~chambln/public-inbox

This thread contains a patchset. You're looking at the original emails, but you may wish to use the patch review UI. Review patch
2 2

[PATCH] lessgmi -- A self-contained pager for GMI files

Peter Marinov <pmar21@sonic.net>
Details
Message ID
<20210521032941.GA189533@presidio-x1>
DKIM signature
missing
Download raw message
Patch: +110 -8
* Implemented in the most-AWK-compatible script.

  Tested to work with `mawk` (default on Ubuntu) and `gawk` (GNU Awk
  installed separately)

* Implements wrapping of long lines

  It takes care of proper folding and formatting of quote sections
  (starting with "> ") and free text sections, keeps code sections
  verbatim (no folding applied)

* Placed the script as a HERE document for an entirely self-contained
  `lessgmi` pager
---
 contrib/lessgmi | 118 ++++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 110 insertions(+), 8 deletions(-)

diff --git a/contrib/lessgmi b/contrib/lessgmi
index 922cb95..b615eab 100755
--- a/contrib/lessgmi
+++ b/contrib/lessgmi
@@ -1,11 +1,113 @@
#!/bin/sh
#!/bin/sh -Ceu

# A wrapper around `less` that can color .gmi files

if command -v gawk >/dev/null 2>&1; then
    # TODO: Official install is probably /usr/share/gmi?
    gawk -f contrib/gmi_color_gawk.awk -- "$@"
else
    # TODO: fall back to plain `awk`, for the moment fall-back on plain `cat`
    cat -- "$@"
fi | less -r
# ANSI Escape Sequences at
# https://gist.github.com/fnky/458719343aabd01cfb17a3a4f7296797

AWK_GMI_SCRIPT=$(cat) <<'EOF'
function print_folded(long_line, prefix) {
  # Split words into an array
  split(long_line, list_of_words)

  line_len = 0
  num_words = length(list_of_words)

  for(i = 1; i <= num_words; i++) {
    # Check if we go beyond width of the block
    if ((line_len + length(list_of_words[i])) > 78) {
      line_len = 0
      printf("\n")
    }

    # Print prefix
    if (line_len == 0)
      printf(prefix)

    line_len += length(list_of_words[i]) + 1
    printf("%s ", list_of_words[i])
  }

  # Handle case of empty lines
  if (length(long_line) == 0)
    printf(prefix)

  # The line ends with a new-line
  printf("\n")
}

BEGIN {
  code_section = 0

  ANSI_NO_COLOR = "\033[0m"
  ANSI_URL1 = "\033[0;36m"
  ANSI_URL2 = "\033[0;35m\033[1m"
  ANSI_QUOTE = "\033[0;34m"
  ANSI_CODE = "\033[0;32m"
  ANSI_HEADING = "\033[0m\033[1m"
}

# URL
/^=> / {
  if (code_section == 0)
  {
    # Split line into URL elements
    split($0, e)

    # Print the elements with individual colors
    printf("%s%s %s%s,", ANSI_URL1, e[1], e[2], ANSI_NO_COLOR)
    printf("\n\t%s%s", ANSI_URL2, e[3])
    printf("%s\n", ANSI_NO_COLOR)
    next
  }
}

# Quote
/^>/ {
  if (code_section == 0)
  {
    # Strip leading "> "
    line = substr($0, 2)

    # Print folded block + setting color + using "> " as a line prefix
    print_folded(line, ANSI_QUOTE "> ")

    # Switch off color at the end of the quote block
    printf("%s", ANSI_NO_COLOR)
    next
  }
}

# Code section
/^```/ {
  if (code_section == 0)
      # Activate color for code section
      printf("%s```\n", ANSI_CODE)
  else
      # Remove color at the end of the code section
      printf("%s```%s\n", ANSI_CODE, ANSI_NO_COLOR)
  code_section = 1 - code_section
  next
}

# Heading
/^#.*/ {
  if (code_section == 0)
  {
    printf("%s%s%s\n", ANSI_HEADING, $0, ANSI_NO_COLOR)
    next
  }
}

# Everything else, the text of the file that is not prefixed by any markup
{
  if (code_section == 0)
    # Outside code section, fold long lines, prefix = switch off the color
    print_folded($0, ANSI_NO_COLOR)
  else
    # Inside code section, print line unmodified (no folding), apply color for code
    printf("%s%s\n", ANSI_CODE, $0)
}
EOF

awk "$AWK_GMI_SCRIPT" | less -r
-- 
2.25.1
Details
Message ID
<871ra065np.fsf@cosine.blue>
In-Reply-To
<20210521032941.GA189533@presidio-x1> (view parent)
DKIM signature
missing
Download raw message
Patch: +1 -1
Hi Peter,

> * Implemented in the most-AWK-compatible script.
> 
>   Tested to work with `mawk` (default on Ubuntu) and `gawk` (GNU Awk
>   installed separately)

Works fine for me with gawk and original-awk.  However, testing with
mawk 1.3.3 I get the following error:

    mawk: line 6: illegal reference to array list_of_words

Any idea why that might be?

