You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

ucl_parser.c 68KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798991001011021031041051061071081091101111121131141151161171181191201211221231241251261271281291301311321331341351361371381391401411421431441451461471481491501511521531541551561571581591601611621631641651661671681691701711721731741751761771781791801811821831841851861871881891901911921931941951961971981992002012022032042052062072082092102112122132142152162172182192202212222232242252262272282292302312322332342352362372382392402412422432442452462472482492502512522532542552562572582592602612622632642652662672682692702712722732742752762772782792802812822832842852862872882892902912922932942952962972982993003013023033043053063073083093103113123133143153163173183193203213223233243253263273283293303313323333343353363373383393403413423433443453463473483493503513523533543553563573583593603613623633643653663673683693703713723733743753763773783793803813823833843853863873883893903913923933943953963973983994004014024034044054064074084094104114124134144154164174184194204214224234244254264274284294304314324334344354364374384394404414424434444454464474484494504514524534544554564574584594604614624634644654664674684694704714724734744754764774784794804814824834844854864874884894904914924934944954964974984995005015025035045055065075085095105115125135145155165175185195205215225235245255265275285295305315325335345355365375385395405415425435445455465475485495505515525535545555565575585595605615625635645655665675685695705715725735745755765775785795805815825835845855865875885895905915925935945955965975985996006016026036046056066076086096106116126136146156166176186196206216226236246256266276286296306316326336346356366376386396406416426436446456466476486496506516526536546556566576586596606616626636646656666676686696706716726736746756766776786796806816826836846856866876886896906916926936946956966976986997007017027037047057067077087097107117127137147157167177187197207217227237247257267277287297307317327337347357367377387397407417427437447457467477487497507517527537547557567577587597607617627637647657667677687697707717727737747757767777787797807817827837847857867877887897907917927937947957967977987998008018028038048058068078088098108118128138148158168178188198208218228238248258268278288298308318328338348358368378388398408418428438448458468478488498508518528538548558568578588598608618628638648658668678688698708718728738748758768778788798808818828838848858868878888898908918928938948958968978988999009019029039049059069079089099109119129139149159169179189199209219229239249259269279289299309319329339349359369379389399409419429439449459469479489499509519529539549559569579589599609619629639649659669679689699709719729739749759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165116611671168116911701171117211731174117511761177117811791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276127712781279128012811282128312841285128612871288128912901291129212931294129512961297129812991300130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426142714281429143014311432143314341435143614371438143914401441144214431444144514461447144814491450145114521453145414551456145714581459146014611462146314641465146614671468146914701471147214731474147514761477147814791480148114821483148414851486148714881489149014911492149314941495149614971498149915001501150215031504150515061507150815091510151115121513151415151516151715181519152015211522152315241525152615271528152915301531153215331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591159215931594159515961597159815991600160116021603160416051606160716081609161016111612161316141615161616171618161916201621162216231624162516261627162816291630163116321633163416351636163716381639164016411642164316441645164616471648164916501651165216531654165516561657165816591660166116621663166416651666166716681669167016711672167316741675167616771678167916801681168216831684168516861687168816891690169116921693169416951696169716981699170017011702170317041705170617071708170917101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762176317641765176617671768176917701771177217731774177517761777177817791780178117821783178417851786178717881789179017911792179317941795179617971798179918001801180218031804180518061807180818091810181118121813181418151816181718181819182018211822182318241825182618271828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885188618871888188918901891189218931894189518961897189818991900190119021903190419051906190719081909191019111912191319141915191619171918191919201921192219231924192519261927192819291930193119321933193419351936193719381939194019411942194319441945194619471948194919501951195219531954195519561957195819591960196119621963196419651966196719681969197019711972197319741975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005200620072008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920402041204220432044204520462047204820492050205120522053205420552056205720582059206020612062206320642065206620672068206920702071207220732074207520762077207820792080208120822083208420852086208720882089209020912092209320942095209620972098209921002101210221032104210521062107210821092110211121122113211421152116211721182119212021212122212321242125212621272128212921302131213221332134213521362137213821392140214121422143214421452146214721482149215021512152215321542155215621572158215921602161216221632164216521662167216821692170217121722173217421752176217721782179218021812182218321842185218621872188218921902191219221932194219521962197219821992200220122022203220422052206220722082209221022112212221322142215221622172218221922202221222222232224222522262227222822292230223122322233223422352236223722382239224022412242224322442245224622472248224922502251225222532254225522562257225822592260226122622263226422652266226722682269227022712272227322742275227622772278227922802281228222832284228522862287228822892290229122922293229422952296229722982299230023012302230323042305230623072308230923102311231223132314231523162317231823192320232123222323232423252326232723282329233023312332233323342335233623372338233923402341234223432344234523462347234823492350235123522353235423552356235723582359236023612362236323642365236623672368236923702371237223732374237523762377237823792380238123822383238423852386238723882389239023912392239323942395239623972398239924002401240224032404240524062407240824092410241124122413241424152416241724182419242024212422242324242425242624272428242924302431243224332434243524362437243824392440244124422443244424452446244724482449245024512452245324542455245624572458245924602461246224632464246524662467246824692470247124722473247424752476247724782479248024812482248324842485248624872488248924902491249224932494249524962497249824992500250125022503250425052506250725082509251025112512251325142515251625172518251925202521252225232524252525262527252825292530253125322533253425352536253725382539254025412542254325442545254625472548254925502551255225532554255525562557255825592560256125622563256425652566256725682569257025712572257325742575257625772578257925802581258225832584258525862587258825892590259125922593259425952596259725982599260026012602260326042605260626072608260926102611261226132614261526162617261826192620262126222623262426252626262726282629263026312632263326342635263626372638263926402641264226432644264526462647264826492650265126522653265426552656265726582659266026612662266326642665266626672668266926702671267226732674267526762677267826792680268126822683268426852686268726882689269026912692269326942695269626972698269927002701270227032704270527062707270827092710271127122713271427152716271727182719272027212722272327242725272627272728272927302731273227332734273527362737273827392740274127422743274427452746274727482749275027512752275327542755275627572758275927602761276227632764276527662767276827692770277127722773277427752776277727782779278027812782278327842785278627872788278927902791279227932794279527962797279827992800280128022803280428052806280728082809281028112812281328142815281628172818281928202821282228232824282528262827282828292830283128322833283428352836283728382839284028412842284328442845284628472848284928502851285228532854285528562857285828592860286128622863286428652866286728682869287028712872287328742875287628772878287928802881288228832884288528862887288828892890289128922893289428952896289728982899290029012902290329042905290629072908290929102911291229132914291529162917291829192920292129222923292429252926292729282929293029312932293329342935293629372938293929402941294229432944294529462947294829492950295129522953295429552956295729582959296029612962296329642965296629672968296929702971297229732974297529762977297829792980298129822983298429852986298729882989299029912992299329942995299629972998299930003001300230033004300530063007300830093010301130123013301430153016301730183019302030213022302330243025302630273028302930303031303230333034303530363037303830393040304130423043304430453046304730483049305030513052305330543055305630573058305930603061306230633064306530663067306830693070307130723073307430753076307730783079308030813082308330843085308630873088308930903091309230933094309530963097309830993100310131023103310431053106310731083109311031113112311331143115311631173118311931203121312231233124312531263127312831293130313131323133313431353136
  1. /* Copyright (c) 2013, Vsevolod Stakhov
  2. * All rights reserved.
  3. *
  4. * Redistribution and use in source and binary forms, with or without
  5. * modification, are permitted provided that the following conditions are met:
  6. * * Redistributions of source code must retain the above copyright
  7. * notice, this list of conditions and the following disclaimer.
  8. * * Redistributions in binary form must reproduce the above copyright
  9. * notice, this list of conditions and the following disclaimer in the
  10. * documentation and/or other materials provided with the distribution.
