X-Git-Url: http://sjero.net/git/?p=wget;a=blobdiff_plain;f=src%2Frecur.c;h=66ef2e0fd9b566d0c9c809562daa2d4ee32ddb70;hp=24b80ad475878203680ec355f9ea818758e00b0c;hb=b014f8fae9291e7504c0cca2dd8b9a0035466c03;hpb=b30a0dd817886f77a64be9218c5e5399bcbc2e67 diff --git a/src/recur.c b/src/recur.c index 24b80ad4..66ef2e0f 100644 --- a/src/recur.c +++ b/src/recur.c @@ -51,7 +51,6 @@ as that of the covered work. */ #include "html-url.h" #include "css-url.h" #include "spider.h" -#include "iri.h" /* Functions for maintaining the URL queue. */ @@ -61,7 +60,7 @@ struct queue_element { int depth; /* the depth */ bool html_allowed; /* whether the document is allowed to be treated as HTML. */ - char *remote_encoding; + struct iri *iri; /* sXXXav */ bool css_allowed; /* whether the document is allowed to be treated as CSS. */ struct queue_element *next; /* next element in queue */ @@ -95,12 +94,12 @@ url_queue_delete (struct url_queue *queue) into it. */ static void -url_enqueue (struct url_queue *queue, +url_enqueue (struct url_queue *queue, struct iri *i, const char *url, const char *referer, int depth, bool html_allowed, bool css_allowed) { struct queue_element *qel = xnew (struct queue_element); - char *charset = get_current_charset (); + qel->iri = i; qel->url = url; qel->referer = referer; qel->depth = depth; @@ -108,19 +107,17 @@ url_enqueue (struct url_queue *queue, qel->css_allowed = css_allowed; qel->next = NULL; - if (charset) - qel->remote_encoding = xstrdup (charset); - else - qel->remote_encoding = NULL; - ++queue->count; if (queue->count > queue->maxcount) queue->maxcount = queue->count; - DEBUGP (("Enqueuing %s at depth %d\n", url, depth)); + DEBUGP (("Enqueuing %s at depth %d\n", + quotearg_n_style (0, escape_quoting_style, url), depth)); DEBUGP (("Queue count %d, maxcount %d.\n", queue->count, queue->maxcount)); - /*printf ("[Enqueuing %s with %s\n", url, qel->remote_encoding);*/ + if (i) + DEBUGP (("[IRI Enqueuing %s with %s\n", quote_n (0, url), + i->uri_encoding ? quote_n (1, i->uri_encoding) : "None")); if (queue->tail) queue->tail->next = qel; @@ -134,7 +131,7 @@ url_enqueue (struct url_queue *queue, succeeded, or false if the queue is empty. */ static bool -url_dequeue (struct url_queue *queue, +url_dequeue (struct url_queue *queue, struct iri **i, const char **url, const char **referer, int *depth, bool *html_allowed, bool *css_allowed) { @@ -147,10 +144,7 @@ url_dequeue (struct url_queue *queue, if (!queue->head) queue->tail = NULL; - set_remote_charset (qel->remote_encoding); - if (qel->remote_encoding) - xfree (qel->remote_encoding); - + *i = qel->iri; *url = qel->url; *referer = qel->referer; *depth = qel->depth; @@ -159,7 +153,8 @@ url_dequeue (struct url_queue *queue, --queue->count; - DEBUGP (("Dequeuing %s at depth %d\n", qel->url, qel->depth)); + DEBUGP (("Dequeuing %s at depth %d\n", + quotearg_n_style (0, escape_quoting_style, qel->url), qel->depth)); DEBUGP (("Queue count %d, maxcount %d.\n", queue->count, queue->maxcount)); xfree (qel); @@ -167,9 +162,9 @@ url_dequeue (struct url_queue *queue, } static bool download_child_p (const struct urlpos *, struct url *, int, - struct url *, struct hash_table *); -static bool descend_redirect_p (const char *, const char *, int, - struct url *, struct hash_table *); + struct url *, struct hash_table *, struct iri *); +static bool descend_redirect_p (const char *, struct url *, int, + struct url *, struct hash_table *, struct iri *); /* Retrieve a part of the web beginning with START_URL. This used to @@ -194,7 +189,7 @@ static bool descend_redirect_p (const char *, const char *, int, options, add it to the queue. */ uerr_t -retrieve_tree (const char *start_url) +retrieve_tree (struct url *start_url_parsed, struct iri *pi) { uerr_t status = RETROK; @@ -206,24 +201,27 @@ retrieve_tree (const char *start_url) struct hash_table *blacklist; int up_error_code; - struct url *start_url_parsed; + struct iri *i = iri_new (); - set_ugly_no_encode (true); - start_url_parsed= url_parse (start_url, &up_error_code); - set_ugly_no_encode (false); - if (!start_url_parsed) +#define COPYSTR(x) (x) ? xstrdup(x) : NULL; + /* Duplicate pi struct if not NULL */ + if (pi) { - logprintf (LOG_NOTQUIET, "%s: %s.