GNU bug report logs - #76108
Call `modify_text' only on the text being replaced in `subst-char-in-region'

Previous Next

Full log

View this message in rfc822 format

From: Thuna <thuna.cing <at> gmail.com>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 76108 <at> debbugs.gnu.org
Subject: bug#76108: Call `modify_text' only on the text being replaced in `subst-char-in-region'
Date: Fri, 07 Feb 2025 16:19:36 +0100

>> I'm assuming that there isn't any particular reason why `modify_text'
>> should be called on the whole paragraph, but if there is, then that's ok
>> too, I suppose.  I don't see any reason why this should be breaking
>> stuff, but I haven't actually checked.
>
> The reason is that running the modification hooks N times (to replace
> N chars) would be expensive.

What if we called `modify_text' for each contiguous region of FROMCHARs?
It should be faster than calling it for each character and even
occasionally (though I wouldn't count on it) faster than calling it over
the whole region since it would avoid running hooks unnecessarily.  Do
you know of a way we could benchmark it?

> The `modification-hooks` text property is weaker than the
> `modification-hooks` overlay property: if you use an overlay here, then
> you should be able to get the behavior your test wants by checking in
> your `hook` the value of the second argument (which indicates if it's
> the "before change" or the "after change" case): while the "before
> change" call covers the whole region (actually, IIRC it covers the
> region between the first substituted chars and the end argument), the
> "after change" should be more specific and cover only the region between
> the first and the last chars that were substituted.

And that is the intended and desired behavior?  Personally I find it
quite unintuitive to have `modification-hooks' be "Hooks that are called
when some characters in region _might_ be modified".

I don't think that it's just me, either; in buffer-tests.el's
`overlay-modification-hooks' there is a comment stating
> Replacing text never calls `insert-in-front-hooks' or
> `insert-behind-hooks'.  It calls `modification-hooks' if the overlay
> covers any text that has changed.
which seems to suggest that the tests there were also written with the
understanding that `modification-hooks' should only be called for
characters actually being replaced.  Of course, those tests are for
`replace-match', but still, I see no reason why the two should have
different semantics (sans performance considerations, of course).

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.