  11. *
  12. * THIS SOFTWARE IS PROVIDED ''AS IS'' AND ANY
  13. * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  14. * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
  15. * DISCLAIMED. IN NO EVENT SHALL AUTHOR BE LIABLE FOR ANY
  16. * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
  17. * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  18. * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
  19. * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  20. * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
  21. * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  22. */
  23. #include <math.h>
  24. #include "ucl.h"
  25. #include "ucl_internal.h"
  26. #include "ucl_chartable.h"
  27. /**
  28. * @file ucl_parser.c
  29. * The implementation of ucl parser
  30. */
  31. struct ucl_parser_saved_state {
  32. unsigned int line;
  33. unsigned int column;
  34. size_t remain;
  35. const unsigned char *pos;
  36. };
  37. /**
  38. * Move up to len characters
  39. * @param parser
  40. * @param begin
  41. * @param len
  42. * @return new position in chunk
  43. */
  44. #define ucl_chunk_skipc(chunk, p) \
  45. do { \
  46. if (*(p) == '\n') { \
  47. (chunk)->line ++; \
  48. (chunk)->column = 0; \
  49. } \
  50. else (chunk)->column ++; \
  51. (p++); \
  52. (chunk)->pos ++; \
  53. (chunk)->remain --; \
  54. } while (0)
  55. static inline void
  56. ucl_set_err (struct ucl_parser *parser, int code, const char *str, UT_string **err)
  57. {
  58. const char *fmt_string, *filename;
  59. struct ucl_chunk *chunk = parser->chunks;
  60. if (parser->cur_file) {
  61. filename = parser->cur_file;
  62. }
  63. else {
  64. filename = "<unknown>";
  65. }
  66. if (chunk->pos < chunk->end) {
  67. if (isgraph (*chunk->pos)) {
  68. fmt_string = "error while parsing %s: "
  69. "line: %d, column: %d - '%s', character: '%c'";
  70. }
  71. else {
  72. fmt_string = "error while parsing %s: "
  73. "line: %d, column: %d - '%s', character: '0x%02x'";
  74. }
  75. ucl_create_err (err, fmt_string,
  76. filename, chunk->line, chunk->column,
  77. str, *chunk->pos);
  78. }
  79. else {
  80. ucl_create_err (err, "error while parsing %s: at the end of chunk: %s",
  81. filename, str);
  82. }
  83. parser->err_code = code;
  84. parser->state = UCL_STATE_ERROR;
  85. }
  86. static void
  87. ucl_save_comment (struct ucl_parser *parser, const char *begin, size_t len)
  88. {
  89. ucl_object_t *nobj;
  90. if (len > 0 && begin != NULL) {
  91. nobj = ucl_object_fromstring_common (begin, len, 0);
  92. if (parser->last_comment) {
  93. /* We need to append data to an existing object */
  94. DL_APPEND (parser->last_comment, nobj);
  95. }
  96. else {
  97. parser->last_comment = nobj;
  98. }
  99. }
  100. }
  101. static void
  102. ucl_attach_comment (struct ucl_parser *parser, ucl_object_t *obj, bool before)
  103. {
  104. if (parser->last_comment) {
  105. ucl_object_insert_key (parser->comments, parser->last_comment,
  106. (const char *)&obj, sizeof (void *), true);
  107. if (before) {
  108. parser->last_comment->flags |= UCL_OBJECT_INHERITED;
  109. }
  110. parser->last_comment = NULL;
  111. }
  112. }
  113. /**
  114. * Skip all comments from the current pos resolving nested and multiline comments
  115. * @param parser
  116. * @return
  117. */
  118. static bool
  119. ucl_skip_comments (struct ucl_parser *parser)
  120. {
  121. struct ucl_chunk *chunk = parser->chunks;
  122. const unsigned char *p, *beg = NULL;
  123. int comments_nested = 0;
  124. bool quoted = false;
  125. p = chunk->pos;
  126. start:
  127. if (chunk->remain > 0 && *p == '#') {
  128. if (parser->state != UCL_STATE_SCOMMENT &&
  129. parser->state != UCL_STATE_MCOMMENT) {
  130. beg = p;
  131. while (p < chunk->end) {
  132. if (*p == '\n') {
  133. if (parser->flags & UCL_PARSER_SAVE_COMMENTS) {
  134. ucl_save_comment (parser, beg, p - beg);
  135. beg = NULL;
  136. }
  137. ucl_chunk_skipc (chunk, p);
  138. goto start;
  139. }
  140. ucl_chunk_skipc (chunk, p);
  141. }
  142. }
  143. }
  144. else if (chunk->remain >= 2 && *p == '/') {
  145. if (p[1] == '*') {
  146. beg = p;
  147. ucl_chunk_skipc (chunk, p);
  148. comments_nested ++;
  149. ucl_chunk_skipc (chunk, p);
  150. while (p < chunk->end) {
  151. if (*p == '"' && *(p - 1) != '\\') {
  152. quoted = !quoted;
  153. }
  154. if (!quoted) {
  155. if (*p == '*') {
  156. ucl_chunk_skipc (chunk, p);
  157. if (*p == '/') {
  158. comments_nested --;
  159. if (comments_nested == 0) {
  160. if (parser->flags & UCL_PARSER_SAVE_COMMENTS) {
  161. ucl_save_comment (parser, beg, p - beg + 1);
  162. beg = NULL;
  163. }
  164. ucl_chunk_skipc (chunk, p);
  165. goto start;
  166. }
  167. }
  168. ucl_chunk_skipc (chunk, p);
  169. }
  170. else if (p[0] == '/' && chunk->remain >= 2 && p[1] == '*') {
  171. comments_nested ++;
  172. ucl_chunk_skipc (chunk, p);
  173. ucl_chunk_skipc (chunk, p);
  174. continue;
  175. }
  176. }
  177. ucl_chunk_skipc (chunk, p);
  178. }
  179. if (comments_nested != 0) {
  180. ucl_set_err (parser, UCL_ENESTED,
  181. "unfinished multiline comment", &parser->err);
  182. return false;
  183. }
  184. }
  185. }
  186. if (beg && p > beg && (parser->flags & UCL_PARSER_SAVE_COMMENTS)) {
  187. ucl_save_comment (parser, beg, p - beg);
  188. }
  189. return true;
  190. }
  191. /**
  192. * Return multiplier for a character
  193. * @param c multiplier character
  194. * @param is_bytes if true use 1024 multiplier
  195. * @return multiplier
  196. */
  197. static inline unsigned long
  198. ucl_lex_num_multiplier (const unsigned char c, bool is_bytes) {
  199. const struct {
  200. char c;
  201. long mult_normal;
  202. long mult_bytes;
  203. } multipliers[] = {
  204. {'m', 1000 * 1000, 1024 * 1024},
  205. {'k', 1000, 1024},
  206. {'g', 1000 * 1000 * 1000, 1024 * 1024 * 1024}
  207. };
  208. int i;
  209. for (i = 0; i < 3; i ++) {
  210. if (tolower (c) == multipliers[i].c) {
  211. if (is_bytes) {
  212. return multipliers[i].mult_bytes;
  213. }
  214. return multipliers[i].mult_normal;
  215. }
  216. }
  217. return 1;
  218. }
  219. /**
  220. * Return multiplier for time scaling
  221. * @param c
  222. * @return
  223. */
  224. static inline double
  225. ucl_lex_time_multiplier (const unsigned char c) {
  226. const struct {
  227. char c;
  228. double mult;
  229. } multipliers[] = {
  230. {'m', 60},
  231. {'h', 60 * 60},
  232. {'d', 60 * 60 * 24},
  233. {'w', 60 * 60 * 24 * 7},
  234. {'y', 60 * 60 * 24 * 365}
  235. };
  236. int i;
  237. for (i = 0; i < 5; i ++) {
  238. if (tolower (c) == multipliers[i].c) {
  239. return multipliers[i].mult;
  240. }
  241. }
  242. return 1;
  243. }
  244. /**
  245. * Return true if a character is a end of an atom
  246. * @param c
  247. * @return
  248. */
  249. static inline bool
  250. ucl_lex_is_atom_end (const unsigned char c)
  251. {
  252. return ucl_test_character (c, UCL_CHARACTER_VALUE_END);
  253. }
  254. static inline bool
  255. ucl_lex_is_comment (const unsigned char c1, const unsigned char c2)
  256. {
  257. if (c1 == '/') {
  258. if (c2 == '*') {
  259. return true;
  260. }
  261. }
  262. else if (c1 == '#') {
  263. return true;
  264. }
  265. return false;
  266. }
  267. /**
  268. * Check variable found
  269. * @param parser
  270. * @param ptr
  271. * @param remain
  272. * @param out_len
  273. * @param strict
  274. * @param found
  275. * @return
  276. */
  277. static inline const char *
  278. ucl_check_variable_safe (struct ucl_parser *parser, const char *ptr, size_t remain,
  279. size_t *out_len, bool strict, bool *found)
  280. {
  281. struct ucl_variable *var;
  282. unsigned char *dst;
  283. size_t dstlen;
  284. bool need_free = false;
  285. LL_FOREACH (parser->variables, var) {
  286. if (strict) {
  287. if (remain == var->var_len) {
  288. if (memcmp (ptr, var->var, var->var_len) == 0) {
  289. *out_len += var->value_len;
  290. *found = true;
