allbery@axcess.UUCP (Brandon S. Allbery) (01/16/88)
I have a function which wants to narrow the current buffer to a region only if it is not already narrowed. (The narrowing separates parts of a buffer which are edited with different major modes.) If the buffer is already narrowed, I want the function to signal an error. The only way I have figured out so far to do this is the following piece of ugliness: (let ((old-min (point-min)) (old-max (point-max))) (unwind-protect (progn (widen) (if (not (and (= old-min (point-min)) (= old-max (point-max)))) (error "You already have a C section in this file."))) (narrow-to-region old-min old-max))) I personally consider stuff like "unwind-protect" to be for use in special situations -- and checking for narrowing isn't one of them, or shouldn't be. I guess it's done in the C code rather than Lisp (since the format %n is used in the modeline rather than checking for a variable's value), but why isn't some special function provided ( (narrowp) or something ) to do this? -- ___ ________________, Brandon S. Allbery cbosgd \ ' \/ __ __, __, aXcess Company mandrill| __ | /__> <__ <__ 6615 Center St. #A1-105 !ncoast! / ` | \__. .__> .__> Mentor, OH 44060-4101 necntc | axcess!allbery \___/\________________. Moderator, comp.sources.misc hoptoad/
wolfgang@mgm.mit.edu (Wolfgang Rupprecht) (01/18/88)
In article <138@axcess.UUCP> allbery@axcess.UUCP (Brandon S. Allbery) writes: >I have a function which wants to narrow the current buffer to a region only if >it is not already narrowed. [...] but why isn't >some special function provided ( (narrowp) or something ) to do this? (defun narowp () "Return t if buffer is narrowed." (/= 1 (point-min))) --- Wolfgang Rupprecht ARPA: wolfgang@mgm.mit.edu (IP 18.82.0.114) Independent Consultant UUCP: {mit-eddie!mgm.mit.edu,mirror!mit-mgm}!wolfgang VOICE: Hey_Wolfgang!_(617)_267-4365
quiroz@cs.rochester.edu (Cesar Quiroz) (01/18/88)
Brandon Allbery wants a function defined to check if the current
buffer is under narrowing. I take he was thinking of a C primitive
or such, to complement the other narrowing functions. Wolfgang
Rupprecht suggests this function:
:(defun narowp ()
:"Return t if buffer is narrowed."
: (/= 1 (point-min)))
This doesn't work, as a little reflection shows. The narrowing
might permit (= 1 (point-min))! Think of narrowing to the first
page, for instance. The original idea (to check if either point-min
or point-max changed) is better. A possible implementation,
although I suspect Brandon doesn't need this code:
(defun narrowp ()
"True if the current buffer is narrowed.
The value actually returned is a list of the end points of the
restriction."
(let ((min (point-min))
(max (point-max)))
(if (and (= min 1) (= max (1+ (buffer-size))))
nil
(list min max))))
This still fails if the restriction coincides with the whole
buffer...
--
Cesar Augusto Quiroz Gonzalez
Department of Computer Science ...allegra!rochester!quiroz
University of Rochester or
Rochester, NY 14627 quiroz@cs.rochester.eduwolfgang@mgm.mit.edu (Wolfgang Rupprecht) (01/18/88)
%$#* news software wouldn't let me cancel my article.
No checking for point-min + point-max isn't enough to tell if a buffer
is narrowed. The only real way is probably to add some C code.
This might just do the trick. To editfns.c add:
DEFUN ("narrowp", Fnarrowp, Snarrowp, 0, 0, "",
"Returns t if buffer is narrowed, else returns nil.")
()
{
return (bf_head_clip > 1 || bf_tail_clip > 0) ? Qt: Qnil;
}
Towards the end of the file, add this extra line to syms_of_editfns:
void
syms_of_editfns ()
{
[...]
defsubr (&SNarrowp); /* add this line only ! */
}
Now recompile + dump.
(Now to see why nntp-rn screws up and refuses to let me cancel my
previous article.)
