Can reproduce using ISEARCH with
#v+
w3m -config /dev/null -o keymap_file=/dev/null ~/word-joiner-w3m.html
#v-

----- Forwarded message from Vincent Lefevre <vincent@vinc17.net> -----

Subject: Bug#1100487: w3m: bad display with image and U+2060 WORD JOINER
From: Vincent Lefevre <vincent@vinc17.net>
To: Debian Bug Tracking System <submit@bugs.debian.org>
Reply-To: Vincent Lefevre <vincent@vinc17.net>, 1100487@bugs.debian.org
X-Original-To: mail@rkta.de
List-Id: <w3m.tracker.debian.org>
Message-ID: <20250314122253.GA4191353@cventin.lip.ens-lyon.fr>

On Fri, Mar 07, 2025 at 08:00:26PM +0100, bptato@chawan.net wrote:
> From: bptato <nincsnevem662@gmail.com>
> 
> * </dl>, <dt> now closes the last open <dd>
> * remove redundant if/else in <dd> processing
> 
> This is mainly aimed at fixing the rendering of doc-jp/FAQ.html, which
> does not close its dd tags and hence was rendered rather strangely since
> my previous modifications to dl rendering.
> ---
>  file.c        | 20 ++++++++++++++------
>  t/dl.expected | 11 +++++++++++
>  t/dl.html     | 14 ++++++++++++++
>  3 files changed, 39 insertions(+), 6 deletions(-)

On Tue, Feb 25, 2025 at 05:52:36AM +0000, lesssalt wrote:
> This patch provides link hint functionality by using the existing
> getLinkNumberStr function. It adds the ability to display and input link
> numbers using alphanumeric characters, with the new displayLinkAlpha
> option.

Thanks. Will take me some time to do a proper review. Two remarks from a
quick look:

Might be worth it to put all the code (accept for the changes to
anchor.c, fm.h and rc.c) into a separate file, e.g. lhints.[ch].

Which part of the buffer is kept in the view when issuing a command
seems to be dependent on the cursor position. As someone whose cursor is

I finally fixed a bug when using w3m with -dump_source.  Users are
strongly advised to either user the current master branch with commit
32193d61 (Fix -dump_source on compressed streams, 2025-03-05) or to
backport this patch.

JFTR, doing a proper release and going back to proper versioning is on
my roadmap.

On Fri, Mar 07, 2025 at 02:34:15AM +0000, remph wrote:
> Drat, of course I forgot to commit (:
> Here's the diff -- is that acceptable or should I resubmit the patch?

Please send a v2.

On Fri, Jan 03, 2025 at 11:31:24AM +0000, skejg wrote:
> On Fri Jan 3, 2025 at 10:01 +0100, Rene Kita <mail@rkta.de> wrote:
> 
> > When you put the printf statement in a CGI script w3m will interpret the
> > output as a response from a webserver.  'w3m-control:' is a special
> > header line.  Header as in header send from a web server like
> > 'content-encoding'.  Only when it's seen in that context w3m will treat
> > it special.  In the former case it's just some text.
> 
> Thanks! Suspected something like this :)
> 
> It's probably not possible to simulate a header from a mapping? For
> instance, adding some, well, seasoning to the printf statement?

On Thu, Jan 02, 2025 at 03:01:13PM +0000, skejg wrote:
> On Thu Jan 2, 2025 at 10:28 +0100, Rene Kita <mail@rkta.de> wrote:
> 
> > You can access it from in the script.
> > [...]
> > W3M_CURRENT_LINK
> > [...]
> > W3M_URL
> 
> My question was not about accessing variables (I know I can access them
> with *SHELL commands, and, OK, I learned that I can't pass them with
> GOTO command) but more about why can't I trick w3m into executing
> a command via a mapping the way I can do via a script. E.g., why a
> binding like this doesn't work:

On Thu, Jan 02, 2025 at 07:18:10AM +0000, skejg wrote:
> 
> Hi!
> 
> 
> Just for the record, you seem can't have a mapping, which would end with
> a minus sign ('-'). A single minus sign works just fine, though.
> 
> To overcome that, you add one more minus to the mapping. E.g., to map
> the two key sequence `a-`, you create a mapping for `a--` instead.
> 
> Probably, w3m treats one of these minuses as a junction between keys (as
> it would be used in something like `C-a`), and therefore doesn't add it
> to an actual mapping. Some sort of a guard could probably solve that but

On Thu, Jan 02, 2025 at 06:35:16AM +0000, skejg wrote:
> 
> Hi!
> 
> 
> Let's say I have a cgi script, x.cgi, which takes an URL as a
> $QUERY_STRING, and then passes the text/html output to w3m. How do
> I map it?
> 
> I can't do this:
> 
>    keymap x GOTO file:///cgi-bin/x.cgi?$W3M_URL
> 
> or this:

On Sat, Dec 28, 2024 at 10:20:47PM +0100, Matěj Cepl wrote:
> On Sat Dec 28, 2024 at 7:36 PM CET, Rene Kita wrote:
> > Then I think reading from stdin should be the default for the script.
> > When I receive such an email I will pipe it.
> 
> The problem is that I don’t know how to make script check that
> there is something on stdin, and if there isn’t, then just go
> with command line arguments. We have to tell it explicitly that
> there is something on stdin.
> 
>     INSTR="$(cat)"
> 
> has a nasty habit of waiting forever for input.