  291. return (ptr + var->var_len);
  292. }
  293. }
  294. }
  295. else {
  296. if (remain >= var->var_len) {
  297. if (memcmp (ptr, var->var, var->var_len) == 0) {
  298. *out_len += var->value_len;
  299. *found = true;
  300. return (ptr + var->var_len);
  301. }
  302. }
  303. }
  304. }
  305. /* XXX: can only handle ${VAR} */
  306. if (!(*found) && parser->var_handler != NULL && strict) {
  307. /* Call generic handler */
  308. if (parser->var_handler (ptr, remain, &dst, &dstlen, &need_free,
  309. parser->var_data)) {
  310. *found = true;
  311. if (need_free) {
  312. free (dst);
  313. }
  314. return (ptr + remain);
  315. }
  316. }
  317. return ptr;
  318. }
  319. /**
  320. * Check for a variable in a given string
  321. * @param parser
  322. * @param ptr
  323. * @param remain
  324. * @param out_len
  325. * @param vars_found
  326. * @return
  327. */
  328. static const char *
  329. ucl_check_variable (struct ucl_parser *parser, const char *ptr,
  330. size_t remain, size_t *out_len, bool *vars_found)
  331. {
  332. const char *p, *end, *ret = ptr;
  333. bool found = false;
  334. if (*ptr == '{') {
  335. /* We need to match the variable enclosed in braces */
  336. p = ptr + 1;
  337. end = ptr + remain;
  338. while (p < end) {
  339. if (*p == '}') {
  340. ret = ucl_check_variable_safe (parser, ptr + 1, p - ptr - 1,
  341. out_len, true, &found);
  342. if (found) {
  343. /* {} must be excluded actually */
  344. ret ++;
  345. if (!*vars_found) {
  346. *vars_found = true;
  347. }
  348. }
  349. else {
  350. *out_len += 2;
  351. }
  352. break;
  353. }
  354. p ++;
  355. }
  356. }
  357. else if (*ptr != '$') {
  358. /* Not count escaped dollar sign */
  359. ret = ucl_check_variable_safe (parser, ptr, remain, out_len, false, &found);
  360. if (found && !*vars_found) {
  361. *vars_found = true;
  362. }
  363. if (!found) {
  364. (*out_len) ++;
  365. }
  366. }
  367. else {
  368. ret ++;
  369. (*out_len) ++;
  370. }
  371. return ret;
  372. }
  373. /**
  374. * Expand a single variable
  375. * @param parser
  376. * @param ptr
  377. * @param remain
  378. * @param dest
  379. * @return
  380. */
  381. static const char *
  382. ucl_expand_single_variable (struct ucl_parser *parser, const char *ptr,
  383. size_t remain, unsigned char **dest)
  384. {
  385. unsigned char *d = *dest, *dst;
  386. const char *p = ptr + 1, *ret;
  387. struct ucl_variable *var;
  388. size_t dstlen;
  389. bool need_free = false;
  390. bool found = false;
  391. bool strict = false;
  392. ret = ptr + 1;
  393. remain --;
  394. if (*p == '$') {
  395. *d++ = *p++;
  396. *dest = d;
  397. return p;
  398. }
  399. else if (*p == '{') {
  400. p ++;
  401. strict = true;
  402. ret += 2;
  403. remain -= 2;
  404. }
  405. LL_FOREACH (parser->variables, var) {
  406. if (remain >= var->var_len) {
  407. if (memcmp (p, var->var, var->var_len) == 0) {
  408. memcpy (d, var->value, var->value_len);
  409. ret += var->var_len;
  410. d += var->value_len;
  411. found = true;
  412. break;
  413. }
  414. }
  415. }
  416. if (!found) {
  417. if (strict && parser->var_handler != NULL) {
  418. if (parser->var_handler (p, remain, &dst, &dstlen, &need_free,
  419. parser->var_data)) {
  420. memcpy (d, dst, dstlen);
  421. ret += remain;
  422. d += dstlen;
  423. found = true;
  424. if (need_free) {
  425. free (dst);
  426. }
  427. }
  428. }
  429. /* Leave variable as is */
  430. if (!found) {
  431. if (strict) {
  432. /* Copy '${' */
  433. memcpy (d, ptr, 2);
  434. d += 2;
  435. ret --;
  436. }
  437. else {
  438. memcpy (d, ptr, 1);
  439. d ++;
  440. }
  441. }
  442. }
  443. *dest = d;
  444. return ret;
  445. }
  446. /**
  447. * Expand variables in string
  448. * @param parser
  449. * @param dst
  450. * @param src
  451. * @param in_len
  452. * @return
  453. */
  454. static ssize_t
  455. ucl_expand_variable (struct ucl_parser *parser, unsigned char **dst,
  456. const char *src, size_t in_len)
  457. {
  458. const char *p, *end = src + in_len;
  459. unsigned char *d;
  460. size_t out_len = 0;
  461. bool vars_found = false;
  462. if (parser->flags & UCL_PARSER_DISABLE_MACRO) {
  463. *dst = NULL;
  464. return in_len;
  465. }
  466. p = src;
  467. while (p != end) {
  468. if (*p == '$') {
  469. p = ucl_check_variable (parser, p + 1, end - p - 1, &out_len, &vars_found);
  470. }
  471. else {
  472. p ++;
  473. out_len ++;
  474. }
  475. }
  476. if (!vars_found) {
  477. /* Trivial case */
  478. *dst = NULL;
  479. return in_len;
  480. }
  481. *dst = UCL_ALLOC (out_len + 1);
  482. if (*dst == NULL) {
  483. return in_len;
  484. }
  485. d = *dst;
  486. p = src;
  487. while (p != end) {
  488. if (*p == '$') {
  489. p = ucl_expand_single_variable (parser, p, end - p, &d);
  490. }
  491. else {
  492. *d++ = *p++;
  493. }
  494. }
  495. *d = '\0';
  496. return out_len;
  497. }
  498. /**
  499. * Store or copy pointer to the trash stack
  500. * @param parser parser object
  501. * @param src src string
  502. * @param dst destination buffer (trash stack pointer)
  503. * @param dst_const const destination pointer (e.g. value of object)
  504. * @param in_len input length
  505. * @param need_unescape need to unescape source (and copy it)
  506. * @param need_lowercase need to lowercase value (and copy)
  507. * @param need_expand need to expand variables (and copy as well)
  508. * @param unescape_squote unescape single quoted string
  509. * @return output length (excluding \0 symbol)
  510. */
  511. static inline ssize_t
  512. ucl_copy_or_store_ptr (struct ucl_parser *parser,
  513. const unsigned char *src, unsigned char **dst,
  514. const char **dst_const, size_t in_len,
  515. bool need_unescape, bool need_lowercase, bool need_expand,
  516. bool unescape_squote)
  517. {
  518. ssize_t ret = -1, tret;
  519. unsigned char *tmp;
  520. if (need_unescape || need_lowercase ||
  521. (need_expand && parser->variables != NULL) ||
  522. !(parser->flags & UCL_PARSER_ZEROCOPY)) {
  523. /* Copy string */
  524. *dst = UCL_ALLOC (in_len + 1);
  525. if (*dst == NULL) {
  526. ucl_set_err (parser, UCL_EINTERNAL, "cannot allocate memory for a string",
  527. &parser->err);
  528. return false;
  529. }
  530. if (need_lowercase) {
  531. ret = ucl_strlcpy_tolower (*dst, src, in_len + 1);
  532. }
  533. else {
  534. ret = ucl_strlcpy_unsafe (*dst, src, in_len + 1);
  535. }
  536. if (need_unescape) {
  537. if (!unescape_squote) {
  538. ret = ucl_unescape_json_string (*dst, ret);
  539. }
  540. else {
  541. ret = ucl_unescape_squoted_string (*dst, ret);
  542. }
  543. }
  544. if (need_expand) {
  545. tmp = *dst;
  546. tret = ret;
  547. ret = ucl_expand_variable (parser, dst, tmp, ret);
  548. if (*dst == NULL) {
  549. /* Nothing to expand */
  550. *dst = tmp;
  551. ret = tret;
  552. }
  553. else {
  554. /* Free unexpanded value */
  555. UCL_FREE (in_len + 1, tmp);
  556. }
  557. }
  558. *dst_const = *dst;
  559. }
  560. else {
  561. *dst_const = src;
  562. ret = in_len;
  563. }
  564. return ret;
  565. }
  566. /**
  567. * Create and append an object at the specified level
  568. * @param parser
  569. * @param is_array
  570. * @param level
  571. * @return
  572. */
  573. static inline ucl_object_t *
  574. ucl_parser_add_container (ucl_object_t *obj, struct ucl_parser *parser,
  575. bool is_array, uint32_t level, bool has_obrace)
  576. {
  577. struct ucl_stack *st;
  578. ucl_object_t *nobj;
  579. if (obj == NULL) {
  580. nobj = ucl_object_new_full (is_array ? UCL_ARRAY : UCL_OBJECT, parser->chunks->priority);
  581. if (nobj == NULL) {
  582. goto enomem0;
  583. }
  584. } else {
  585. if (obj->type == (is_array ? UCL_OBJECT : UCL_ARRAY)) {
  586. /* Bad combination for merge: array and object */
  587. ucl_set_err (parser, UCL_EMERGE,
  588. "cannot merge an object with an array",
  589. &parser->err);
  590. return NULL;
  591. }
  592. nobj = obj;
  593. nobj->type = is_array ? UCL_ARRAY : UCL_OBJECT;