\n", start_url, - url_error (up_error_code)); - return URLERROR; + i->uri_encoding = COPYSTR (pi->uri_encoding); + i->content_encoding = COPYSTR (pi->content_encoding); + i->utf8_encode = pi->utf8_encode; } + else + set_uri_encoding (i, opt.locale, true); +#undef COPYSTR queue = url_queue_new (); blacklist = make_string_hash_table (0); /* Enqueue the starting URL. Use start_url_parsed->url rather than just URL so we enqueue the canonical form of the URL. */ - url_enqueue (queue, xstrdup (start_url_parsed->url), NULL, 0, true, false); + url_enqueue (queue, i, xstrdup (start_url_parsed->url), NULL, 0, true, + false); string_set_add (blacklist, start_url_parsed->url); while (1) @@ -242,7 +240,7 @@ retrieve_tree (const char *start_url) /* Get the next URL from the queue... */ - if (!url_dequeue (queue, + if (!url_dequeue (queue, (struct iri **) &i, (const char **)&url, (const char **)&referer, &depth, &html_allowed, &css_allowed)) break; @@ -280,10 +278,12 @@ retrieve_tree (const char *start_url) } else { - int dt = 0; + int dt = 0, url_err; char *redirected = NULL; + struct url *url_parsed = url_parse (url, &url_err, i, true); - status = retrieve_url (url, &file, &redirected, referer, &dt, false); + status = retrieve_url (url_parsed, url, &file, &redirected, referer, + &dt, false, i, true); if (html_allowed && file && status == RETROK && (dt & RETROKF) && (dt & TEXTHTML)) @@ -310,8 +310,8 @@ retrieve_tree (const char *start_url) want to follow it. */ if (descend) { - if (!descend_redirect_p (redirected, url, depth, - start_url_parsed, blacklist)) + if (!descend_redirect_p (redirected, url_parsed, depth, + start_url_parsed, blacklist, i)) descend = false; else /* Make sure that the old pre-redirect form gets @@ -322,6 +322,12 @@ retrieve_tree (const char *start_url) xfree (url); url = redirected; } + else + { + xfree (url); + url = xstrdup (url_parsed->url); + } + url_free(url_parsed); } if (opt.spider) @@ -363,7 +369,7 @@ retrieve_tree (const char *start_url) bool meta_disallow_follow = false; struct urlpos *children = is_css ? get_urls_css_file (file, url) : - get_urls_html (file, url, &meta_disallow_follow); + get_urls_html (file, url, &meta_disallow_follow, i); if (opt.use_robots && meta_disallow_follow) { @@ -374,9 +380,8 @@ retrieve_tree (const char *start_url) if (children) { struct urlpos *child = children; - set_ugly_no_encode (true); - struct url *url_parsed = url_parse (url, NULL); - set_ugly_no_encode (false); + struct url *url_parsed = url_parse (url, NULL, i, true); + struct iri *ci; char *referer_url = url; bool strip_auth = (url_parsed != NULL && url_parsed->user != NULL); @@ -393,9 +398,11 @@ retrieve_tree (const char *start_url) if (dash_p_leaf_HTML && !child->link_inline_p) continue; if (download_child_p (child, url_parsed, depth, start_url_parsed, - blacklist)) + blacklist, i)) { - url_enqueue (queue, xstrdup (child->url->url), + ci = iri_new (); + set_uri_encoding (ci, i->content_encoding, false); + url_enqueue (queue, ci, xstrdup (child->url->url), xstrdup (referer_url), depth + 1, child->link_expect_html, child->link_expect_css); @@ -440,6 +447,7 @@ retrieve_tree (const char *start_url) xfree (url); xfree_null (referer); xfree_null (file); + iri_free (i); } /* If anything is left of the queue due to a premature exit, free it @@ -448,17 +456,17 @@ retrieve_tree (const char *start_url) char *d1, *d2; int d3; bool d4, d5; - while (url_dequeue (queue, + struct iri *d6; + while (url_dequeue (queue, (struct iri **)&d6, (const char **)&d1, (const char **)&d2, &d3, &d4, &d5)) { + iri_free (d6); xfree (d1); xfree_null (d2); } } url_queue_delete (queue); - if (start_url_parsed) - url_free (start_url_parsed); string_set_free (blacklist); if (opt.quota && total_downloaded_bytes > opt.quota) @@ -479,7 +487,8 @@ retrieve_tree (const char *start_url) static bool download_child_p (const struct urlpos *upos, struct url *parent, int depth, - struct url *start_url_parsed, struct hash_table *blacklist) + struct url *start_url_parsed, struct hash_table *blacklist, + struct iri *iri) { struct url *u = upos->url; const char *url = u->url; @@ -620,7 +629,7 @@ download_child_p (const struct urlpos *upos, struct url *parent, int depth, if (!specs) { char *rfile; - if (res_retrieve_file (url, &rfile)) + if (res_retrieve_file (url, &rfile, iri)) { specs = res_parse_from_file (rfile); @@ -674,28 +683,25 @@ download_child_p (const struct urlpos *upos, struct url *parent, int depth, it is merely a simple-minded wrapper around download_child_p. */ static bool -descend_redirect_p (const char *redirected, const char *original, int depth, - struct url *start_url_parsed, struct hash_table *blacklist) +descend_redirect_p (const char *redirected, struct url *orig_parsed, int depth, + struct url *start_url_parsed, struct hash_table *blacklist, + struct iri *iri) { - struct url *orig_parsed, *new_parsed; + struct url *new_parsed; struct urlpos *upos; bool success; - set_ugly_no_encode (true); - orig_parsed = url_parse (original, NULL); assert (orig_parsed != NULL); - new_parsed = url_parse (redirected, NULL); + new_parsed = url_parse (redirected, NULL, NULL, false); assert (new_parsed != NULL); - set_ugly_no_encode (false); upos = xnew0 (struct urlpos); upos->url = new_parsed; success = download_child_p (upos, orig_parsed, depth, - start_url_parsed, blacklist); + start_url_parsed, blacklist, iri); - url_free (orig_parsed); url_free (new_parsed); xfree (upos);