emacs 30.5.0 editing epub

emacs 30.5.0 editing epub

[H.-J. Heitländer]

Hi there,

trying to get my feet wet with emacs and lisp. ... well it seems I am 
drowning...

The problem:

I am editing epub archives and  any saves that I am doing result in

<snip>

***backtrace***

Debugger entered--Lisp error: (wrong-type-argument stringp nil)

doc-view--revert-buffer(#f(compiled-function (&rest args) #<bytecode 
-0x1d096aa50772e2b6>) t t)

apply(doc-view--revert-buffer #f(compiled-function (&rest args) 
#<bytecode -0x1d096aa50772e2b6>) (t t))

#f(advice doc-view--revert-buffer :around #f(compiled-function (&rest 
args) #<bytecode -0x1d096aa50772e2b6>))(t t)

archive--mode-revert(#f(advice doc-view--revert-buffer :around 
#f(compiled-function (&rest args) #<bytecode -0x1d096aa50772e2b6>)) nil nil)

apply(archive--mode-revert #f(advice doc-view--revert-buffer :around 
#f(compiled-function (&rest args) #<bytecode -0x1d096aa50772e2b6>)) (nil 
nil))

#f(advice archive--mode-revert :around #f(advice doc-view--revert-buffer 
:around #f(compiled-function (&rest args) #<bytecode 
-0x1d096aa50772e2b6>)))(nil nil)

revert-buffer()

archive-write-file-member()

run-hook-with-args-until-success(archive-write-file-member)

basic-save-buffer(t)

save-buffer(1)

funcall-interactively(save-buffer 1)

command-execute(save-buffer)

</snip>


For the moment I can live with that as in reality the epub is being saved.

My problem is that from that moment on the display of the epub subfile 
(<name>.html) loses the utf-8 display. (oh, and every other file is 
being displayed in the same - wrong - way.)

The content is shown as (example)

<snip>

\342\200\234Nicely done. Now let\342\200\231s get that equipment on 
board.\342\200\234 </p>

</snip>

It should have been shown as

<snip>

"Nicely done. Now let's get that equipment on board." </p>

</snip>

I searched the internet without any result. Don't know which part of 
documentation to read.

Question: How do I switch the display back to show the contents in the 
expected way???


TX

Heiner

Re: emacs 30.5.0 editing epub

[Michael Heerdegen]

H.-J. Heitländer <Heiner.Heitlaender@posteo.de> writes:

> The problem:
>
> I am editing epub archives and  any saves that I am doing result in
>
> <snip>
>
> ***backtrace***
>
> Debugger entered--Lisp error: (wrong-type-argument stringp nil)
>
> doc-view--revert-buffer(#f(compiled-function (&rest args) #<bytecode
> -0x1d096aa50772e2b6>) t t)

No idea why this is happening... but your backtrace is not very
informative because it was created using compiled Elisp code.

Could you please try to recreate a backtrace using interpreted Elisp?
The only thing you need to do is to load the .el files of the according
libraries ("doc-view.el" and "files.el" should suffice) and then do the
same as before.

TIA,

Michael.

Re: emacs 30.5.0 editing epub

[H.-J. Heitländer]

Hi Michael,

thanks for getting involved.

Nevertheless I would be quite content with a temporary solution by being 
able to get emacs to redisplay in the correct form.


Thank you

Heiner


Appended backtrace:

load-file: arc-mode.el, doc-view.el, files.el

here we go:

<snip>

Debugger entered--Lisp error: (wrong-type-argument stringp nil)

write-region(nil nil nil)

(if (equal buffer-file-name doc-view--buffer-file-name) nil 
(doc-view-make-safe-dir doc-view-cache-directory) (write-region nil nil 
doc-view--buffer-file-name))

(let ((revert-buffer-preserve-modes t)) (apply orig-fun args) (if (equal 
buffer-file-name doc-view--buffer-file-name) nil (doc-view-make-safe-dir 
doc-view-cache-directory) (write-region nil nil 
doc-view--buffer-file-name)))

(closure ((args t t) (orig-fun . #f(compiled-function (&rest args) 
#<bytecode -0x1d33bbea90aba2b6>)) revert-buffer-preserve-modes) nil (let 
((revert-buffer-preserve-modes t)) (apply orig-fun args) (if (equal 
buffer-file-name doc-view--buffer-file-name) nil (doc-view-make-safe-dir 
doc-view-cache-directory) (write-region nil nil 
doc-view--buffer-file-name))))()

funcall((closure ((args t t) (orig-fun . #f(compiled-function (&rest 
args) #<bytecode -0x1d33bbea90aba2b6>)) revert-buffer-preserve-modes) 
nil (let ((revert-buffer-preserve-modes t)) (apply orig-fun args) (if 
(equal buffer-file-name doc-view--buffer-file-name) nil 
(doc-view-make-safe-dir doc-view-cache-directory) (write-region nil nil 
doc-view--buffer-file-name)))))

(if (and (eq 'pdf doc-view-doc-type) (executable-find "pdfinfo")) (if (= 
0 (call-process "pdfinfo" nil nil nil doc-view--buffer-file-name)) 
(funcall --cl-revert--) (if (called-interactively-p 'interactive) (progn 
(message "Can't revert right now because the file is corrupt...")))) 
(funcall --cl-revert--))

(let* ((--cl-revert-- #'(lambda nil (let ((revert-buffer-preserve-modes 
t)) (apply orig-fun args) (if (equal buffer-file-name 
doc-view--buffer-file-name) nil (doc-view-make-safe-dir 
doc-view-cache-directory) (write-region nil nil 
doc-view--buffer-file-name)))))) (if (and (eq 'pdf doc-view-doc-type) 
(executable-find "pdfinfo")) (if (= 0 (call-process "pdfinfo" nil nil 
nil doc-view--buffer-file-name)) (funcall --cl-revert--) (if 
(called-interactively-p 'interactive) (progn (message "Can't revert 
right now because the file is corrupt...")))) (funcall --cl-revert--)))

doc-view--revert-buffer(#f(compiled-function (&rest args) #<bytecode 
-0x1d33bbea90aba2b6>) t t)

apply(doc-view--revert-buffer #f(compiled-function (&rest args) 
#<bytecode -0x1d33bbea90aba2b6>) (t t))

#f(advice doc-view--revert-buffer :around #f(compiled-function (&rest 
args) #<bytecode -0x1d33bbea90aba2b6>))(t t)

apply(#f(advice doc-view--revert-buffer :around #f(compiled-function 
(&rest args) #<bytecode -0x1d33bbea90aba2b6>)) t t nil)

(let ((coding-system-for-read 'no-conversion)) (apply orig-fun t t (cdr 
(cdr args))))

