rpm/epub2txt2
SHA256
1
0
Fork 0
forked from pool/epub2txt2
Code Pull Requests Activity

Accepting request 992660 from utilities

Browse Source 
Automatic submission by obs-autosubmit

OBS-URL: https://build.opensuse.org/request/show/992660
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/epub2txt2?expand=0&rev=5
This commit is contained in:
Dominique Leuenberger 2022-08-04 11:23:48 +00:00 committed by Git OBS Bridge
commit 8a3f44e74b
5 changed files with 173 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
02f69e6243d6c96f78da45fb710a265e5aee2fb5.tar.gz
View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:b700901f8af180fb0e7d06a5d1b5491dad8a2ae88e141212464c9b7e62a95ecf
size 79888

156
54b41e87156bc823d5938749d71c4c57adc75b1b.patch Normal file
View File

@ -0,0 +1,156 @@
From 54b41e87156bc823d5938749d71c4c57adc75b1b Mon Sep 17 00:00:00 2001
From: Kevin Boone <kevin@railwayterrace.com>
Date: Thu, 30 Jun 2022 15:41:59 +0100
Subject: [PATCH] Fixed handling of URL-encoded spine hrefs
---
README.md | 1 +
src/epub2txt.c | 3 ++-
src/util.c | 44 +++++++++++++++++++++++++++++++++++++++++++-
src/util.h | 6 +++++-
src/xhtml.c | 5 +++++
5 files changed, 56 insertions(+), 3 deletions(-)
diff --git a/README.md b/README.md
index 4fbcafc..546242f 100644
--- a/README.md
+++ b/README.md
@@ -244,6 +244,7 @@ covered.
Date | Change
-----|-------
+?,&nbsp;Jun&nbsp;2022 | Fixed handling of URL-encoded spine href's
2.06,&nbsp;Jun&nbsp;2022 | Fixed bug in invoking unzip
2.05,&nbsp;Apr&nbsp;2022 | Fixed bug with empty metadata tags
2.04,&nbsp;Apr&nbsp;2022 | Improved handling of UTF-8 BOMs
diff --git a/src/epub2txt.c b/src/epub2txt.c
index 7e9f4a1..72e0504 100644
--- a/src/epub2txt.c
+++ b/src/epub2txt.c
@@ -312,7 +312,8 @@ List *epub2txt_get_items (const char *opf, char **error)
char *val2 = r3->attributes[p].value;
if (strcmp (name2, "href") == 0)
{
- list_append (ret, strdup (val2));
+ char *decoded_val2 = decode_url (val2);
+ list_append (ret, decoded_val2);
}
}
}
diff --git a/src/util.c b/src/util.c
index 853343c..16e7431 100644
--- a/src/util.c
+++ b/src/util.c
@@ -1,12 +1,14 @@
/*============================================================================
epub2txt v2
util.c
- Copyright (c)2022 Marco Bonelli, GPL v3.0
+ Copyright (c)2022 Marco Bonelli, Kevin Boone, GPL v3.0
============================================================================*/
#include <errno.h>
#include <string.h>
#include <unistd.h>
+#include <stdlib.h>
+#include <ctype.h>
#include <signal.h>
#include <sys/wait.h>
#include "util.h"
@@ -16,6 +18,7 @@
run_command
Run an helper command through fork + execvp, wait for it to finish and return
its status. Log execvp errors, and abort execution if abort_on_error is TRUE.
+(Marco Bonelli)
*==========================================================================*/
int run_command (const char *const argv[], BOOL abort_on_error)
{
@@ -39,3 +42,42 @@ int run_command (const char *const argv[], BOOL abort_on_error)
waitpid (pid, &status, 0);
return status;
}
+
+/*==========================================================================
+ Decode %xx in URL-type strings. The caller must free the resulting
+ string, which will be no longer than the input.
+ (Kevin Boone)
+*==========================================================================*/
+char *decode_url (const char *url)
+ {
+ char *ret = malloc (strlen (url) + 2);
+
+ int len = 0;
+ for (; *url; len++)
+ {
+ if (*url == '%' && url[1] && url[2] &&
+ isxdigit(url[1]) && isxdigit(url[2]))
+ {
+ char url1 = url[1];
+ char url2 = url[2];
+ url1 -= url1 <= '9' ? '0' : (url1 <= 'F' ? 'A' : 'a')-10;
+ url2 -= url2 <= '9' ? '0' : (url2 <= 'F' ? 'A' : 'a')-10;
+ ret[len] = 16 * url1 + url2;
+ url += 3;
+ continue;
+ }
+ else if (*url == '+')
+ {
+ /* I have not tested this piece of the function, because I have not
+ seen any instances of '+' (meaning space) in a spine href */
+ url += 1;
+ ret[len] = ' ';
+ }
+ ret[len] = *url++;
+ }
+ ret[len] = '\0';
+
+ return ret;
+ }
+
+
diff --git a/src/util.h b/src/util.h
index 2685a02..6b0c197 100644
--- a/src/util.h
+++ b/src/util.h
@@ -1,7 +1,7 @@
/*============================================================================
epub2txt v2
util.h
- Copyright (c)2022 Marco Bonelli, GPL v3.0
+ Copyright (c)2022 Marco Bonelli, Kevin Boone GPL v3.0
============================================================================*/
#pragma once
@@ -9,3 +9,7 @@
#include "defs.h"
int run_command (const char *const argv[], BOOL abort_on_error);
+
+/** Decode %xx in URL-type strings. The caller must free the resulting
+ string, which will be no longer than the input. */
+char *decode_url (const char *url);
diff --git a/src/xhtml.c b/src/xhtml.c
index 1338882..fbfceae 100644
--- a/src/xhtml.c
+++ b/src/xhtml.c
@@ -530,6 +530,8 @@ WString *xhtml_transform_char (uint32_t c, BOOL to_ascii)
============================================================================*/
WString *xhtml_translate_entity (const WString *entity)
{
+ /* Program flow in this function is very ugly, and prone to memory
+ leaks when modified. The whole thing needs to be rewritten */
char out[20];
IN
char *in = wstring_to_utf8 (entity);
@@ -569,8 +571,11 @@ WString *xhtml_translate_entity (const WString *entity)
WString *ret = wstring_create_empty();
wstring_append_c (ret, (uint32_t)v);
OUT
+ free (s);
+ free (in);
return ret;
}
+ free (s);
}
else
{