> A PROBLEM:
> 
> `lessgmi` can only receive content via a pipe, I simply couldn't make
> it operate via a file from the command line, I hope you can make that
> work so it is a proper pager (I imagine people might use it to open
> local .gmi files)
> 
> Example:
> cat README.gmi | lessgmi  = Works
> lessgmi READM.gmi         = Doesn't

That can be done like so:

diff --git a/contrib/lessgmi b/contrib/lessgmi
index b615eab..98a351c 100755
--- a/contrib/lessgmi
+++ b/contrib/lessgmi
@@ -110,4 +110,4 @@ BEGIN {
}
EOF

awk "$AWK_GMI_SCRIPT" | less -r
awk -- "$AWK_GMI_SCRIPT" "$@" | less -r
--

Having such a large script in a heredoc feels wrong somehow, but at
least it’s all together in one executable.  I mean, it’s like 99% Awk
and 1% shell.  There is just one command:

   awk "$AWK_GMI_SCRIPT" | less -r

Surely it makes more sense to invoke less from inside Awk, if that’s
possible?

Also, I think we have our creative differences when it comes to the
colouring and typesetting of Gemtext.  And that’s fine.  We can
distribute a number of gemtext pagers along with gmi, and/or separately.
Users are at liberty to use whatever pager they like.

Going back to my earlier idea about the pager being a separate program
in its own right, I’m going to extract the shorter Awk script I wrote
out of gmi itself and into its own executable, akin to what you’ve done
here.  Then I’ll find a way to make it clear to users how to use their
preferred pager both within and without gmi.

Greg.
Peter Marinov <pmar21@sonic.net>
Details
Message ID
<20210522045030.GA192671@presidio-x1>
In-Reply-To
<871ra065np.fsf@cosine.blue> (view parent)
DKIM signature
missing
Download raw message
On Fri, May 21, 2021 at 06:02:18AM +0100, Gregory Chamberlain wrote:
> Works fine for me with gawk and original-awk.  However, testing with
> mawk 1.3.3 I get the following error:
> 
>     mawk: line 6: illegal reference to array list_of_words

Hello Greg,

I'm using v1.3.4 of `mawk`. I tried running it on my work machine too
(a Mac, with `brew install mawk`) and there was a problem with the
HERE document in the script.

What OS are you using?

Could you try the new patch I've sent you just minutes ago, it has a
split `lessgmi` and the awk script?


> > A PROBLEM:
> > 
> > `lessgmi` can only receive content via a pipe, I simply couldn't make
> > it operate via a file from the command line, I hope you can make that
> > work so it is a proper pager (I imagine people might use it to open
> > local .gmi files)
> > 
> > Example:
> > cat README.gmi | lessgmi  = Works
> > lessgmi READM.gmi         = Doesn't
> 
> That can be done like so:
> 
> diff --git a/contrib/lessgmi b/contrib/lessgmi
> index b615eab..98a351c 100755
> --- a/contrib/lessgmi
> +++ b/contrib/lessgmi
> @@ -110,4 +110,4 @@ BEGIN {
>  }
>  EOF
>  
> -awk "$AWK_GMI_SCRIPT" | less -r
> +awk -- "$AWK_GMI_SCRIPT" "$@" | less -r
> --

Thanks, this worked on Linux but failed on the Mac. I'm not sure why,
for one thing `bash` on the Mac is an ancient version. I don't know, I
plaied with it and couln't figure it out.


> Having such a large script in a heredoc feels wrong somehow, but at
> least it’s all together in one executable.  I mean, it’s like 99% Awk
> and 1% shell.  There is just one command:
> 
>     awk "$AWK_GMI_SCRIPT" | less -r
> 
> Surely it makes more sense to invoke less from inside Awk, if that’s
> possible?

I agree, the HERE document was too much. I've split the script as before.

I think you can invoke external programs from awk but I don't see how
you can pipe into `less` with this method, I only imagine if that goes
via a temp file, I feel that is less elegant.

I've used your technique from pandoc-rss about how to reach the
auxiliary awk script from `lessgmi`, if you add this to the `make
install` it should work from ~/.local or /usr or any other PREFIX


> Also, I think we have our creative differences when it comes to the
> colouring and typesetting of Gemtext.  And that’s fine.  We can
> distribute a number of gemtext pagers along with gmi, and/or separately.
> Users are at liberty to use whatever pager they like.

So, I agree that the colors are an individual thing, that's why I
added a way for it to be configured via an environment variable or it
falls back on the default colors.

To try it use this as a starting point in your shell (bash) prompt:
export GMI_COLORS="*e[0;36m *e[0;35m*e[1m *e[0;34m *e[0;32m *e[0m*e[1m"

 
> Going back to my earlier idea about the pager being a separate program
> in its own right, I’m going to extract the shorter Awk script I wrote
> out of gmi itself and into its own executable, akin to what you’ve done
> here.  Then I’ll find a way to make it clear to users how to use their
> preferred pager both within and without gmi.

I think the overall structure is your call. The idea of an individual
pager `lessgmi` is useful standalone or invoked from `gmi` (by default
or via a setting)

--pe
https://hangar118.sdf.org
Reply to thread Export thread (mbox)