(let ((no (archive-get-lineno))) (setq archive-files nil) (let 
((coding-system-for-read 'no-conversion)) (apply orig-fun t t (cdr (cdr 
args)))) (archive-mode) (goto-char archive-file-list-start) 
(archive-next-line no))

archive--mode-revert(#f(advice doc-view--revert-buffer :around 
#f(compiled-function (&rest args) #<bytecode -0x1d33bbea90aba2b6>)) nil nil)

apply(archive--mode-revert #f(advice doc-view--revert-buffer :around 
#f(compiled-function (&rest args) #<bytecode -0x1d33bbea90aba2b6>)) (nil 
nil))

#f(advice archive--mode-revert :around #f(advice doc-view--revert-buffer 
:around #f(compiled-function (&rest args) #<bytecode 
-0x1d33bbea90aba2b6>)))(nil nil)

funcall(#f(advice archive--mode-revert :around #f(advice 
doc-view--revert-buffer :around #f(compiled-function (&rest args) 
#<bytecode -0x1d33bbea90aba2b6>))) nil nil)

(let ((revert-buffer-in-progress-p t) (revert-buffer-preserve-modes 
preserve-modes) (state (and (boundp 'read-only-mode--state) (list 
read-only-mode--state)))) (funcall (or revert-buffer-function 
#'revert-buffer--default) ignore-auto noconfirm) (if state (progn (setq 
buffer-read-only (car state)) (set (make-local-variable 
'read-only-mode--state) (car state)))))

revert-buffer()

(if (not archive-remote) (revert-buffer) (archive-maybe-update nil))

(save-restriction (message "Updating archive...") (widen) (let ((writer 
(save-current-buffer (set-buffer archive-superior-buffer) (archive-name 
"write-file-member"))) (archive (save-current-buffer (set-buffer 
archive-superior-buffer) (archive-maybe-copy (buffer-file-name))))) (if 
(fboundp writer) (funcall writer archive archive-subfile-mode) 
(archive-*-write-file-member archive archive-subfile-mode (symbol-value 
writer))) (set-buffer-modified-p nil) (message "Updating 
archive...done")) (set-buffer archive-superior-buffer) (if (not 
archive-remote) (revert-buffer) (archive-maybe-update nil)))

(save-excursion (save-restriction (message "Updating archive...") 
(widen) (let ((writer (save-current-buffer (set-buffer 
archive-superior-buffer) (archive-name "write-file-member"))) (archive 
(save-current-buffer (set-buffer archive-superior-buffer) 
(archive-maybe-copy (buffer-file-name))))) (if (fboundp writer) (funcall 
writer archive archive-subfile-mode) (archive-*-write-file-member 
archive archive-subfile-mode (symbol-value writer))) 
(set-buffer-modified-p nil) (message "Updating archive...done")) 
(set-buffer archive-superior-buffer) (if (not archive-remote) 
(revert-buffer) (archive-maybe-update nil))))

archive-write-file-member()

run-hook-with-args-until-success(archive-write-file-member)

(or (run-hook-with-args-until-success 'local-write-file-hooks) 
(run-hook-with-args-until-success 'write-file-functions) (let ((dir 
(file-name-directory (expand-file-name buffer-file-name)))) (if 
(file-exists-p dir) nil (if (y-or-n-p (format-message "Directory `%s' 
does not exist; create? " dir)) (make-directory dir t) (error 
"Canceled"))) (setq setmodes (basic-save-buffer-1))))

(if (run-hook-with-args-until-success 'write-contents-functions) nil (or 
buffer-file-name (let ((filename (expand-file-name (read-file-name "File 
to save in: " nil (expand-file-name ...))))) (if (file-exists-p 
filename) (if (file-directory-p filename) (error "%s is a directory" 
filename) (if (y-or-n-p (format-message "File `%s' exists; overwrite? " 
filename)) nil (error "Canceled")))) (set-visited-file-name filename))) 
(vc-before-save) (or (run-hook-with-args-until-success 
'local-write-file-hooks) (run-hook-with-args-until-success 
'write-file-functions) (let ((dir (file-name-directory (expand-file-name 
buffer-file-name)))) (if (file-exists-p dir) nil (if (y-or-n-p 
(format-message "Directory `%s' does not exist; create? " dir)) 
(make-directory dir t) (error "Canceled"))) (setq setmodes 
(basic-save-buffer-1)))))

(save-restriction (widen) (save-excursion (and (> (point-max) 
(point-min)) (not find-file-literally) (null buffer-read-only) (/= 
(char-after (1- (point-max))) 10) (not (and (eq selective-display t) (= 
(char-after (1- ...)) 13))) (or (eq require-final-newline t) (eq 
require-final-newline 'visit-save) (and require-final-newline (y-or-n-p 
(format "Buffer %s does not end in newline. Add one? " (buffer-name))))) 
(save-excursion (goto-char (point-max)) (insert 10)))) (condition-case 
err (run-hooks 'before-save-hook) ((debug error) (message "Before-save 
hook error: %S" err) nil)) (if (run-hook-with-args-until-success 
'write-contents-functions) nil (or buffer-file-name (let ((filename 
(expand-file-name (read-file-name "File to save in: " nil ...)))) (if 
(file-exists-p filename) (if (file-directory-p filename) (error "%s is a 
directory" filename) (if (y-or-n-p ...) nil (error "Canceled")))) 
(set-visited-file-name filename))) (vc-before-save) (or 
(run-hook-with-args-until-success 'local-write-file-hooks) 
(run-hook-with-args-until-success 'write-file-functions) (let ((dir 
(file-name-directory (expand-file-name buffer-file-name)))) (if 
(file-exists-p dir) nil (if (y-or-n-p (format-message "Directory `%s' 
does not exist; create? " dir)) (make-directory dir t) (error 
"Canceled"))) (setq setmodes (basic-save-buffer-1))))) (if 
buffer-file-name (progn (if save-buffer-coding-system (setq 
save-buffer-coding-system last-coding-system-used) (setq 
buffer-file-coding-system last-coding-system-used)) (setq 
buffer-file-number (file-attribute-file-identifier (file-attributes 
buffer-file-name))) (if setmodes (condition-case nil (progn (if 
(condition-case err ... ...) nil (set-file-extended-attributes 
buffer-file-name ...))) (error nil))) (vc-after-save))) 
(delete-auto-save-file-if-necessary recent-save))