---
Wolfgang Rupprecht ARPA: wolfgang@mgm.mit.edu (IP 18.82.0.114)
Independent Consultant UUCP: {mit-eddie!mgm.mit.edu,mirror!mit-mgm}!wolfgang
VOICE: Hey_Wolfgang!_(617)_267-4365wesommer@athena.mit.edu (William E. Sommerfeld) (01/18/88)
In article <138@axcess.UUCP> allbery@axcess.UUCP (Brandon S. Allbery) writes: >I have a function which wants to narrow the current buffer to a region only if >it is not already narrowed. ... If the buffer is already narrowed, I >want the function to signal an error. > >The only way I have figured out so far to do this is the following piece of >ugliness: > (gross use of unwind-protect deleted in the interest of ) Rob Krawitz will probably also respond to this, but... The other suggestions on how to detect a current restriction makes sense, but there is there is a special form, save-restriction, which could be used to cleanly implement constructs similar to what you have above (which you're probably want to use in the yacc/bison mode): (save-restriction form1 form2 ...) is sort of equivalent to (let ((old-min (point-min)) (old-max (point-max))) (unwind-protect (progn form1 form2 ...) (widen) (narrow-to-region old-min old-max))) (the difference being that old-min and old-max won't be visible inside the execution of the internal forms). Another form related to this which is real useful is save-excursion, which saves `point' and `mark' on entry, and _always_ restores them when the form is exited. The two are sort of like gsave and grestore in postscript, except that they also interact well with exception handling. Look in rmail.el and info.el to see a lot of use of both of these forms. - Bill
allbery@axcess.UUCP (Brandon S. Allbery) (01/24/88)
In article <2336@bloom-beacon.MIT.EDU>, wolfgang@mgm.mit.edu (Wolfgang Rupprecht) writes: +--------------- | In article <138@axcess.UUCP> allbery@axcess.UUCP (Brandon S. Allbery) writes: | >I have a function which wants to narrow the current buffer to a region if | >it is not already narrowed. [...] but why isn't | >some special function provided ( (narrowp) or something ) to do this? | | (defun narowp () | "Return t if buffer is narrowed." | (/= 1 (point-min))) +--------------- Nice try, but if you've seen my Yacc/Bison mode for Emacs, you might notice that the first thing it does is narrow to the first two sections of a file (i.e. make the pure C code at the end of the file invisible). So (point-min) will return 1 regardless. -- ___ ________________, Brandon S. Allbery cbosgd \ ' \/ __ __, __, aXcess Company mandrill| __ | /__> <__ <__ 6615 Center St. #A1-105 !ncoast! / ` | \__. .__> .__> Mentor, OH 44060-4101 necntc | axcess!allbery \___/\________________. Moderator, comp.sources.misc hoptoad/
allbery@axcess.UUCP (Brandon S. Allbery) (01/24/88)
In article <5959@sol.ARPA>, quiroz@cs.rochester.edu (Cesar Quiroz) writes: +--------------- | page, for instance. The original idea (to check if either point-min | or point-max changed) is better. A possible implementation, | although I suspect Brandon doesn't need this code: > ... | (if (and (= min 1) (= max (1+ (buffer-size)))) > ... | This still fails if the restriction coincides with the whole | buffer... +--------------- As far as I can see, if the narrowed region is the whole buffer the C code doesn't consider it narrowed. (At least, the mode line doesn't.) In any case, I just want to know if the whole buffer is visible or not, and narrowing isn't pushed/popped in the C code, so everything works for my purposes. (I'm not upset about narrowing not being pushed/popped; looks like a Lisp wrapper around narrow-to-region and widen would provide that easily if I wanted it.) Yes, I was thinking of a C primitive. However, (buffer-size) looks like it will do the job. Tell me, is there any more complete way to find out all these little functions (the partially complete gnu functions manual isn't really useful and the help stuff virtually requires you to know what you're looking for beforehand) than to sit down and start reading ALL the Emacs code? Anyway, thanks for the help. Someone might want to replace my gross hack to do this if they saved a copy of the yacc-mode I sent out. -- ___ ________________, Brandon S. Allbery cbosgd \ ' \/ __ __, __, aXcess Company mandrill| __ | /__> <__ <__ 6615 Center St. #A1-105 !ncoast! / ` | \__. .__> .__> Mentor, OH 44060-4101 necntc | axcess!allbery \___/\________________. Moderator, comp.sources.misc hoptoad/