/cgit/auctex.git/commit/?id=d56502e7dcba118c62d38f6fe2f5a855834e6816

Thanks again!
Tassilo


___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

Ok, I should have read the *Help* more carefully, it states


   (cl-pushnew X PLACE): insert X at the head of the list if not already there.

and I understood "there" as meaning "head of the list", instead of just "list". 
Of course the next sentence in the docstring clarifies the point, but I 
overlooked it.

Maybe it could be rephrased as:

   (cl-pushnew X PLACE): insert X at the head of list PLACE unless X is already 
an element of PLACE.

If you agree with the rephrasal, I can implement the change.

 Vincent.


> From: monn...@iro.umontreal.ca
> To: vincent@hotmail.fr
> CC: t...@gnu.org; auctex@gnu.org
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> Date: Mon, 26 Oct 2015 13:34:57 -0400
>
>> interest. Some time ago Stefan Monnier changed add-to-list by pushnew, I
>> don't know what the motivation was, probably somebody had complained
>> that add-to-list was taking too much time.
>
> add-to-list doesn't work on lexical variables.
>
>> However these are not functionally equivalent, pushnew will skip the
>> replacement only if the node name is already there on top of the list,
>> while add-to-list will skip it if it is already there at any place.
>
> Not so: (cl-)pushnew looks at the whole list as well.
>
> I don't have much opinion on this function in general, since I don't
> use it. All I care about is to not re-introduce the
> add-to-list abomination (it's fine for use in a user's ~/.emacs to add
> some element to a global variable, but definitely not when operating on
> a let-bound variable).
>
>
> Stefan
  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

Ok, I should have read the *Help* more carefully, it states


   (cl-pushnew X PLACE): insert X at the head of the list if not already there.

and I understood "there" as meaning "head of the list", instead of just "list". 
Of course the second sentence clarifies the point, but I overlooked it.

Maybe it could be rephrased as:

   (cl-pushnew X PLACE): insert X at the head of the list PLACE unless X is 
already an element of PLACE.

If you agree with the rephrasal, I can implement the change.

 Vincent.


> From: monn...@iro.umontreal.ca
> To: vincent@hotmail.fr
> CC: t...@gnu.org; auctex@gnu.org
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> Date: Mon, 26 Oct 2015 13:34:57 -0400
>
>> interest. Some time ago Stefan Monnier changed add-to-list by pushnew, I
>> don't know what the motivation was, probably somebody had complained
>> that add-to-list was taking too much time.
>
> add-to-list doesn't work on lexical variables.
>
>> However these are not functionally equivalent, pushnew will skip the
>> replacement only if the node name is already there on top of the list,
>> while add-to-list will skip it if it is already there at any place.
>
> Not so: (cl-)pushnew looks at the whole list as well.
>
> I don't have much opinion on this function in general, since I don't
> use it. All I care about is to not re-introduce the
> add-to-list abomination (it's fine for use in a user's ~/.emacs to add
> some element to a global variable, but definitely not when operating on
> a let-bound variable).
>
>
> Stefan
  
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Stefan Monnier]

> interest. Some time ago Stefan Monnier changed add-to-list by pushnew, I
> don't know what the motivation was, probably somebody had complained
> that add-to-list was taking too much time.

add-to-list doesn't work on lexical variables.

> However these are not functionally equivalent, pushnew will skip the
> replacement only if the node name is already there on top of the list,
> while add-to-list will skip it if it is already there at any place.

Not so: (cl-)pushnew looks at the whole list as well.

I don't have much opinion on this function in general, since I don't
use it.  All I care about is to not re-introduce the
add-to-list abomination (it's fine for use in a user's ~/.emacs to add
some element to a global variable, but definitely not when operating on
a let-bound variable).


Stefan

___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex

Re: [AUCTeX] Patch to tex-info.el to support node name completion in @..ref commands, and solve a few other pbs

[Vincent Belaïche]

Hello Tassilo,

To Stefan also concerning add-to-list -> pushnew change...

Feedback inserted below...


> From: t...@gnu.org
> To: vincent@hotmail.fr
> Subject: Re: [AUCTeX] Patch to tex-info.el to support node name completion in 
> @..ref commands, and solve a few other pbs
> Date: Mon, 26 Oct 2015 14:16:20 +0100
>
> Vincent Belaïche  writes:
>
> Hi Vincent,
>
>> Ooops... sorry, I hadn't noticed that you also had switched to git.
>
> Hey, that has been in April 2013! :-)

Ok, I understand what has happened. I am using some script to do all the
refreshing of the work area, and installation, when I had modified this
script from SVN to GIT quite some time ago I was not so familiar with
git at that time, and I wrote `git fetch' instead of `git pull' for the
case of already existing work area...

>
>> I will do the checks ASAP and let you know.
>
> Great, thank you. I think the only hunk that didn't apply is the one
> adapting `Texinfo-insert-node' so have a special look on that.
>

I had a look at function Texinfo-make-node-list, which is the one of
interest. Some time ago Stefan Monnier changed add-to-list by pushnew, I
don't know what the motivation was, probably somebody had complained
that add-to-list was taking too much time.

However these are not functionally equivalent, pushnew will skip the
replacement only if the node name is already there on top of the list,
while add-to-list will skip it if it is already there at any place.

So, I assume that the motivation was to avoid any duplicate in the
resulting list. Anyways node names are supposed to be unique, so whether
it be pushnew or add-to-list that should make a difference only in the
case of boguous Texinfo source code

Another problem is that the user is not warned if there are duplicate
node names, why doing this check if then things go by silently.

So I propose to revise the code as follows:
(defun Texinfo-make-node-list ( nodes)
  ;; Build list of nodes in current buffer.
  ;; (What about using `imenu--index-alist'?)
  ;; FIXME: Support multi-file documents.
  (let ((htable (make-hash-table :test 'equal)) hentry pt key msg)
    (dolist (n nodes)
  (puthash (car n) t htable))
    (save-excursion
  (goto-char (point-min))
  (while (re-search-forward "^@node\\b" nil t)
    (skip-chars-forward "[:blank:]")
    (setq key  (Texinfo-nodename-de-escape
          (buffer-substring-no-properties
           (setq pt (point)) (progn (skip-chars-forward "^\r\n,")
                  (skip-chars-backward "[:blank:]")
                  (point
      hentry (gethash key htable))
    (cond
     ((integerp hentry)
      (push (format "AUCTeX/Texinfo:%d: Duplicate node name `%s' at line %d" 
(line-number-at-pos hentry) key (line-number-at-pos pt)) msg))
     ((eq hentry t)
      (puthash key pt htable))
     (t
      (puthash key pt htable)
      (push (list key) nodes)
  (if msg (warn "AUCTeX/Texinfo...\n%s" (mapconcat 'identity (nreverse msg) 
"\n"
  nodes)

The changes are as follows:

- use hash table, to keep both speed advantage of pushnew and thorough
  duplicate check of add-to-list

- warn the user if a duplicate is found

> Bye,
> Tassilo

Bye,
   Vincent
___
auctex mailing list
auctex@gnu.org
https://lists.gnu.org/mailman/listinfo/auctex