(let ((recent-save (recent-auto-save-p)) setmodes) (or (null 
buffer-file-name) (verify-visited-file-modtime (current-buffer)) (not 
(file-exists-p buffer-file-name)) (yes-or-no-p (format "%s has changed 
since visited or saved. Save anywa..." (file-name-nondirectory 
buffer-file-name))) (user-error "Save not confirmed")) (save-restriction 
(widen) (save-excursion (and (> (point-max) (point-min)) (not 
find-file-literally) (null buffer-read-only) (/= (char-after (1- 
(point-max))) 10) (not (and (eq selective-display t) (= (char-after ...) 
13))) (or (eq require-final-newline t) (eq require-final-newline 
'visit-save) (and require-final-newline (y-or-n-p (format "Buffer %s 
does not end in newline. Add one? " ...)))) (save-excursion (goto-char 
(point-max)) (insert 10)))) (condition-case err (run-hooks 
'before-save-hook) ((debug error) (message "Before-save hook error: %S" 
err) nil)) (if (run-hook-with-args-until-success 
'write-contents-functions) nil (or buffer-file-name (let ((filename 
(expand-file-name ...))) (if (file-exists-p filename) (if 
(file-directory-p filename) (error "%s is a directory" filename) (if ... 
nil ...))) (set-visited-file-name filename))) (vc-before-save) (or 
(run-hook-with-args-until-success 'local-write-file-hooks) 
(run-hook-with-args-until-success 'write-file-functions) (let ((dir 
(file-name-directory ...))) (if (file-exists-p dir) nil (if (y-or-n-p 
...) (make-directory dir t) (error "Canceled"))) (setq setmodes 
(basic-save-buffer-1))))) (if buffer-file-name (progn (if 
save-buffer-coding-system (setq save-buffer-coding-system 
last-coding-system-used) (setq buffer-file-coding-system 
last-coding-system-used)) (setq buffer-file-number 
(file-attribute-file-identifier (file-attributes buffer-file-name))) (if 
setmodes (condition-case nil (progn (if ... nil ...)) (error nil))) 
(vc-after-save))) (delete-auto-save-file-if-necessary recent-save)) 
(run-hooks 'after-save-hook))

(if (or (buffer-modified-p) (and buffer-file-name (not (file-exists-p 
buffer-file-name)))) (let ((recent-save (recent-auto-save-p)) setmodes) 
(or (null buffer-file-name) (verify-visited-file-modtime 
(current-buffer)) (not (file-exists-p buffer-file-name)) (yes-or-no-p 
(format "%s has changed since visited or saved. Save anywa..." 
(file-name-nondirectory buffer-file-name))) (user-error "Save not 
confirmed")) (save-restriction (widen) (save-excursion (and (> 
(point-max) (point-min)) (not find-file-literally) (null 
buffer-read-only) (/= (char-after (1- ...)) 10) (not (and (eq 
selective-display t) (= ... 13))) (or (eq require-final-newline t) (eq 
require-final-newline 'visit-save) (and require-final-newline (y-or-n-p 
...))) (save-excursion (goto-char (point-max)) (insert 10)))) 
(condition-case err (run-hooks 'before-save-hook) ((debug error) 
(message "Before-save hook error: %S" err) nil)) (if 
(run-hook-with-args-until-success 'write-contents-functions) nil (or 
buffer-file-name (let ((filename ...)) (if (file-exists-p filename) (if 
... ... ...)) (set-visited-file-name filename))) (vc-before-save) (or 
(run-hook-with-args-until-success 'local-write-file-hooks) 
(run-hook-with-args-until-success 'write-file-functions) (let ((dir 
...)) (if (file-exists-p dir) nil (if ... ... ...)) (setq setmodes 
(basic-save-buffer-1))))) (if buffer-file-name (progn (if 
save-buffer-coding-system (setq save-buffer-coding-system 
last-coding-system-used) (setq buffer-file-coding-system 
last-coding-system-used)) (setq buffer-file-number 
(file-attribute-file-identifier (file-attributes buffer-file-name))) (if 
setmodes (condition-case nil (progn ...) (error nil))) (vc-after-save))) 
(delete-auto-save-file-if-necessary recent-save)) (run-hooks 
'after-save-hook)) (or noninteractive (not called-interactively) 
(files--message "(No changes need to be saved)")))

(save-current-buffer (if (buffer-base-buffer) (set-buffer 
(buffer-base-buffer))) (if (or (buffer-modified-p) (and buffer-file-name 
(not (file-exists-p buffer-file-name)))) (let ((recent-save 
(recent-auto-save-p)) setmodes) (or (null buffer-file-name) 
(verify-visited-file-modtime (current-buffer)) (not (file-exists-p 
buffer-file-name)) (yes-or-no-p (format "%s has changed since visited or 
saved. Save anywa..." (file-name-nondirectory buffer-file-name))) 
(user-error "Save not confirmed")) (save-restriction (widen) 
(save-excursion (and (> (point-max) (point-min)) (not 
find-file-literally) (null buffer-read-only) (/= (char-after ...) 10) 
(not (and ... ...)) (or (eq require-final-newline t) (eq 
require-final-newline ...) (and require-final-newline ...)) 
(save-excursion (goto-char ...) (insert 10)))) (condition-case err 
(run-hooks 'before-save-hook) ((debug error) (message "Before-save hook 
error: %S" err) nil)) (if (run-hook-with-args-until-success 
'write-contents-functions) nil (or buffer-file-name (let (...) (if ... 
...) (set-visited-file-name filename))) (vc-before-save) (or 
(run-hook-with-args-until-success 'local-write-file-hooks) 
(run-hook-with-args-until-success 'write-file-functions) (let (...) (if 
... nil ...) (setq setmodes ...)))) (if buffer-file-name (progn (if 
save-buffer-coding-system (setq save-buffer-coding-system 
last-coding-system-used) (setq buffer-file-coding-system 
last-coding-system-used)) (setq buffer-file-number 
(file-attribute-file-identifier ...)) (if setmodes (condition-case nil 
... ...)) (vc-after-save))) (delete-auto-save-file-if-necessary 
recent-save)) (run-hooks 'after-save-hook)) (or noninteractive (not 
called-interactively) (files--message "(No changes need to be saved)"))))

basic-save-buffer(t)

(let ((modp (buffer-modified-p)) (make-backup-files (or (and 
make-backup-files (not (eq arg 0))) (memq arg '(16 64))))) (and modp 
(memq arg '(16 64)) (setq buffer-backed-up nil)) (if (and modp 
(buffer-file-name) (not noninteractive) (not save-silently)) (message 
"Saving file %s..." (buffer-file-name))) (basic-save-buffer 
(called-interactively-p 'any)) (and modp (memq arg '(4 64)) (setq 
buffer-backed-up nil)))

save-buffer(1)

funcall-interactively(save-buffer 1)

command-execute(save-buffer)