  594. }
  595. if (!is_array) {
  596. if (nobj->value.ov == NULL) {
  597. nobj->value.ov = ucl_hash_create (parser->flags & UCL_PARSER_KEY_LOWERCASE);
  598. if (nobj->value.ov == NULL) {
  599. goto enomem1;
  600. }
  601. }
  602. parser->state = UCL_STATE_KEY;
  603. } else {
  604. parser->state = UCL_STATE_VALUE;
  605. }
  606. st = UCL_ALLOC (sizeof (struct ucl_stack));
  607. if (st == NULL) {
  608. goto enomem1;
  609. }
  610. st->obj = nobj;
  611. if (level >= UINT16_MAX) {
  612. ucl_set_err (parser, UCL_ENESTED,
  613. "objects are nesting too deep (over 65535 limit)",
  614. &parser->err);
  615. if (nobj != obj) {
  616. ucl_object_unref (obj);
  617. }
  618. return NULL;
  619. }
  620. st->e.params.level = level;
  621. st->e.params.line = parser->chunks->line;
  622. st->chunk = parser->chunks;
  623. if (has_obrace) {
  624. st->e.params.flags = UCL_STACK_HAS_OBRACE;
  625. }
  626. else {
  627. st->e.params.flags = 0;
  628. }
  629. LL_PREPEND (parser->stack, st);
  630. parser->cur_obj = nobj;
  631. return nobj;
  632. enomem1:
  633. if (nobj != obj)
  634. ucl_object_unref (nobj);
  635. enomem0:
  636. ucl_set_err (parser, UCL_EINTERNAL, "cannot allocate memory for an object",
  637. &parser->err);
  638. return NULL;
  639. }
  640. int
  641. ucl_maybe_parse_number (ucl_object_t *obj,
  642. const char *start, const char *end, const char **pos,
  643. bool allow_double, bool number_bytes, bool allow_time)
  644. {
  645. const char *p = start, *c = start;
  646. char *endptr;
  647. bool got_dot = false, got_exp = false, need_double = false,
  648. is_time = false, valid_start = false, is_hex = false,
  649. is_neg = false;
  650. double dv = 0;
  651. int64_t lv = 0;
  652. if (*p == '-') {
  653. is_neg = true;
  654. c ++;
  655. p ++;
  656. }
  657. while (p < end) {
  658. if (is_hex && isxdigit (*p)) {
  659. p ++;
  660. }
  661. else if (isdigit (*p)) {
  662. valid_start = true;
  663. p ++;
  664. }
  665. else if (!is_hex && (*p == 'x' || *p == 'X')) {
  666. is_hex = true;
  667. allow_double = false;
  668. c = p + 1;
  669. }
  670. else if (allow_double) {
  671. if (p == c) {
  672. /* Empty digits sequence, not a number */
  673. *pos = start;
  674. return EINVAL;
  675. }
  676. else if (*p == '.') {
  677. if (got_dot) {
  678. /* Double dots, not a number */
  679. *pos = start;
  680. return EINVAL;
  681. }
  682. else {
  683. got_dot = true;
  684. need_double = true;
  685. p ++;
  686. }
  687. }
  688. else if (*p == 'e' || *p == 'E') {
  689. if (got_exp) {
  690. /* Double exp, not a number */
  691. *pos = start;
  692. return EINVAL;
  693. }
  694. else {
  695. got_exp = true;
  696. need_double = true;
  697. p ++;
  698. if (p >= end) {
  699. *pos = start;
  700. return EINVAL;
  701. }
  702. if (!isdigit (*p) && *p != '+' && *p != '-') {
  703. /* Wrong exponent sign */
  704. *pos = start;
  705. return EINVAL;
  706. }
  707. else {
  708. p ++;
  709. }
  710. }
  711. }
  712. else {
  713. /* Got the end of the number, need to check */
  714. break;
  715. }
  716. }
  717. else {
  718. break;
  719. }
  720. }
  721. if (!valid_start) {
  722. *pos = start;
  723. return EINVAL;
  724. }
  725. errno = 0;
  726. if (need_double) {
  727. dv = strtod (c, &endptr);
  728. }
  729. else {
  730. if (is_hex) {
  731. lv = strtoimax (c, &endptr, 16);
  732. }
  733. else {
  734. lv = strtoimax (c, &endptr, 10);
  735. }
  736. }
  737. if (errno == ERANGE) {
  738. *pos = start;
  739. return ERANGE;
  740. }
  741. /* Now check endptr */
  742. if (endptr == NULL || ucl_lex_is_atom_end (*endptr) || *endptr == '\0') {
  743. p = endptr;
  744. goto set_obj;
  745. }
  746. if (endptr < end && endptr != start) {
  747. p = endptr;
  748. switch (*p) {
  749. case 'm':
  750. case 'M':
  751. case 'g':
  752. case 'G':
  753. case 'k':
  754. case 'K':
  755. if (end - p >= 2) {
  756. if (p[1] == 's' || p[1] == 'S') {
  757. /* Milliseconds */
  758. if (!need_double) {
  759. need_double = true;
  760. dv = lv;
  761. }
  762. is_time = true;
  763. if (p[0] == 'm' || p[0] == 'M') {
  764. dv /= 1000.;
  765. }
  766. else {
  767. dv *= ucl_lex_num_multiplier (*p, false);
  768. }
  769. p += 2;
  770. goto set_obj;
  771. }
  772. else if (number_bytes || (p[1] == 'b' || p[1] == 'B')) {
  773. /* Bytes */
  774. if (need_double) {
  775. need_double = false;
  776. lv = dv;
  777. }
  778. lv *= ucl_lex_num_multiplier (*p, true);
  779. p += 2;
  780. goto set_obj;
  781. }
  782. else if (ucl_lex_is_atom_end (p[1])) {
  783. if (need_double) {
  784. dv *= ucl_lex_num_multiplier (*p, false);
  785. }
  786. else {
  787. lv *= ucl_lex_num_multiplier (*p, number_bytes);
  788. }
  789. p ++;
  790. goto set_obj;
  791. }
  792. else if (allow_time && end - p >= 3) {
  793. if (tolower (p[0]) == 'm' &&
  794. tolower (p[1]) == 'i' &&
  795. tolower (p[2]) == 'n') {
  796. /* Minutes */
  797. if (!need_double) {
  798. need_double = true;
  799. dv = lv;
  800. }
  801. is_time = true;
  802. dv *= 60.;
  803. p += 3;
  804. goto set_obj;
  805. }
  806. }
  807. }
  808. else {
  809. if (need_double) {
  810. dv *= ucl_lex_num_multiplier (*p, false);
  811. }
  812. else {
  813. lv *= ucl_lex_num_multiplier (*p, number_bytes);
  814. }
  815. p ++;
  816. goto set_obj;
  817. }
  818. break;
  819. case 'S':
  820. case 's':
  821. if (allow_time &&
  822. (p == end - 1 || ucl_lex_is_atom_end (p[1]))) {
  823. if (!need_double) {
  824. need_double = true;
  825. dv = lv;
  826. }
  827. p ++;
  828. is_time = true;
  829. goto set_obj;
  830. }
  831. break;
  832. case 'h':
  833. case 'H':
  834. case 'd':
  835. case 'D':
  836. case 'w':
  837. case 'W':
  838. case 'Y':
  839. case 'y':
  840. if (allow_time &&
  841. (p == end - 1 || ucl_lex_is_atom_end (p[1]))) {
  842. if (!need_double) {
  843. need_double = true;
  844. dv = lv;
  845. }
  846. is_time = true;
  847. dv *= ucl_lex_time_multiplier (*p);
  848. p ++;
  849. goto set_obj;
  850. }
  851. break;
  852. case '\t':
  853. case ' ':
  854. while (p < end && ucl_test_character(*p, UCL_CHARACTER_WHITESPACE)) {
  855. p++;
  856. }
  857. if (ucl_lex_is_atom_end(*p))
  858. goto set_obj;
  859. break;
  860. }
  861. }
  862. else if (endptr == end) {
  863. /* Just a number at the end of chunk */
  864. p = endptr;
  865. goto set_obj;
  866. }
  867. *pos = c;
  868. return EINVAL;
  869. set_obj:
  870. if (obj != NULL) {
  871. if (allow_double && (need_double || is_time)) {
  872. if (!is_time) {
  873. obj->type = UCL_FLOAT;
  874. }
  875. else {
  876. obj->type = UCL_TIME;
  877. }
  878. obj->value.dv = is_neg ? (-dv) : dv;
  879. }
  880. else {
  881. obj->type = UCL_INT;
  882. obj->value.iv = is_neg ? (-lv) : lv;
  883. }
  884. }
  885. *pos = p;
  886. return 0;
  887. }
  888. /**
  889. * Parse possible number
  890. * @param parser
  891. * @param chunk
  892. * @param obj
  893. * @return true if a number has been parsed
  894. */
  895. static bool
  896. ucl_lex_number (struct ucl_parser *parser,
  897. struct ucl_chunk *chunk, ucl_object_t *obj)
  898. {
  899. const unsigned char *pos;
  900. int ret;
  901. ret = ucl_maybe_parse_number (obj, chunk->pos, chunk->end, (const char **)&pos,
  902. true, false, ((parser->flags & UCL_PARSER_NO_TIME) == 0));
  903. if (ret == 0) {
  904. chunk->remain -= pos - chunk->pos;
  905. chunk->column += pos - chunk->pos;
  906. chunk->pos = pos;
  907. return true;
  908. }
  909. else if (ret == ERANGE) {
  910. ucl_set_err (parser, UCL_ESYNTAX, "numeric value out of range",
  911. &parser->err);
  912. }
  913. return false;
  914. }
  915. /**
  916. * Parse quoted string with possible escapes
  917. * @param parser
  918. * @param chunk
  919. * @param need_unescape
  920. * @param ucl_escape
  921. * @param var_expand
  922. * @return true if a string has been parsed
  923. */
  924. static bool
  925. ucl_lex_json_string (struct ucl_parser *parser,