3
ac4e73fa79202ccc106b36c06c20e36c37345a58.tar.gz Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:b067c613c4cb4e3069a06cd90fa8b80c7ec9e187c296c66337daf5b14a87c259
size 80852

9
epub2txt2.changes
View File

@ -1,3 +1,12 @@
-------------------------------------------------------------------
Thu Jul 28 03:10:37 UTC 2022 - Soc Virnyl Estela <socvirnyl.estela@gmail.com>
- Update to version 2.06:
* No changelog available
- Added 54b41e87156bc823d5938749d71c4c57adc75b1b.patch:
* Fixed handling of URL-encoded spine hrefs
-------------------------------------------------------------------
Sun May 1 08:20:56 UTC 2022 - Soc Virnyl Estela <socvirnyl.estela@gmail.com>

9
epub2txt2.spec
View File

@ -16,16 +16,17 @@
#
%global _commit_hash "02f69e6243d6c96f78da45fb710a265e5aee2fb5"
%global _commit_hash ac4e73fa79202ccc106b36c06c20e36c37345a58
Name: epub2txt2
Version: 2.04
Version: 2.06
Release: 0
Summary: Simple command-line utility for extracting text from EPUB documents
License: GPL-3.0-only
Group: Productivity/Text/Utilities
URL: https://github.com/kevinboone/epub2txt2
Source0: https://github.com/kevinboone/epub2txt2/archive/02f69e6243d6c96f78da45fb710a265e5aee2fb5.tar.gz
Source0: https://github.com/kevinboone/epub2txt2/archive/%{_commit_hash}.tar.gz
Patch0: 54b41e87156bc823d5938749d71c4c57adc75b1b.patch
BuildRequires: gcc
BuildRequires: make
Requires: unzip
@ -34,7 +35,7 @@ Requires: unzip
Simple command-line utility for extracting text from EPUB documents
%prep
%setup -n %{name}-%{_commit_hash}
%autosetup -n %{name}-%{_commit_hash} -p1
%build
%make_build