</snip>






Am 16.03.23 um 06:05 schrieb Michael Heerdegen:
> H.-J. Heitländer<Heiner.Heitlaender@posteo.de>  writes:
>
>> The problem:
>>
>> I am editing epub archives and  any saves that I am doing result in
>>
>> <snip>
>>
>> ***backtrace***
>>
>> Debugger entered--Lisp error: (wrong-type-argument stringp nil)
>>
>> doc-view--revert-buffer(#f(compiled-function (&rest args) #<bytecode
>> -0x1d096aa50772e2b6>) t t)
> No idea why this is happening... but your backtrace is not very
> informative because it was created using compiled Elisp code.
>
> Could you please try to recreate a backtrace using interpreted Elisp?
> The only thing you need to do is to load the .el files of the according
> libraries ("doc-view.el" and "files.el" should suffice) and then do the
> same as before.
>
> TIA,
>
> Michael.
>
>

Re: emacs 30.5.0 editing epub

[Stephen Berman]

On Wed, 15 Mar 2023 18:35:55 +0000 H.-J. Heitländer <Heiner.Heitlaender@posteo.de> wrote:

> Hi there,
>
> trying to get my feet wet with emacs and lisp. ... well it seems I am
> drowning...
>
> The problem:
>
> I am editing epub archives and  any saves that I am doing result in
>
> <snip>
>
> ***backtrace***
[...]
> For the moment I can live with that as in reality the epub is being saved.
>
> My problem is that from that moment on the display of the epub subfile
> (<name>.html) loses the utf-8 display. (oh, and every other file is being
> displayed in the same - wrong - way.)
>
> The content is shown as (example)
>
> <snip>
>
> \342\200\234Nicely done. Now let\342\200\231s get that equipment on
> board.\342\200\234 </p>
>
> </snip>
>
> It should have been shown as
>
> <snip>
>
> "Nicely done. Now let's get that equipment on board." </p>
>
> </snip>
>
> I searched the internet without any result. Don't know which part of
> documentation to read.
>
> Question: How do I switch the display back to show the contents in the
> expected way???

It looks like the buffer is being displayed in the raw-text coding
system (is the first character in the mode line "t"?).  \342\200\234 is
the raw-byte sequence of the unicode character LEFT DOUBLE QUOTATION
MARK (#x201c) and \342\200\231 of RIGHT SINGLE QUOTATION MARK (#x2019).
Does typing `C-x RET r' and at the prompt entering `utf-8' and at the
next prompt `yes' fix the display?

Steve Berman

Re: emacs 30.5.0 editing epub

[H.-J. Heitländer]

Hi Stephen,

mode line displays

<snip>

=:**-

</snip>

When entering the command C-x RET r ... utf-8  the "answer" is

"wrong type argument: strongp, nil".


Testcase 2)

kill buffer up to the top epub

reopen epub

open epub subfile (...html)

C-x RET r ... utf-8

in minibufer: Revert buffer from file xxx.html? (y or n)

Answer: y

Minibuffer: Cannot revert noexistent file xxx.html



Other directory buffers display with "coding system (default nil)" as 
soon as I perform a refresh of the directory display with "g". To cure 
this phenomen I have to kill / restart emacs.

So no easy solution / workaround at the moment.

TX
Heiner



Am 16.03.23 um 10:22 schrieb Stephen Berman:
> On Wed, 15 Mar 2023 18:35:55 +0000 H.-J. Heitländer<Heiner.Heitlaender@posteo.de>  wrote:
>
>> Hi there,
>>
>> trying to get my feet wet with emacs and lisp. ... well it seems I am
>> drowning...
>>
>> The problem:
>>
>> I am editing epub archives and  any saves that I am doing result in
>>
>> <snip>
>>
>> ***backtrace***
> [...]
>> For the moment I can live with that as in reality the epub is being saved.
>>
>> My problem is that from that moment on the display of the epub subfile
>> (<name>.html) loses the utf-8 display. (oh, and every other file is being
>> displayed in the same - wrong - way.)
>>
>> The content is shown as (example)
>>
>> <snip>
>>
>> \342\200\234Nicely done. Now let\342\200\231s get that equipment on
>> board.\342\200\234 </p>
>>
>> </snip>
>>
>> It should have been shown as
>>
>> <snip>
>>
>> "Nicely done. Now let's get that equipment on board." </p>
>>
>> </snip>
>>
>> I searched the internet without any result. Don't know which part of
>> documentation to read.
>>
>> Question: How do I switch the display back to show the contents in the
>> expected way???
> It looks like the buffer is being displayed in the raw-text coding
> system (is the first character in the mode line "t"?).  \342\200\234 is
> the raw-byte sequence of the unicode character LEFT DOUBLE QUOTATION
> MARK (#x201c) and \342\200\231 of RIGHT SINGLE QUOTATION MARK (#x2019).
> Does typing `C-x RET r' and at the prompt entering `utf-8' and at the
> next prompt `yes' fix the display?
>
> Steve Berman

Re: emacs 30.5.0 editing epub

[Yuri Khan]

On Thu, 16 Mar 2023 at 20:17, H.-J. Heitländer
<Heiner.Heitlaender@posteo.de> wrote:

> kill buffer up to the top epub
>
> reopen epub
>
> open epub subfile (...html)
>
> C-x RET r ... utf-8
>
> in minibufer: Revert buffer from file xxx.html? (y or n)
>
> Answer: y
>
> Minibuffer: Cannot revert noexistent file xxx.html

You keep mentioning “editing epub”. Are you even sure it’s supposed to work?

An EPUB file is a zip archive, typically containing XHTML pages and
some metadata. Emacs covers some basic scenarios where you can browse
the archive, visit files inside, even edit and save them (and they get
re-compressed and updated in the archive).

Reverting, on the other hand, does not work, probably because of the
way archive-mode is implemented. So when it picks the wrong encoding,
you cannot fix it by reverting.

You could probably achieve better results if you first extract the
contents of the EPUB archive into a real directory on your file
system, then edit files there. If necessary, re-pack the modified
files.

I see ‘archive--extract-file’ initially sets the coding system to
‘no-conversion’, possibly because it needs that for decompression. It
is supposed to re-decide on a coding system later in
‘archive-set-buffer-as-visiting-file’, but for some reason it keeps
‘no-conversion’ in your case.

Re: emacs 30.5.0 editing epub

[H.-J. Heitländer]

Hi Yuri,

yes I keep mentioning "editing EPUB" because

a) that's my primary use case at the moment (in combination with 
"calibre") and

b) it worked some time ago. (emacs 29.5 and earlier as far as I remember)

Even in this buggy scenario it is much more comfortable for me to handle 
EPUBs this way  than doing it by extracting the zip file and rezipping 
it. Especially as I lack an "idiot-safe" workflow for unzipping / rezipping.