  926. struct ucl_chunk *chunk,
  927. bool *need_unescape,
  928. bool *ucl_escape,
  929. bool *var_expand)
  930. {
  931. const unsigned char *p = chunk->pos;
  932. unsigned char c;
  933. int i;
  934. while (p < chunk->end) {
  935. c = *p;
  936. if (c < 0x1F) {
  937. /* Unmasked control character */
  938. if (c == '\n') {
  939. ucl_set_err (parser, UCL_ESYNTAX, "unexpected newline",
  940. &parser->err);
  941. }
  942. else {
  943. ucl_set_err (parser, UCL_ESYNTAX, "unexpected control character",
  944. &parser->err);
  945. }
  946. return false;
  947. }
  948. else if (c == '\\') {
  949. ucl_chunk_skipc (chunk, p);
  950. c = *p;
  951. if (p >= chunk->end) {
  952. ucl_set_err (parser, UCL_ESYNTAX, "unfinished escape character",
  953. &parser->err);
  954. return false;
  955. }
  956. else if (ucl_test_character (c, UCL_CHARACTER_ESCAPE)) {
  957. if (c == 'u') {
  958. ucl_chunk_skipc (chunk, p);
  959. for (i = 0; i < 4 && p < chunk->end; i ++) {
  960. if (!isxdigit (*p)) {
  961. ucl_set_err (parser, UCL_ESYNTAX, "invalid utf escape",
  962. &parser->err);
  963. return false;
  964. }
  965. ucl_chunk_skipc (chunk, p);
  966. }
  967. if (p >= chunk->end) {
  968. ucl_set_err (parser, UCL_ESYNTAX,
  969. "unfinished escape character",
  970. &parser->err);
  971. return false;
  972. }
  973. }
  974. else {
  975. ucl_chunk_skipc (chunk, p);
  976. }
  977. }
  978. *need_unescape = true;
  979. *ucl_escape = true;
  980. continue;
  981. }
  982. else if (c == '"') {
  983. ucl_chunk_skipc (chunk, p);
  984. return true;
  985. }
  986. else if (ucl_test_character (c, UCL_CHARACTER_UCL_UNSAFE)) {
  987. *ucl_escape = true;
  988. }
  989. else if (c == '$') {
  990. *var_expand = true;
  991. }
  992. ucl_chunk_skipc (chunk, p);
  993. }
  994. ucl_set_err (parser, UCL_ESYNTAX,
  995. "no quote at the end of json string",
  996. &parser->err);
  997. return false;
  998. }
  999. /**
  1000. * Process single quoted string
  1001. * @param parser
  1002. * @param chunk
  1003. * @param need_unescape
  1004. * @return
  1005. */
  1006. static bool
  1007. ucl_lex_squoted_string (struct ucl_parser *parser,
  1008. struct ucl_chunk *chunk, bool *need_unescape)
  1009. {
  1010. const unsigned char *p = chunk->pos;
  1011. unsigned char c;
  1012. while (p < chunk->end) {
  1013. c = *p;
  1014. if (c == '\\') {
  1015. ucl_chunk_skipc (chunk, p);
  1016. if (p >= chunk->end) {
  1017. ucl_set_err (parser, UCL_ESYNTAX,
  1018. "unfinished escape character",
  1019. &parser->err);
  1020. return false;
  1021. }
  1022. else {
  1023. ucl_chunk_skipc (chunk, p);
  1024. }
  1025. *need_unescape = true;
  1026. continue;
  1027. }
  1028. else if (c == '\'') {
  1029. ucl_chunk_skipc (chunk, p);
  1030. return true;
  1031. }
  1032. ucl_chunk_skipc (chunk, p);
  1033. }
  1034. ucl_set_err (parser, UCL_ESYNTAX,
  1035. "no quote at the end of single quoted string",
  1036. &parser->err);
  1037. return false;
  1038. }
  1039. static void
  1040. ucl_parser_append_elt (struct ucl_parser *parser, ucl_hash_t *cont,
  1041. ucl_object_t *top,
  1042. ucl_object_t *elt)
  1043. {
  1044. ucl_object_t *nobj;
  1045. if ((parser->flags & UCL_PARSER_NO_IMPLICIT_ARRAYS) == 0) {
  1046. /* Implicit array */
  1047. top->flags |= UCL_OBJECT_MULTIVALUE;
  1048. DL_APPEND (top, elt);
  1049. parser->stack->obj->len ++;
  1050. }
  1051. else {
  1052. if ((top->flags & UCL_OBJECT_MULTIVALUE) != 0) {
  1053. /* Just add to the explicit array */
  1054. ucl_array_append (top, elt);
  1055. }
  1056. else {
  1057. /* Convert to an array */
  1058. nobj = ucl_object_typed_new (UCL_ARRAY);
  1059. nobj->key = top->key;
  1060. nobj->keylen = top->keylen;
  1061. nobj->flags |= UCL_OBJECT_MULTIVALUE;
  1062. ucl_array_append (nobj, top);
  1063. ucl_array_append (nobj, elt);
  1064. ucl_hash_replace (cont, top, nobj);
  1065. }
  1066. }
  1067. }
  1068. bool
  1069. ucl_parser_process_object_element (struct ucl_parser *parser, ucl_object_t *nobj)
  1070. {
  1071. ucl_hash_t *container;
  1072. ucl_object_t *tobj = NULL, *cur;
  1073. char errmsg[256];
  1074. container = parser->stack->obj->value.ov;
  1075. DL_FOREACH (parser->stack->obj, cur) {
  1076. tobj = __DECONST (ucl_object_t *, ucl_hash_search_obj (cur->value.ov, nobj));
  1077. if (tobj != NULL) {
  1078. break;
  1079. }
  1080. }
  1081. if (tobj == NULL) {
  1082. container = ucl_hash_insert_object (container, nobj,
  1083. parser->flags & UCL_PARSER_KEY_LOWERCASE);
  1084. if (container == NULL) {
  1085. return false;
  1086. }
  1087. nobj->prev = nobj;
  1088. nobj->next = NULL;
  1089. parser->stack->obj->len ++;
  1090. }
  1091. else {
  1092. unsigned priold = ucl_object_get_priority (tobj),
  1093. prinew = ucl_object_get_priority (nobj);
  1094. switch (parser->chunks->strategy) {
  1095. case UCL_DUPLICATE_APPEND:
  1096. /*
  1097. * The logic here is the following:
  1098. *
  1099. * - if we have two objects with the same priority, then we form an
  1100. * implicit or explicit array
  1101. * - if a new object has bigger priority, then we overwrite an old one
  1102. * - if a new object has lower priority, then we ignore it
  1103. */
  1104. /* Special case for inherited objects */
  1105. if (tobj->flags & UCL_OBJECT_INHERITED) {
  1106. prinew = priold + 1;
  1107. }
  1108. if (priold == prinew) {
  1109. ucl_parser_append_elt (parser, container, tobj, nobj);
  1110. }
  1111. else if (priold > prinew) {
  1112. /*
  1113. * We add this new object to a list of trash objects just to ensure
  1114. * that it won't come to any real object
  1115. * XXX: rather inefficient approach
  1116. */
  1117. DL_APPEND (parser->trash_objs, nobj);
  1118. }
  1119. else {
  1120. ucl_hash_replace (container, tobj, nobj);
  1121. ucl_object_unref (tobj);
  1122. }
  1123. break;
  1124. case UCL_DUPLICATE_REWRITE:
  1125. /* We just rewrite old values regardless of priority */
  1126. ucl_hash_replace (container, tobj, nobj);
  1127. ucl_object_unref (tobj);
  1128. break;
  1129. case UCL_DUPLICATE_ERROR:
  1130. snprintf(errmsg, sizeof(errmsg),
  1131. "duplicate element for key '%s' found",
  1132. nobj->key);
  1133. ucl_set_err (parser, UCL_EMERGE, errmsg, &parser->err);
  1134. return false;
  1135. case UCL_DUPLICATE_MERGE:
  1136. /*
  1137. * Here we do have some old object so we just push it on top of objects stack
  1138. * Check priority and then perform the merge on the remaining objects
  1139. */
  1140. if (tobj->type == UCL_OBJECT || tobj->type == UCL_ARRAY) {
  1141. ucl_object_unref (nobj);
  1142. nobj = tobj;
  1143. }
  1144. else if (priold == prinew) {
  1145. ucl_parser_append_elt (parser, container, tobj, nobj);
  1146. }
  1147. else if (priold > prinew) {
  1148. /*
  1149. * We add this new object to a list of trash objects just to ensure
  1150. * that it won't come to any real object
  1151. * XXX: rather inefficient approach
  1152. */
  1153. DL_APPEND (parser->trash_objs, nobj);
  1154. }
  1155. else {
  1156. ucl_hash_replace (container, tobj, nobj);
  1157. ucl_object_unref (tobj);
  1158. }
  1159. break;
  1160. }
  1161. }
  1162. parser->stack->obj->value.ov = container;
  1163. parser->cur_obj = nobj;
  1164. ucl_attach_comment (parser, nobj, false);
  1165. return true;
  1166. }
  1167. /**
  1168. * Parse a key in an object
  1169. * @param parser
  1170. * @param chunk
  1171. * @param next_key
  1172. * @param end_of_object
  1173. * @return true if a key has been parsed
  1174. */
  1175. static bool
  1176. ucl_parse_key (struct ucl_parser *parser, struct ucl_chunk *chunk,
  1177. bool *next_key, bool *end_of_object)
  1178. {
  1179. const unsigned char *p, *c = NULL, *end, *t;
  1180. const char *key = NULL;
  1181. bool got_quote = false, got_eq = false, got_semicolon = false,
  1182. need_unescape = false, ucl_escape = false, var_expand = false,
  1183. got_content = false, got_sep = false;
  1184. ucl_object_t *nobj;
  1185. ssize_t keylen;
  1186. p = chunk->pos;
  1187. if (*p == '.') {
