Vsevolod Stakhov
9 years ago
2 changed files with 30 additions and 10 deletions
+ 26
- 3
src/libserver/html.c
View File
| static gboolean | static gboolean | ||||
| rspamd_html_process_tag (rspamd_mempool_t *pool, struct html_content *hc, | rspamd_html_process_tag (rspamd_mempool_t *pool, struct html_content *hc, | ||||
| struct html_tag *tag, GNode **cur_level) | |||||
| struct html_tag *tag, GNode **cur_level, gboolean *balanced) | |||||
| { | { | ||||
| GNode *nnode; | GNode *nnode; | ||||
| nnode = g_node_new (tag); | nnode = g_node_new (tag); | ||||
| if (tag->params) { | |||||
| rspamd_mempool_add_destructor (pool, | |||||
| (rspamd_mempool_destruct_t) g_list_free, | |||||
| tag->params); | |||||
| } | |||||
| if (tag->flags & FL_CLOSING) { | if (tag->flags & FL_CLOSING) { | ||||
| if (!*cur_level) { | if (!*cur_level) { | ||||
| debug_task ("bad parent node"); | debug_task ("bad parent node"); | ||||
| g_node_destroy (nnode); | |||||
| return FALSE; | return FALSE; | ||||
| } | } | ||||
| g_node_append (*cur_level, nnode); | g_node_append (*cur_level, nnode); | ||||
| if (!rspamd_html_check_balance (nnode, cur_level)) { | if (!rspamd_html_check_balance (nnode, cur_level)) { | ||||
| debug_task ( | debug_task ( | ||||
| "mark part as unbalanced as it has not pairable closing tags"); | "mark part as unbalanced as it has not pairable closing tags"); | ||||
| hc->flags |= RSPAMD_HTML_FLAG_UNBALANCED; | hc->flags |= RSPAMD_HTML_FLAG_UNBALANCED; | ||||
| *balanced = FALSE; | |||||
| } | |||||
| else { | |||||
| *balanced = TRUE; | |||||
| } | } | ||||
| } | } | ||||
| else { | else { | ||||
| { | { | ||||
| const guchar *p, *c, *end, *tag_start = NULL, *savep = NULL; | const guchar *p, *c, *end, *tag_start = NULL, *savep = NULL; | ||||
| guchar t; | guchar t; | ||||
| gboolean closing = FALSE, need_decode = FALSE, save_space = FALSE; | |||||
| gboolean closing = FALSE, need_decode = FALSE, save_space = FALSE, balanced; | |||||
| GByteArray *dest; | GByteArray *dest; | ||||
| guint obrace = 0, ebrace = 0; | guint obrace = 0, ebrace = 0; | ||||
| GNode *cur_level = NULL; | GNode *cur_level = NULL; | ||||
| savep = NULL; | savep = NULL; | ||||
| if (cur_tag != NULL) { | if (cur_tag != NULL) { | ||||
| if (rspamd_html_process_tag (pool, hc, cur_tag, &cur_level)) { | |||||
| balanced = TRUE; | |||||
| if (rspamd_html_process_tag (pool, hc, cur_tag, &cur_level, | |||||
| &balanced)) { | |||||
| state = content_write; | state = content_write; | ||||
| need_decode = FALSE; | need_decode = FALSE; | ||||
| } | } | ||||
| else { | else { | ||||
| state = content_ignore; | state = content_ignore; | ||||
| } | } | ||||
| if ((cur_tag->id == Tag_P || cur_tag->id == Tag_BR || | |||||
| cur_tag->id == Tag_HR) && balanced) { | |||||
| /* Insert newline */ | |||||
| g_byte_array_append (dest, "\r\n", 2); | |||||
| save_space = FALSE; | |||||
| } | |||||
| } | } | ||||
| else { | else { | ||||
| state = content_write; | state = content_write; | ||||
| } | } | ||||
| p++; | p++; | ||||
| c = p; | c = p; | ||||
| cur_tag = NULL; | cur_tag = NULL; |
+ 4
- 7
test/lua/unit/html.lua
View File
| <b>stuff</p>? | <b>stuff</p>? | ||||
| </body> | </body> | ||||
| </html> | </html> | ||||
| ]], 'Hello, world! test data stuff?'}, | |||||
| ]], "Hello, world! test\r\ndata\r\nstuff?"}, | |||||
| {[[ | {[[ | ||||
| <?xml version="1.0" encoding="iso-8859-1"?> | <?xml version="1.0" encoding="iso-8859-1"?> | ||||
| <!DOCTYPE html | <!DOCTYPE html | ||||
| </p> | </p> | ||||
| </body> | </body> | ||||
| </html>]], 'Hello, world!'}, | |||||
| </html>]], '\r\nHello, world!\r\n'}, | |||||
| {[[ | {[[ | ||||
| <!DOCTYPE html> | <!DOCTYPE html> | ||||
| <html lang="en"> | <html lang="en"> | ||||
| --></head> | --></head> | ||||
| <body> | <body> | ||||
| <!-- page content --> | <!-- page content --> | ||||
| Hello, world! <b>test</b> | |||||
| <p>data<> | |||||
| </P> | |||||
| <b>stuff</p>? | |||||
| Hello, world! | |||||
| </body> | </body> | ||||
| </html> | </html> | ||||
| ]], 'Hello, world! test data stuff?'}, | |||||
| ]], 'Hello, world!'}, | |||||
| } | } | ||||
| for _,c in ipairs(cases) do | for _,c in ipairs(cases) do |
Loading…