So I shall live with this situation because I can't correct it.

| I see ‘archive--extract-file’ initially sets the coding system to
| ‘no-conversion’, possibly because it needs that for decompression. It
| is supposed to re-decide on a coding system later in
| ‘archive-set-buffer-as-visiting-file’, but for some reason it keeps
| ‘no-conversion’ in your case.

There seems to exist a bug in the migration path somewhere from version 
29 to 30. Where do I address the bug report; as I am definetely too dumb 
for debugging it myself?

Cheers and thanks for your patience

Heiner



Am 16.03.23 um 16:23 schrieb Yuri Khan:
> On Thu, 16 Mar 2023 at 20:17, H.-J. Heitländer
> <Heiner.Heitlaender@posteo.de> wrote:
>
>> kill buffer up to the top epub
>>
>> reopen epub
>>
>> open epub subfile (...html)
>>
>> C-x RET r ... utf-8
>>
>> in minibufer: Revert buffer from file xxx.html? (y or n)
>>
>> Answer: y
>>
>> Minibuffer: Cannot revert noexistent file xxx.html
> You keep mentioning “editing epub”. Are you even sure it’s supposed to work?
>
> An EPUB file is a zip archive, typically containing XHTML pages and
> some metadata. Emacs covers some basic scenarios where you can browse
> the archive, visit files inside, even edit and save them (and they get
> re-compressed and updated in the archive).
>
> Reverting, on the other hand, does not work, probably because of the
> way archive-mode is implemented. So when it picks the wrong encoding,
> you cannot fix it by reverting.
>
> You could probably achieve better results if you first extract the
> contents of the EPUB archive into a real directory on your file
> system, then edit files there. If necessary, re-pack the modified
> files.
>
> I see ‘archive--extract-file’ initially sets the coding system to
> ‘no-conversion’, possibly because it needs that for decompression. It
> is supposed to re-decide on a coding system later in
> ‘archive-set-buffer-as-visiting-file’, but for some reason it keeps
> ‘no-conversion’ in your case.
>

Re: emacs 30.5.0 editing epub

[Stefan Monnier via Users list for the GNU Emacs text editor]

H.-J. Heitländer [2023-03-16 16:54:03] wrote:
> Where do I address the bug report;

Help => How to Report a Bug


        Stefan

Re: emacs 30.5.0 editing epub

[Michael Heerdegen]

H.-J. Heitländer <Heiner.Heitlaender@posteo.de> writes:

> Nevertheless I would be quite content with a temporary solution by
> being able to get emacs to redisplay in the correct form.

Please report this as a bug.  Finding a temporary solution might not be
easier than finding out what happens.

All you need to do is either use M-x report-emacs-bug as described in
(info "(emacs) Checklist") (a subnode of the info page that Stefan's
suggested menu item leads to), or, if you don't yet have/ don't want to set
up sending mails from inside Emacs, just write a mail like this to

  bug-gnu-emacs@gnu.org

and include your backtrace (thanks for it) and a short description of
what you did (most importantly, a short recipe: in particular, what
buffer was current when you tried to save).  A few lines suffice.

> archive--mode-revert(#f(advice doc-view--revert-buffer :around
> #f(compiled-function (&rest args) #<bytecode -0x1d33bbea90aba2b6>))
> nil nil)

This looks odd: the two advices `archive--mode-revert' and
`doc-view--revert-buffer' should not both be installed in a buffer at
the same time.


Michael.

Re: emacs 30.5.0 editing epub

[Michael Heerdegen]

H.-J. Heitländer <Heiner.Heitlaender@posteo.de> writes:

> Hi Michael,
>
> thanks for getting involved.
>
> Nevertheless I would be quite content with a temporary solution by
> being able to get emacs to redisplay in the correct form.

BTW, I managed to copy an pdf-File into an epub Archive.  Opened the
archive and opened the pdf using doc-view.  I was not able to provoke an
error.  Saving worked successfully however I tried.

Michael.

editing a PDF [Re: emacs 30.5.0 editing epub]

[gebser]

On 3/16/23 9:13 PM, Michael Heerdegen wrote:
> I managed to copy an pdf-File into an epub Archive.  Opened the
> archive and opened the pdf using doc-view.  I was not able to provoke an
> error.  Saving worked successfully however I tried.
> 
> Michael.

How does one edit a PDF file in emacs?  When I downloaded 
https://www.gnu.org.ua/software/direvent/manual/direvent.pdf, loaded it 
into emacs, and tried to edit it, I could see it it and read it, but I 
couldn't even highlight a part of the text so I could copy it... nor 
could I make any changes to it.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Michael Heerdegen]

"gebser@mousecar.com" <gebser@mousecar.com> writes:

> How does one edit a PDF file in emacs?  When I downloaded
> https://www.gnu.org.ua/software/direvent/manual/direvent.pdf, loaded
> it into emacs, and tried to edit it, I could see it it and read it,
> but I couldn't even highlight a part of the text so I could copy
> it... nor could I make any changes to it.

I don't think you want to edit pdf in Emacs: the file contents are not
human readable, and Emacs doesn't provide features to allow graphical
editing like those pdf editors (at least not built in).

epub files can contain human readable parts like html, however.

What I had written:

>> I managed to copy an pdf-File into an epub Archive.  Opened the
>> archive and opened the pdf using doc-view.  I was not able to provoke an
>> error.  Saving worked successfully however I tried.

was only happening because I was trying around how to provoke the
reported error.

Michael.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[gebser]

On 3/19/23 7:59 PM, Michael Heerdegen wrote:
> "gebser@mousecar.com" <gebser@mousecar.com> writes:
> 
>> How does one edit a PDF file in emacs?  When I downloaded
>> https://www.gnu.org.ua/software/direvent/manual/direvent.pdf, loaded
>> it into emacs, and tried to edit it, I could see it it and read it,
>> but I couldn't even highlight a part of the text so I could copy
>> it... nor could I make any changes to it.
> 
> I don't think you want to edit pdf in Emacs: the file contents are not
> human readable, and Emacs doesn't provide features to allow graphical
> editing like those pdf editors (at least not built in).
> 
> epub files can contain human readable parts like html, however.
> 
> What I had written:
> 
>>> I managed to copy an pdf-File into an epub Archive.  Opened the
>>> archive and opened the pdf using doc-view.  I was not able to provoke an
>>> error.  Saving worked successfully however I tried.
> 
> was only happening because I was trying around how to provoke the
> reported error.
> 
> Michael.

Yep, of course that was the paragraph I was reading too.