  1188. /* It is macro actually */
  1189. if (!(parser->flags & UCL_PARSER_DISABLE_MACRO)) {
  1190. ucl_chunk_skipc (chunk, p);
  1191. }
  1192. parser->prev_state = parser->state;
  1193. parser->state = UCL_STATE_MACRO_NAME;
  1194. *end_of_object = false;
  1195. return true;
  1196. }
  1197. while (p < chunk->end) {
  1198. /*
  1199. * A key must start with alpha, number, '/' or '_' and end with space character
  1200. */
  1201. if (c == NULL) {
  1202. if (chunk->remain >= 2 && ucl_lex_is_comment (p[0], p[1])) {
  1203. if (!ucl_skip_comments (parser)) {
  1204. return false;
  1205. }
  1206. p = chunk->pos;
  1207. }
  1208. else if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  1209. ucl_chunk_skipc (chunk, p);
  1210. }
  1211. else if (ucl_test_character (*p, UCL_CHARACTER_KEY_START)) {
  1212. /* The first symbol */
  1213. c = p;
  1214. ucl_chunk_skipc (chunk, p);
  1215. got_content = true;
  1216. }
  1217. else if (*p == '"') {
  1218. /* JSON style key */
  1219. c = p + 1;
  1220. got_quote = true;
  1221. got_content = true;
  1222. ucl_chunk_skipc (chunk, p);
  1223. }
  1224. else if (*p == '}') {
  1225. /* We have actually end of an object */
  1226. *end_of_object = true;
  1227. return true;
  1228. }
  1229. else if (*p == '.') {
  1230. ucl_chunk_skipc (chunk, p);
  1231. parser->prev_state = parser->state;
  1232. parser->state = UCL_STATE_MACRO_NAME;
  1233. return true;
  1234. }
  1235. else {
  1236. /* Invalid identifier */
  1237. ucl_set_err (parser, UCL_ESYNTAX, "key must begin with a letter",
  1238. &parser->err);
  1239. return false;
  1240. }
  1241. }
  1242. else {
  1243. /* Parse the body of a key */
  1244. if (!got_quote) {
  1245. if (ucl_test_character (*p, UCL_CHARACTER_KEY)) {
  1246. got_content = true;
  1247. ucl_chunk_skipc (chunk, p);
  1248. }
  1249. else if (ucl_test_character (*p, UCL_CHARACTER_KEY_SEP)) {
  1250. end = p;
  1251. break;
  1252. }
  1253. else {
  1254. ucl_set_err (parser, UCL_ESYNTAX, "invalid character in a key",
  1255. &parser->err);
  1256. return false;
  1257. }
  1258. }
  1259. else {
  1260. /* We need to parse json like quoted string */
  1261. if (!ucl_lex_json_string (parser, chunk, &need_unescape, &ucl_escape, &var_expand)) {
  1262. return false;
  1263. }
  1264. /* Always escape keys obtained via json */
  1265. end = chunk->pos - 1;
  1266. p = chunk->pos;
  1267. break;
  1268. }
  1269. }
  1270. }
  1271. if (p >= chunk->end && got_content) {
  1272. ucl_set_err (parser, UCL_ESYNTAX, "unfinished key", &parser->err);
  1273. return false;
  1274. }
  1275. else if (!got_content) {
  1276. return true;
  1277. }
  1278. *end_of_object = false;
  1279. /* We are now at the end of the key, need to parse the rest */
  1280. while (p < chunk->end) {
  1281. if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE)) {
  1282. ucl_chunk_skipc (chunk, p);
  1283. }
  1284. else if (*p == '=') {
  1285. if (!got_eq && !got_semicolon) {
  1286. ucl_chunk_skipc (chunk, p);
  1287. got_eq = true;
  1288. }
  1289. else {
  1290. ucl_set_err (parser, UCL_ESYNTAX, "unexpected '=' character",
  1291. &parser->err);
  1292. return false;
  1293. }
  1294. }
  1295. else if (*p == ':') {
  1296. if (!got_eq && !got_semicolon) {
  1297. ucl_chunk_skipc (chunk, p);
  1298. got_semicolon = true;
  1299. }
  1300. else {
  1301. ucl_set_err (parser, UCL_ESYNTAX, "unexpected ':' character",
  1302. &parser->err);
  1303. return false;
  1304. }
  1305. }
  1306. else if (chunk->remain >= 2 && ucl_lex_is_comment (p[0], p[1])) {
  1307. /* Check for comment */
  1308. if (!ucl_skip_comments (parser)) {
  1309. return false;
  1310. }
  1311. p = chunk->pos;
  1312. }
  1313. else {
  1314. /* Start value */
  1315. break;
  1316. }
  1317. }
  1318. if (p >= chunk->end && got_content) {
  1319. ucl_set_err (parser, UCL_ESYNTAX, "unfinished key", &parser->err);
  1320. return false;
  1321. }
  1322. got_sep = got_semicolon || got_eq;
  1323. if (!got_sep) {
  1324. /*
  1325. * Maybe we have more keys nested, so search for termination character.
  1326. * Possible choices:
  1327. * 1) key1 key2 ... keyN [:=] value <- we treat that as error
  1328. * 2) key1 ... keyN {} or [] <- we treat that as nested objects
  1329. * 3) key1 value[;,\n] <- we treat that as linear object
  1330. */
  1331. t = p;
  1332. *next_key = false;
  1333. while (ucl_test_character (*t, UCL_CHARACTER_WHITESPACE)) {
  1334. t ++;
  1335. }
  1336. /* Check first non-space character after a key */
  1337. if (*t != '{' && *t != '[') {
  1338. while (t < chunk->end) {
  1339. if (*t == ',' || *t == ';' || *t == '\n' || *t == '\r') {
  1340. break;
  1341. }
  1342. else if (*t == '{' || *t == '[') {
  1343. *next_key = true;
  1344. break;
  1345. }
  1346. t ++;
  1347. }
  1348. }
  1349. }
  1350. /* Create a new object */
  1351. nobj = ucl_object_new_full (UCL_NULL, parser->chunks->priority);
  1352. if (nobj == NULL) {
  1353. return false;
  1354. }
  1355. keylen = ucl_copy_or_store_ptr (parser, c, &nobj->trash_stack[UCL_TRASH_KEY],
  1356. &key, end - c, need_unescape, parser->flags & UCL_PARSER_KEY_LOWERCASE,
  1357. false, false);
  1358. if (keylen == -1) {
  1359. ucl_object_unref (nobj);
  1360. return false;
  1361. }
  1362. else if (keylen == 0) {
  1363. ucl_set_err (parser, UCL_ESYNTAX, "empty keys are not allowed", &parser->err);
  1364. ucl_object_unref (nobj);
  1365. return false;
  1366. }
  1367. nobj->key = key;
  1368. nobj->keylen = keylen;
  1369. if (!ucl_parser_process_object_element (parser, nobj)) {
  1370. return false;
  1371. }
  1372. if (ucl_escape) {
  1373. nobj->flags |= UCL_OBJECT_NEED_KEY_ESCAPE;
  1374. }
  1375. return true;
  1376. }
  1377. /**
  1378. * Parse a cl string
  1379. * @param parser
  1380. * @param chunk
  1381. * @param var_expand
  1382. * @param need_unescape
  1383. * @return true if a key has been parsed
  1384. */
  1385. static bool
  1386. ucl_parse_string_value (struct ucl_parser *parser,
  1387. struct ucl_chunk *chunk, bool *var_expand, bool *need_unescape)
  1388. {
  1389. const unsigned char *p;
  1390. enum {
  1391. UCL_BRACE_ROUND = 0,
  1392. UCL_BRACE_SQUARE,
  1393. UCL_BRACE_FIGURE
  1394. };
  1395. int braces[3][2] = {{0, 0}, {0, 0}, {0, 0}};
  1396. p = chunk->pos;
  1397. while (p < chunk->end) {
  1398. /* Skip pairs of figure braces */
  1399. if (*p == '{') {
  1400. braces[UCL_BRACE_FIGURE][0] ++;
  1401. }
  1402. else if (*p == '}') {
  1403. braces[UCL_BRACE_FIGURE][1] ++;
  1404. if (braces[UCL_BRACE_FIGURE][1] <= braces[UCL_BRACE_FIGURE][0]) {
  1405. /* This is not a termination symbol, continue */
  1406. ucl_chunk_skipc (chunk, p);
  1407. continue;
  1408. }
  1409. }
  1410. /* Skip pairs of square braces */
  1411. else if (*p == '[') {
  1412. braces[UCL_BRACE_SQUARE][0] ++;
  1413. }
  1414. else if (*p == ']') {
  1415. braces[UCL_BRACE_SQUARE][1] ++;
  1416. if (braces[UCL_BRACE_SQUARE][1] <= braces[UCL_BRACE_SQUARE][0]) {
  1417. /* This is not a termination symbol, continue */
  1418. ucl_chunk_skipc (chunk, p);
  1419. continue;
  1420. }
  1421. }
  1422. else if (*p == '$') {
  1423. *var_expand = true;
  1424. }
  1425. else if (*p == '\\') {
  1426. *need_unescape = true;
  1427. ucl_chunk_skipc (chunk, p);
  1428. if (p < chunk->end) {
  1429. ucl_chunk_skipc (chunk, p);
  1430. }
  1431. continue;
  1432. }
  1433. if (ucl_lex_is_atom_end (*p) || (chunk->remain >= 2 && ucl_lex_is_comment (p[0], p[1]))) {
  1434. break;
  1435. }
  1436. ucl_chunk_skipc (chunk, p);
  1437. }
  1438. return true;
  1439. }
  1440. /**
  1441. * Parse multiline string ending with \n{term}\n
  1442. * @param parser
  1443. * @param chunk
  1444. * @param term
  1445. * @param term_len
  1446. * @param beg
  1447. * @param var_expand
  1448. * @return size of multiline string or 0 in case of error
  1449. */
  1450. static int
  1451. ucl_parse_multiline_string (struct ucl_parser *parser,
  1452. struct ucl_chunk *chunk, const unsigned char *term,
