GopherProxy

	improve link references, add option to show full URL inline - webdump - HTML to…
	git clone git://git.codemadness.org/webdump
	Log
	Files
	Refs
	README
	LICENSE
	---
	commit 6365a78f6c050106e64b281d29d8ef550f131bf1
	parent 56ec7ea6c49d79cc3aaf301d2e6040e15d17785a
	Author: Hiltjo Posthuma <[email protected]>
	Date: Fri, 8 Sep 2023 11:25:13 +0200

	improve link references, add option to show full URL inline

	- fix URL references not being visible when only the -l option is specified
	(without -i). Now each option can be specified separately.
	- add -I option to show full URL option inline.

	Diffstat:
	M webdump.1 \| 5 ++++-
	M webdump.c \| 90 +++++++++++++++++------------…

	2 files changed, 53 insertions(+), 42 deletions(-)
	---
	diff --git a/webdump.1 b/webdump.1
	@@ -6,7 +6,7 @@
	.Nd convert HTML to plain-text
	.Sh SYNOPSIS
	.Nm
	-.Op Fl 8ailrx
	+.Op Fl 8aiIlrx
	.Op Fl b Ar baseurl
	.Op Fl s Ar selector
	.Op Fl u Ar selector
	@@ -30,6 +30,9 @@ This is used to make links absolute.
	.It Fl i
	Toggle if link reference numbers are displayed inline or not, by default it is
	not enabled.
	+.It Fl I
	+Toggle if URLs for link reference are displayed inline or not, by default it is
	+not enabled.
	.It Fl l
	Toggle if link references are displayed at the bottom or not, by default it is
	not enabled.
	diff --git a/webdump.c b/webdump.c
	@@ -47,6 +47,7 @@ struct uri {
	static int allowansi = 0; /* allow ANSI escape codes */
	static int showrefbottom = 0; /* show link references at the bottom */
	static int showrefinline = 0; /* show link reference number inline */
	+static int showurlinline = 0; /* show full link reference inline */
	static int linewrap = 0; /* line-wrapping */
	static int termwidth = 77; /* terminal width */
	static int resources = 0; /* write resources line-by-line to fd 3? */
	@@ -1319,46 +1320,49 @@ handleinlinelink(void)
	char buf[4096], *url;
	int b, r;

	- /* show links as reference at the bottom */
	- if ((showrefbottom \|\| resources) && (attr_src.len \|\| attr_href.len)) {
	- /* by default use the original URL */
	- if (attr_src.len)
	- url = attr_src.data;
	- else
	- url = attr_href.data;
	-
	- b = -1;
	- if (uri_hasscheme(url))
	- ; /* already absolute: nothing to do */
	- else if (basehref[0]) /* prefer -b option over <base> */
	- b = uri_parse(basehref, &base);
	- else if (basehrefdoc[0])
	- b = uri_parse(basehrefdoc, &base);
	-
	- if (b != -1 &&
	- uri_parse(url, &olduri) != -1 &&
	- uri_makeabs(&newuri, &olduri, &base) != -1 &&
	- newuri.proto[0]) {
	- r = uri_format(buf, sizeof(buf), &newuri);
	- if (r >= 0 && (size_t)r < sizeof(buf))
	- url = buf;
	- }
	+ if (!showrefbottom && !showrefinline && !showurlinline && !resources)
	+ return; /* there is no need to collect the reference */

	- if (!url[0])
	- return;
	+ if (!attr_src.len && !attr_href.len)
	+ return; /* there is no reference */

	- cur = &nodes[curnode];
	+ /* by default use the original URL */
	+ if (attr_src.len)
	+ url = attr_src.data;
	+ else
	+ url = attr_href.data;
	+
	+ b = -1;
	+ if (uri_hasscheme(url))
	+ ; /* already absolute: nothing to do */
	+ else if (basehref[0]) /* prefer -b option over <base> */
	+ b = uri_parse(basehref, &base);
	+ else if (basehrefdoc[0])
	+ b = uri_parse(basehrefdoc, &base);
	+
	+ if (b != -1 &&
	+ uri_parse(url, &olduri) != -1 &&
	+ uri_makeabs(&newuri, &olduri, &base) != -1 &&
	+ newuri.proto[0]) {
	+ r = uri_format(buf, sizeof(buf), &newuri);
	+ if (r >= 0 && (size_t)r < sizeof(buf))
	+ url = buf;
	+ }

	- if (showrefinline && !(cur->tag.displaytype & DisplayNone)) {
	- string_clear(&nodes_links[curnode]);
	- string_append(&nodes_links[curnode], url, strlen(url));
	- }
	+ if (!url[0])
	+ return;
	+
	+ cur = &nodes[curnode];

	- /* add hidden links directly to the reference,
	- the order doesn't matter */
	- if (cur->tag.displaytype & DisplayNone)
	- addlinkref(url, cur->tag.name, 1);
	+ if (!(cur->tag.displaytype & DisplayNone)) {
	+ string_clear(&nodes_links[curnode]);
	+ string_append(&nodes_links[curnode], url, strlen(url));
	}
	+
	+ /* add hidden links directly to the reference,
	+ the order doesn't matter */
	+ if (cur->tag.displaytype & DisplayNone)
	+ addlinkref(url, cur->tag.name, 1);
	}

	void
	@@ -1574,11 +1578,12 @@ endnode(struct node *cur)
	/* add link and show the link number in the visible order */
	if (!ishidden && nodes_links[curnode].len > 0) {
	addlinkref(nodes_links[curnode].data, cur->tag.name, ishidden);
	-#if 1
	- hprintf("[%zu]", ++linkcount);
	-#else
	- hprintf("[%s: %s]", cur->tag.name, nodes_links[curnode].data);
	-#endif
	+ if (showrefinline)
	+ hprintf("[%zu]", ++linkcount);
	+ if (showurlinline)
	+ hprintf(" [%s: %s]",
	+ !tagcmp(cur->tag.name, "a") ? "link" : cur->ta…
	+ nodes_links[curnode].data);
	}

	handleendtag(&(cur->tag));
	@@ -2014,7 +2019,7 @@ xmlattrstart(XMLParser p, const char t, size_t tl, cons…
	void
	usage(void)
	{
	- fprintf(stderr, "%s [-8ailrx] [-b basehref] [-s selector] [-u selector…
	+ fprintf(stderr, "%s [-8aiIlrx] [-b basehref] [-s selector] [-u selecto…
	exit(1);
	}

	@@ -2038,6 +2043,9 @@ main(int argc, char **argv)
	case 'i':
	showrefinline = !showrefinline;
	break;
	+ case 'I':
	+ showurlinline = !showurlinline;
	+ break;
	case 'l':
	showrefbottom = !showrefbottom;
	break;