It is kinda weird that, with all the many things that emacs can do, it 
can't take the info from doc-view (which obviously understands all the 
pieces of a pdf-- down to its bits-- and how they all go together to 
make a document) and edit it... it's pretty much implausible to believe 
emacs *can't* do that.  But then, I've always thought reality was 
completely implausible.  :^/

Re: emacs 30.5.0 editing epub - finishing remark

[H.-J. Heitländer]

1) debug report has been sent

2) Michael has developed a patch which was near to perfection. With 
minor change it fitted (doc-view version with different LOC)

3) Problem fixed

Thanks to everyone

Heiner

Am 16.03.23 um 19:58 schrieb Stefan Monnier via Users list for the GNU 
Emacs text editor:
> H.-J. Heitländer [2023-03-16 16:54:03] wrote:
>> Where do I address the bug report;
> Help => How to Report a Bug
>
>
>          Stefan
>
>

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Michael Heerdegen]

"gebser@mousecar.com" <gebser@mousecar.com> writes:

> It is kinda weird that, with all the many things that emacs can do, it
> can't take the info from doc-view (which obviously understands all the
> pieces of a pdf-- down to its bits-- and how they all go together to
> make a document) and edit it...

AFAIU it understands nothing at all - it just calls an external
converter to generate a set of png images it then displays.

> it's pretty much implausible to believe emacs *can't* do that.

I think if someone wants to work on something like that the result would
be welcome.  But this probably would have to be done from scratch, more
or less, and not be the smallest effort.

Michael.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Yuri Khan]

On Mon, 20 Mar 2023 at 14:16, gebser@mousecar.com <gebser@mousecar.com> wrote:

> It is kinda weird that, with all the many things that emacs can do, it
> can't take the info from doc-view (which obviously understands all the
> pieces of a pdf-- down to its bits-- and how they all go together to
> make a document) and edit it... it's pretty much implausible to believe
> emacs *can't* do that.  But then, I've always thought reality was
> completely implausible.  :^/

PDF is not really meant for editing. It’s not even a data format.
Rather, it’s an executable program that has instructions like “select
this font” and “display this word in the selected font at this
position on the page” and “make a new page”.

You don’t normally edit executable programs, you compile them from
source. In the same vein, to get a modified PDF, you find the source
document from which it was produced, modify that, and re-export.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Emanuel Berg]

Yuri Khan wrote:

>> It is kinda weird that, with all the many things that emacs
>> can do, it can't take the info from doc-view (which
>> obviously understands all the pieces of a pdf-- down to its
>> bits-- and how they all go together to make a document) and
>> edit it... it's pretty much implausible to believe emacs
>> *can't* do that. But then, I've always thought reality was
>> completely implausible. :^/
>
> PDF is not really meant for editing. It’s not even a data
> format. Rather, it’s an executable program that has
> instructions like “select this font” and “display this word
> in the selected font at this position on the page” and “make
> a new page”.
>
> You don’t normally edit executable programs, you compile
> them from source. In the same vein, to get a modified PDF,
> you find the source document from which it was produced,
> modify that, and re-export.

Still, they are editable at/from the PDF level as well, for
example with xournal. This is used for signing documents,
for example. It should not be confused or compared with
editing the source, that's another thing.

-- 
underground experts united
https://dataswamp.org/~incal

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Ulrich Deiters]

You can try to convert PDF to PostScript, e.g., with pdf2ps, pdftops, …
PostScript is human-readable and can be edited with emacs.

Depending how the PDF code was created, however, the PostScript code
may be difficult to interpret, or even contain blocks of binary code.
Unless you recognize some strings like

	… (some text) show …

which can eventually be modified, it is risky to make changes. As Yuri
stated, PDF (or PostScript) code is meant to be executed. Changes might
result in a buggy code, and then the interpreter or the printer will balk.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[gebser]

On 3/21/23 7:51 AM, Ulrich Deiters wrote:
> You can try to convert PDF to PostScript, e.g., with pdf2ps, pdftops, …
> PostScript is human-readable and can be edited with emacs.
> 
> Depending how the PDF code was created, however, the PostScript code
> may be difficult to interpret, or even contain blocks of binary code.
> Unless you recognize some strings like
> 
>      … (some text) show …
> 
> which can eventually be modified, it is risky to make changes. As Yuri
> stated, PDF (or PostScript) code is meant to be executed. Changes might
> result in a buggy code, and then the interpreter or the printer will balk.
> 
> 

Ugghhh.  You and Yuri have made me dislike PDFs even more than before. 
Yet still those executable(s) could be reverse-engineered.  Better than 
that, what FOSS app(s) would be happy replacement for the 
Acrobat-Industrial Complex?

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Ulrich Deiters]

Am 21.03.23 um 22:57 schrieb gebser@mousecar.com:
> Better than that, what FOSS app(s) would be happy replacement for the 
> Acrobat-Industrial Complex?

This depends on what you want. Printing a document takes
(at least) three stages:

(1) One composes the text using a language that contains formatting
commands. The best open-source choice is LaTeX. LaTeX code is human-
readable and can be edited with emacs; in fact, emacs has a LaTeX
mode. There are also free programs that assist writing LaTeX code,
or which let one compose XML code, which is then translated into LaTeX.

(2) One processes the LaTeX code; the results are DVI, PostScript,
or PDF files. Such files basically consist of instructions of the
type "choose font F, go to location (X, Y), and print character C". 
Therefore these files are long and, if converted to a human-readable
format, very boring. You do not want to edit these.

(3) The intermediate code is processed to produce a bitmap, which is
then transferred to the paper.

So you see, the stage at which editing makes sense is Stage (1).

Exceptions are graphics designers and artists, who work with PostScript
or similar languages directly in order to achieve special effects.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Michael Heerdegen]

Ulrich Deiters <ulrich.deiters@uni-koeln.de> writes:

> So you see, the stage at which editing makes sense is Stage (1).

Apart from that, the few useful things that you can do with existing
pdf's is stuff like reordering pages, changing page sizes or adding page
numbers or merge several files (e.g. pdfsam, pdfchain, page-crunch).

And there is for example Xournal that allows some very simple and
limited editing of pdf files, like adding annotations or inserting your
signature as an image.

Michael.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Jean Louis]

* gebser@mousecar.com <gebser@mousecar.com> [2023-03-22 00:58]:
> Ugghhh.  You and Yuri have made me dislike PDFs even more than before. Yet
> still those executable(s) could be reverse-engineered.  Better than that,
> what FOSS app(s) would be happy replacement for the Acrobat-Industrial
> Complex?

As I do not use proprietary applications, I cannot know what does that
propriety non-free software Acrobat Industrial Complex does. I may
recommend some software here.

Name: flpsed 0.7.3-1
Description     : A WYSIWYG PostScript annotator
URL             : http://flpsed.org/flpsed.html

This software has very simple feature to edit text in PDF documents or
save PS pdocuments to PDF.