  1453. int term_len, unsigned char const **beg,
  1454. bool *var_expand)
  1455. {
  1456. const unsigned char *p, *c, *tend;
  1457. bool newline = false;
  1458. int len = 0;
  1459. p = chunk->pos;
  1460. c = p;
  1461. while (p < chunk->end) {
  1462. if (newline) {
  1463. if (chunk->end - p < term_len) {
  1464. return 0;
  1465. }
  1466. else if (memcmp (p, term, term_len) == 0) {
  1467. tend = p + term_len;
  1468. if (*tend != '\n' && *tend != ';' && *tend != ',') {
  1469. /* Incomplete terminator */
  1470. ucl_chunk_skipc (chunk, p);
  1471. continue;
  1472. }
  1473. len = p - c;
  1474. chunk->remain -= term_len;
  1475. chunk->pos = p + term_len;
  1476. chunk->column = term_len;
  1477. *beg = c;
  1478. break;
  1479. }
  1480. }
  1481. if (*p == '\n') {
  1482. newline = true;
  1483. }
  1484. else {
  1485. if (*p == '$') {
  1486. *var_expand = true;
  1487. }
  1488. newline = false;
  1489. }
  1490. ucl_chunk_skipc (chunk, p);
  1491. }
  1492. return len;
  1493. }
  1494. static inline ucl_object_t*
  1495. ucl_parser_get_container (struct ucl_parser *parser)
  1496. {
  1497. ucl_object_t *t, *obj = NULL;
  1498. if (parser == NULL || parser->stack == NULL || parser->stack->obj == NULL) {
  1499. return NULL;
  1500. }
  1501. if (parser->stack->obj->type == UCL_ARRAY) {
  1502. /* Object must be allocated */
  1503. obj = ucl_object_new_full (UCL_NULL, parser->chunks->priority);
  1504. t = parser->stack->obj;
  1505. if (!ucl_array_append (t, obj)) {
  1506. ucl_object_unref (obj);
  1507. return NULL;
  1508. }
  1509. parser->cur_obj = obj;
  1510. ucl_attach_comment (parser, obj, false);
  1511. }
  1512. else {
  1513. /* Object has been already allocated */
  1514. obj = parser->cur_obj;
  1515. }
  1516. return obj;
  1517. }
  1518. /**
  1519. * Handle value data
  1520. * @param parser
  1521. * @param chunk
  1522. * @return
  1523. */
  1524. static bool
  1525. ucl_parse_value (struct ucl_parser *parser, struct ucl_chunk *chunk)
  1526. {
  1527. const unsigned char *p, *c;
  1528. ucl_object_t *obj = NULL;
  1529. unsigned int stripped_spaces;
  1530. ssize_t str_len;
  1531. bool need_unescape = false, ucl_escape = false, var_expand = false;
  1532. p = chunk->pos;
  1533. /* Skip any spaces and comments */
  1534. if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE) ||
  1535. (chunk->remain >= 2 && ucl_lex_is_comment (p[0], p[1]))) {
  1536. while (p < chunk->end && ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  1537. ucl_chunk_skipc (chunk, p);
  1538. }
  1539. if (!ucl_skip_comments (parser)) {
  1540. return false;
  1541. }
  1542. p = chunk->pos;
  1543. }
  1544. while (p < chunk->end) {
  1545. c = p;
  1546. switch (*p) {
  1547. case '"':
  1548. ucl_chunk_skipc (chunk, p);
  1549. if (!ucl_lex_json_string (parser, chunk, &need_unescape, &ucl_escape,
  1550. &var_expand)) {
  1551. return false;
  1552. }
  1553. obj = ucl_parser_get_container (parser);
  1554. if (!obj) {
  1555. return false;
  1556. }
  1557. str_len = chunk->pos - c - 2;
  1558. obj->type = UCL_STRING;
  1559. if ((str_len = ucl_copy_or_store_ptr (parser, c + 1,
  1560. &obj->trash_stack[UCL_TRASH_VALUE],
  1561. &obj->value.sv, str_len, need_unescape, false,
  1562. var_expand, false)) == -1) {
  1563. return false;
  1564. }
  1565. obj->len = str_len;
  1566. parser->state = UCL_STATE_AFTER_VALUE;
  1567. return true;
  1568. break;
  1569. case '\'':
  1570. ucl_chunk_skipc (chunk, p);
  1571. if (!ucl_lex_squoted_string (parser, chunk, &need_unescape)) {
  1572. return false;
  1573. }
  1574. obj = ucl_parser_get_container (parser);
  1575. if (!obj) {
  1576. return false;
  1577. }
  1578. str_len = chunk->pos - c - 2;
  1579. obj->type = UCL_STRING;
  1580. obj->flags |= UCL_OBJECT_SQUOTED;
  1581. if ((str_len = ucl_copy_or_store_ptr (parser, c + 1,
  1582. &obj->trash_stack[UCL_TRASH_VALUE],
  1583. &obj->value.sv, str_len, need_unescape, false,
  1584. var_expand, true)) == -1) {
  1585. return false;
  1586. }
  1587. obj->len = str_len;
  1588. parser->state = UCL_STATE_AFTER_VALUE;
  1589. return true;
  1590. break;
  1591. case '{':
  1592. obj = ucl_parser_get_container (parser);
  1593. if (obj == NULL) {
  1594. return false;
  1595. }
  1596. /* We have a new object */
  1597. if (parser->stack) {
  1598. obj = ucl_parser_add_container (obj, parser, false,
  1599. parser->stack->e.params.level, true);
  1600. }
  1601. else {
  1602. return false;
  1603. }
  1604. if (obj == NULL) {
  1605. return false;
  1606. }
  1607. ucl_chunk_skipc (chunk, p);
  1608. return true;
  1609. break;
  1610. case '[':
  1611. obj = ucl_parser_get_container (parser);
  1612. if (obj == NULL) {
  1613. return false;
  1614. }
  1615. /* We have a new array */
  1616. if (parser->stack) {
  1617. obj = ucl_parser_add_container (obj, parser, true,
  1618. parser->stack->e.params.level, true);
  1619. }
  1620. else {
  1621. return false;
  1622. }
  1623. if (obj == NULL) {
  1624. return false;
  1625. }
  1626. ucl_chunk_skipc (chunk, p);
  1627. return true;
  1628. break;
  1629. case ']':
  1630. /* We have the array ending */
  1631. if (parser->stack && parser->stack->obj->type == UCL_ARRAY) {
  1632. parser->state = UCL_STATE_AFTER_VALUE;
  1633. return true;
  1634. }
  1635. else {
  1636. goto parse_string;
  1637. }
  1638. break;
  1639. case '<':
  1640. obj = ucl_parser_get_container (parser);
  1641. /* We have something like multiline value, which must be <<[A-Z]+\n */
  1642. if (chunk->end - p > 3) {
  1643. if (memcmp (p, "<<", 2) == 0) {
  1644. p += 2;
  1645. /* We allow only uppercase characters in multiline definitions */
  1646. while (p < chunk->end && *p >= 'A' && *p <= 'Z') {
  1647. p ++;
  1648. }
  1649. if (*p =='\n') {
  1650. /* Set chunk positions and start multiline parsing */
  1651. chunk->remain -= p - c + 1;
  1652. c += 2;
  1653. chunk->pos = p + 1;
  1654. chunk->column = 0;
  1655. chunk->line ++;
  1656. if ((str_len = ucl_parse_multiline_string (parser, chunk, c,
  1657. p - c, &c, &var_expand)) == 0) {
  1658. ucl_set_err (parser, UCL_ESYNTAX,
  1659. "unterminated multiline value", &parser->err);
  1660. return false;
  1661. }
  1662. obj->type = UCL_STRING;
  1663. obj->flags |= UCL_OBJECT_MULTILINE;
  1664. if ((str_len = ucl_copy_or_store_ptr (parser, c,
  1665. &obj->trash_stack[UCL_TRASH_VALUE],
  1666. &obj->value.sv, str_len - 1, false,
  1667. false, var_expand, false)) == -1) {
  1668. return false;
  1669. }
  1670. obj->len = str_len;
  1671. parser->state = UCL_STATE_AFTER_VALUE;
  1672. return true;
  1673. }
  1674. }
  1675. }
  1676. /* Fallback to ordinary strings */
  1677. /* FALLTHRU */
  1678. default:
  1679. parse_string:
  1680. if (obj == NULL) {
  1681. obj = ucl_parser_get_container (parser);
  1682. }
  1683. /* Parse atom */
  1684. if (ucl_test_character (*p, UCL_CHARACTER_VALUE_DIGIT_START)) {
  1685. if (!ucl_lex_number (parser, chunk, obj)) {
  1686. if (parser->state == UCL_STATE_ERROR) {
  1687. return false;
  1688. }
  1689. }
  1690. else {
  1691. parser->state = UCL_STATE_AFTER_VALUE;
  1692. return true;
  1693. }
  1694. /* Fallback to normal string */
  1695. }
  1696. if (!ucl_parse_string_value (parser, chunk, &var_expand,
  1697. &need_unescape)) {
  1698. return false;
  1699. }
  1700. /* Cut trailing spaces */
  1701. stripped_spaces = 0;
  1702. while (ucl_test_character (*(chunk->pos - 1 - stripped_spaces),
  1703. UCL_CHARACTER_WHITESPACE)) {
  1704. stripped_spaces ++;
  1705. }
  1706. str_len = chunk->pos - c - stripped_spaces;
  1707. if (str_len <= 0) {
  1708. ucl_set_err (parser, UCL_ESYNTAX, "string value must not be empty",
  1709. &parser->err);
  1710. return false;
  1711. }
  1712. else if (str_len == 4 && memcmp (c, "null", 4) == 0) {
  1713. obj->len = 0;
  1714. obj->type = UCL_NULL;
  1715. }
  1716. else if (str_len == 3 && memcmp (c, "nan", 3) == 0) {
  1717. obj->len = 0;
  1718. obj->type = UCL_FLOAT;
  1719. obj->value.dv = NAN;