In general when I want to change PDF, I sometimes convert it to
picture and change the picture by using ImageMagick tools. And often I
may import PDF into Gimp software and then edit it as pictures. That
is often for reasons to insert signature in contracts.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

FOSS replacement for PDF [Re: editing a PDF]

[gebser]

On 3/21/23 6:55 PM, Ulrich Deiters wrote:
> Am 21.03.23 um 22:57 schrieb gebser@mousecar.com:
>> Better than that, what FOSS app(s) would be happy replacement for the 
>> Acrobat-Industrial Complex?
> 
> This depends on what you want. Printing a document takes
> (at least) three stages:
> 
> (1) One composes the text using a language that contains formatting
> commands. The best open-source choice is LaTeX. LaTeX code is human-
> readable and can be edited with emacs; in fact, emacs has a LaTeX
> mode. There are also free programs that assist writing LaTeX code,
> or which let one compose XML code, which is then translated into LaTeX.
> 
> (2) One processes the LaTeX code; the results are DVI, PostScript,
> or PDF files. Such files basically consist of instructions of the
> type "choose font F, go to location (X, Y), and print character C". 
> Therefore these files are long and, if converted to a human-readable
> format, very boring. You do not want to edit these.
> 
> (3) The intermediate code is processed to produce a bitmap, which is
> then transferred to the paper.
> 
> So you see, the stage at which editing makes sense is Stage (1).
> 
> Exceptions are graphics designers and artists, who work with PostScript
> or similar languages directly in order to achieve special effects.
> 
> 

Thanks for the very thorough reply.  This question really wasn't for me, 
but rather for people I know and meet who write pdf docs, who take pdf 
as a "standard".  I'm afraid that latex would be far too technological 
for them... and for most people.  Even something like Libreoffice's 
Write would have a more familiar UI-- for that reason alone they'd be 
more like likely to try it.

Re: FOSS replacement for PDF [Re: editing a PDF]

[Yuri Khan]

On Wed, 22 Mar 2023 at 15:26, gebser@mousecar.com <gebser@mousecar.com> wrote:

> Even something like Libreoffice's
> Write would have a more familiar UI-- for that reason alone they'd be
> more like likely to try it.

LibreOffice is okay. And it can export to PDF in case that’s needed as
the final format.

Of course, exporting to PDF is a one-way street, therefore, all actual
editing needs to be done on the source LibreOffice document.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[gebser]

On 3/22/23 3:03 AM, Jean Louis wrote:
> Name: flpsed 0.7.3-1
> Description     : A WYSIWYG PostScript annotator
> URL             :http://flpsed.org/flpsed.html

Unfortunately it isn't found by gnome's "Activities" search.  People 
would need to write a menu on their hand and not take baths.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Yuri Khan]

On Wed, 22 Mar 2023 at 22:46, Emanuel Berg <incal@dataswamp.org> wrote:

> > You don’t normally edit executable programs, you compile
> > them from source. In the same vein, to get a modified PDF,
> > you find the source document from which it was produced,
> > modify that, and re-export.
>
> Still, they are editable at/from the PDF level as well, for
> example with xournal.

That’s not really editing. You can scrawl over the existing text, you
can carefully overpaint existing text with your own, but you cannot
replace one word with a different word and hope the rest of the text
is repositioned to accommodate the change.

(To stretch the executable program analogy, you can append new
instructions to an existing program. They will be executed and may, in
limited cases, have your desired effect.)

Xournal doesn’t even claim it supports editing — the menu item is
called Annotate PDF.

Commercial, proprietary Acrobat (not the Reader variety, the full
suite) can, in limited ways, modify individual runs of text. But if
you overflow a line, it won’t do anything to help you, you just get an
overfull line.

> This is used for signing documents,
> for example.

I’d be very wary of putting my handwritten signature as an annotation
in a PDF document and giving that to another person. For all I know,
it would enable them to extract the annotation containing my signature
and apply it to a different document.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Bob Newell]

I get sent PDFs all the time and it's often frustrating.
However most of the time I need the text, not necessarily the
formatting.  For that, there are various PDF to text
converters which work to some degree (not so great with
multi-column text for the most part).

I also found a PDF to spreadsheet converter which isn't
especially great but better than nothing.  Well, at least a
little better.

I realize this doesn't at all address the original question
about editing PDFs in the true sense of editing.  But in a lot
of cases text extraction may be good enough, depending on your
purpose.

Copy/paste works in many cases as well.  For instance I just
got sent a PDF of a budget spreadsheet, and with not so much
effort was able to make it into an org-mode table.

All bets are off if the PDF is an /image/ of some text.  Then
you've got to get out your friendly OCR software.  In such
cases I write back to the sender and politely ask for
something easier to use.

-- 
Bob Newell
Honolulu, Hawai`i

- Via GNU/Linux/Emacs/Gnus/BBDB

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Jean Louis]

* Yuri Khan <yuri.v.khan@gmail.com> [2023-03-22 19:34]:
> > This is used for signing documents,
> > for example.
> 
> I’d be very wary of putting my handwritten signature as an
> annotation in a PDF document and giving that to another person. For
> all I know, it would enable them to extract the annotation
> containing my signature and apply it to a different document.

It is false sense of safety.

Any of your signature anywhere ever given to anybody, from paper or
anywhere, could be taken, then made picture of it and easily placed in
any kind of documents.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

Re: FOSS replacement for PDF [Re: editing a PDF]

[Jean Louis]

* Yuri Khan <yuri.v.khan@gmail.com> [2023-03-22 12:20]:
> Of course, exporting to PDF is a one-way street, therefore, all actual
> editing needs to be done on the source LibreOffice document.

Okay, though, look at these nice example how it can be converted to
HTML:

pdf2htmlEX by coolwanglu:
https://pdf2htmlex.github.io/pdf2htmlEX/

pdf2htmlex.github.io/pdf2htmlEX/demo/geneve.html:
https://pdf2htmlex.github.io/pdf2htmlEX/demo/geneve.html

pdf2htmlex.github.io/pdf2htmlEX/demo/cheat.html:
https://pdf2htmlex.github.io/pdf2htmlEX/demo/cheat.html

pdf2htmlex.github.io/pdf2htmlEX/demo/issue65_en.html:
https://pdf2htmlex.github.io/pdf2htmlEX/demo/issue65_en.html

pdf2htmlex.github.io/pdf2htmlEX/demo/demo.html:
https://pdf2htmlex.github.io/pdf2htmlEX/demo/demo.html

So far it is possible to convert PDF to plain text provided there was
some text inside, and to HTML.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Jean Louis]

* Bob Newell <bobnewell@bobnewell.net> [2023-03-22 21:49]:
> I get sent PDFs all the time and it's often frustrating.
> However most of the time I need the text, not necessarily the
> formatting.  For that, there are various PDF to text
> converters which work to some degree (not so great with
> multi-column text for the most part).

Following references may be helpful to some people thinking to use
multi-column text in Emacs:

There is (info "(emacs) Two-Column") in Emacs manual.

Example output produced by using RCD Paps packag for Emacs:
https://gnu.support/files/emacs/packages/rcd-paps/2023/03/2023-03-23/2023-03-23-12:27:05.pdf

GNU Emacs package: rcd-paps.el:
https://gnu.support/gnu-emacs/packages/GNU-Emacs-package-rcd-paps-el-76862.html

Any Emacs printing can produce such pages. 

> I realize this doesn't at all address the original question
> about editing PDFs in the true sense of editing.  But in a lot
> of cases text extraction may be good enough, depending on your
> purpose.

LibreaOffice Draw -- open PDF file and you will see you can edit many
things. 

Okular -- you can annotate and insert text

Scribus and Xournal

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

Re: FOSS replacement for PDF [Re: editing a PDF]

[Yuri Khan]

On Thu, 23 Mar 2023 at 16:24, Jean Louis <bugs@gnu.support> wrote:

> Okay, though, look at these nice example how it can be converted to
> HTML:
>
> pdf2htmlEX by coolwanglu:
> https://pdf2htmlex.github.io/pdf2htmlEX/

Yeah that’s somewhat impressive.

You will notice there are stray spaces in some words, possibly due to kerning.

> pdf2htmlex.github.io/pdf2htmlEX/demo/geneve.html:
> https://pdf2htmlex.github.io/pdf2htmlEX/demo/geneve.html

This one is literally unreadable. (If you allow web fonts, it renders
in a hard-to-read font, if you don’t, you get two pages of tofu;
copying to clipboard does not work in any case.)

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Michael Heerdegen]

Yuri Khan <yuri.v.khan@gmail.com> writes:

> I’d be very wary of putting my handwritten signature as an annotation
> in a PDF document and giving that to another person. For all I know,
> it would enable them to extract the annotation containing my signature
> and apply it to a different document.

Anybody can do that with a scan of my signature anyway (that's how I got
my version).  I do have to sign stuff.  I can't prevent criminals from
doing criminal things with my signature in this case I think.

I don't even expect that such pdfs are valid documents btw (as a lot of
such things are actually, like photos of identity cards).

Michael.

Re: FOSS replacement for PDF [Re: editing a PDF]

[Gregory Heytings]

>> Okay, though, look at these nice example how it can be converted to 
>> HTML:
>>
>> pdf2htmlEX by coolwanglu: https://pdf2htmlex.github.io/pdf2htmlEX/
>
> Yeah that’s somewhat impressive.
>

It's impressive, but not for what the OP wanted to do, namely edit the 
text.  The resulting HTML page has the same structure as the PDF: it's a 
sequence of "put this character in that font at that position" 
instructions.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Emanuel Berg]

Yuri Khan wrote:

>>> You don’t normally edit executable programs, you compile
>>> them from source. In the same vein, to get a modified PDF,
>>> you find the source document from which it was produced,
>>> modify that, and re-export.
>>
>> Still, they are editable at/from the PDF level as well, for
>> example with xournal.
>
> That’s not really editing. You can scrawl over the existing
> text, you can carefully overpaint existing text with your
> own, but you cannot replace one word with a different word
> and hope the rest of the text is repositioned to accommodate
> the change.

Well, obviously it's not the same as editing the LaTeX source
and compiling the original document. Actually not only is it
"not the same", it's completely different.

-- 
underground experts united
https://dataswamp.org/~incal

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Emanuel Berg]

Michael Heerdegen wrote:

>> I’d be very wary of putting my handwritten signature as an
>> annotation in a PDF document and giving that to another
>> person. For all I know, it would enable them to extract the
>> annotation containing my signature and apply it to
>> a different document.
>
> Anybody can do that with a scan of my signature anyway
> (that's how I got my version). I do have to sign stuff.
> I can't prevent criminals from doing criminal things with my
> signature in this case I think.

Indeed, as the major of Limbo City, I do that all the time!
But it's my impression these documents don't contain that
much, most of the time, so maybe there is no harm to it.

-- 
underground experts united
https://dataswamp.org/~incal

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Bob Newell]

> Following references may be helpful to some people thinking to use
> multi-column text in Emacs:
>
> There is (info "(emacs) Two-Column") in Emacs manual.

Right, although I was referring to a different problem, namely that if
you do a PDF to text conversion of a PDF with multi-column text, you
typically don't get great results.  Or at least I don't.  I do better
with OCR software that allows me to select a column at a time.  But
then we're getting into multiple manual steps, introducing OCR error,
and in the end we are working with the text, not directly editing the
PDF.  (However extracting text is often enough for me.)

> Example output produced by using RCD Paps packag for Emacs:
> https://gnu.support/files/emacs/packages/rcd-paps/2023/03/2023-03-23/2023-03-23-12:27:05.pdf

I didn't know about the RCD package.  Looks useful and thank you for
the pointer.

> LibreaOffice Draw -- open PDF file and you will see you can edit many
> things.
>
> Okular -- you can annotate and insert text
>
> Scribus and Xournal

Yes, I use Xournal very often and sometimes Gimp in this manner for
signing or annotating business documents received as PDFs.  It is
still not the same as directly editing.  However if the reproduction
is faithfully enough it is indeed possible to make significant changes
and export back out as PDF.

-- 
Bob Newell
Honolulu, Hawai`i

Via Linux/Emacs/Gnus/BBDB.

Re: editing a PDF [Re: emacs 30.5.0 editing epub]

[Michael Heerdegen]

Ulrich Deiters <ulrich.deiters@uni-koeln.de> writes:

> (1) One composes the text using a language that contains formatting
> commands. The best open-source choice is LaTeX. LaTeX code is human-
> readable and can be edited with emacs; in fact, emacs has a LaTeX
> mode. There are also free programs that assist writing LaTeX code,
> or which let one compose XML code, which is then translated into LaTeX.

Let me add: also org-mode can export to pdf.

If it should be LaTex and you don't want to write code yourself, even
not in Emacs with the really good AUCtex package, there are things like
lyx:

  https://www.lyx.org/

- hybrid latex "editors" that mostly feel like using text processing in
LibreOffice (mostly WYSIWYG).  It's a good compromise.

Org is even simpler.  It's not LaTex but still allows to use LaTeX
snippets (e.g. for mathematical stuff) that will be integrated in the
exported document (be it ps, pdf or html).

Michael.