  1720. }
  1721. else if (str_len == 3 && memcmp (c, "inf", 3) == 0) {
  1722. obj->len = 0;
  1723. obj->type = UCL_FLOAT;
  1724. obj->value.dv = INFINITY;
  1725. }
  1726. else if (!ucl_maybe_parse_boolean (obj, c, str_len)) {
  1727. obj->type = UCL_STRING;
  1728. if ((str_len = ucl_copy_or_store_ptr (parser, c,
  1729. &obj->trash_stack[UCL_TRASH_VALUE],
  1730. &obj->value.sv, str_len, need_unescape,
  1731. false, var_expand, false)) == -1) {
  1732. return false;
  1733. }
  1734. obj->len = str_len;
  1735. }
  1736. parser->state = UCL_STATE_AFTER_VALUE;
  1737. return true;
  1738. break;
  1739. }
  1740. }
  1741. return true;
  1742. }
  1743. /**
  1744. * Handle after value data
  1745. * @param parser
  1746. * @param chunk
  1747. * @return
  1748. */
  1749. static bool
  1750. ucl_parse_after_value (struct ucl_parser *parser, struct ucl_chunk *chunk)
  1751. {
  1752. const unsigned char *p;
  1753. bool got_sep = false;
  1754. struct ucl_stack *st;
  1755. p = chunk->pos;
  1756. while (p < chunk->end) {
  1757. if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE)) {
  1758. /* Skip whitespaces */
  1759. ucl_chunk_skipc (chunk, p);
  1760. }
  1761. else if (chunk->remain >= 2 && ucl_lex_is_comment (p[0], p[1])) {
  1762. /* Skip comment */
  1763. if (!ucl_skip_comments (parser)) {
  1764. return false;
  1765. }
  1766. /* Treat comment as a separator */
  1767. got_sep = true;
  1768. p = chunk->pos;
  1769. }
  1770. else if (ucl_test_character (*p, UCL_CHARACTER_VALUE_END)) {
  1771. if (*p == '}' || *p == ']') {
  1772. if (parser->stack == NULL) {
  1773. ucl_set_err (parser, UCL_ESYNTAX,
  1774. "end of array or object detected without corresponding start",
  1775. &parser->err);
  1776. return false;
  1777. }
  1778. if ((*p == '}' && parser->stack->obj->type == UCL_OBJECT) ||
  1779. (*p == ']' && parser->stack->obj->type == UCL_ARRAY)) {
  1780. /* Pop all nested objects from a stack */
  1781. st = parser->stack;
  1782. if (!(st->e.params.flags & UCL_STACK_HAS_OBRACE)) {
  1783. parser->err_code = UCL_EUNPAIRED;
  1784. ucl_create_err (&parser->err,
  1785. "%s:%d object closed with } is not opened with { at line %d",
  1786. chunk->fname ? chunk->fname : "memory",
  1787. parser->chunks->line, st->e.params.line);
  1788. return false;
  1789. }
  1790. parser->stack = st->next;
  1791. UCL_FREE (sizeof (struct ucl_stack), st);
  1792. if (parser->cur_obj) {
  1793. ucl_attach_comment (parser, parser->cur_obj, true);
  1794. }
  1795. while (parser->stack != NULL) {
  1796. st = parser->stack;
  1797. if (st->next == NULL) {
  1798. break;
  1799. }
  1800. else if (st->next->e.params.level == st->e.params.level) {
  1801. break;
  1802. }
  1803. parser->stack = st->next;
  1804. parser->cur_obj = st->obj;
  1805. UCL_FREE (sizeof (struct ucl_stack), st);
  1806. }
  1807. }
  1808. else {
  1809. ucl_set_err (parser, UCL_ESYNTAX,
  1810. "unexpected terminating symbol detected",
  1811. &parser->err);
  1812. return false;
  1813. }
  1814. if (parser->stack == NULL) {
  1815. /* Ignore everything after a top object */
  1816. return true;
  1817. }
  1818. else {
  1819. ucl_chunk_skipc (chunk, p);
  1820. }
  1821. got_sep = true;
  1822. }
  1823. else {
  1824. /* Got a separator */
  1825. got_sep = true;
  1826. ucl_chunk_skipc (chunk, p);
  1827. }
  1828. }
  1829. else {
  1830. /* Anything else */
  1831. if (!got_sep) {
  1832. ucl_set_err (parser, UCL_ESYNTAX, "delimiter is missing",
  1833. &parser->err);
  1834. return false;
  1835. }
  1836. return true;
  1837. }
  1838. }
  1839. return true;
  1840. }
  1841. static bool
  1842. ucl_skip_macro_as_comment (struct ucl_parser *parser,
  1843. struct ucl_chunk *chunk)
  1844. {
  1845. const unsigned char *p, *c;
  1846. enum {
  1847. macro_skip_start = 0,
  1848. macro_has_symbols,
  1849. macro_has_obrace,
  1850. macro_has_quote,
  1851. macro_has_backslash,
  1852. macro_has_sqbrace,
  1853. macro_save
  1854. } state = macro_skip_start, prev_state = macro_skip_start;
  1855. p = chunk->pos;
  1856. c = chunk->pos;
  1857. while (p < chunk->end) {
  1858. switch (state) {
  1859. case macro_skip_start:
  1860. if (!ucl_test_character (*p, UCL_CHARACTER_WHITESPACE)) {
  1861. state = macro_has_symbols;
  1862. }
  1863. else if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  1864. state = macro_save;
  1865. continue;
  1866. }
  1867. ucl_chunk_skipc (chunk, p);
  1868. break;
  1869. case macro_has_symbols:
  1870. if (*p == '{') {
  1871. state = macro_has_sqbrace;
  1872. }
  1873. else if (*p == '(') {
  1874. state = macro_has_obrace;
  1875. }
  1876. else if (*p == '"') {
  1877. state = macro_has_quote;
  1878. }
  1879. else if (*p == '\n') {
  1880. state = macro_save;
  1881. continue;
  1882. }
  1883. ucl_chunk_skipc (chunk, p);
  1884. break;
  1885. case macro_has_obrace:
  1886. if (*p == '\\') {
  1887. prev_state = state;
  1888. state = macro_has_backslash;
  1889. }
  1890. else if (*p == ')') {
  1891. state = macro_has_symbols;
  1892. }
  1893. ucl_chunk_skipc (chunk, p);
  1894. break;
  1895. case macro_has_sqbrace:
  1896. if (*p == '\\') {
  1897. prev_state = state;
  1898. state = macro_has_backslash;
  1899. }
  1900. else if (*p == '}') {
  1901. state = macro_save;
  1902. }
  1903. ucl_chunk_skipc (chunk, p);
  1904. break;
  1905. case macro_has_quote:
  1906. if (*p == '\\') {
  1907. prev_state = state;
  1908. state = macro_has_backslash;
  1909. }
  1910. else if (*p == '"') {
  1911. state = macro_save;
  1912. }
  1913. ucl_chunk_skipc (chunk, p);
  1914. break;
  1915. case macro_has_backslash:
  1916. state = prev_state;
  1917. ucl_chunk_skipc (chunk, p);
  1918. break;
  1919. case macro_save:
  1920. if (parser->flags & UCL_PARSER_SAVE_COMMENTS) {
  1921. ucl_save_comment (parser, c, p - c);
  1922. }
  1923. return true;
  1924. }
  1925. }
  1926. return false;
  1927. }
  1928. /**
  1929. * Handle macro data
  1930. * @param parser
  1931. * @param chunk
  1932. * @param marco
  1933. * @param macro_start
  1934. * @param macro_len
  1935. * @return
  1936. */
  1937. static bool
  1938. ucl_parse_macro_value (struct ucl_parser *parser,
  1939. struct ucl_chunk *chunk, struct ucl_macro *macro,
  1940. unsigned char const **macro_start, size_t *macro_len)
  1941. {
  1942. const unsigned char *p, *c;
  1943. bool need_unescape = false, ucl_escape = false, var_expand = false;
  1944. p = chunk->pos;
  1945. switch (*p) {
  1946. case '"':
  1947. /* We have macro value encoded in quotes */
  1948. c = p;
  1949. ucl_chunk_skipc (chunk, p);
  1950. if (!ucl_lex_json_string (parser, chunk, &need_unescape, &ucl_escape, &var_expand)) {
  1951. return false;
  1952. }
  1953. *macro_start = c + 1;
  1954. *macro_len = chunk->pos - c - 2;
  1955. p = chunk->pos;
  1956. break;
  1957. case '{':
  1958. /* We got a multiline macro body */
  1959. ucl_chunk_skipc (chunk, p);
  1960. /* Skip spaces at the beginning */
  1961. while (p < chunk->end) {
  1962. if (ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  1963. ucl_chunk_skipc (chunk, p);
  1964. }
  1965. else {
  1966. break;
  1967. }
  1968. }
  1969. c = p;
  1970. while (p < chunk->end) {
  1971. if (*p == '}') {
  1972. break;
  1973. }
  1974. ucl_chunk_skipc (chunk, p);
  1975. }
  1976. *macro_start = c;
  1977. *macro_len = p - c;
  1978. ucl_chunk_skipc (chunk, p);
  1979. break;
  1980. default:
  1981. /* Macro is not enclosed in quotes or braces */
  1982. c = p;
  1983. while (p < chunk->end) {
  1984. if (ucl_lex_is_atom_end (*p)) {
  1985. break;
  1986. }
  1987. ucl_chunk_skipc (chunk, p);
  1988. }
  1989. *macro_start = c;
  1990. *macro_len = p - c;
  1991. break;
  1992. }
  1993. /* We are at the end of a macro */
  1994. /* Skip ';' and space characters and return to previous state */
  1995. while (p < chunk->end) {
  1996. if (!ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE) && *p != ';') {
  1997. break;
  1998. }
  1999. ucl_chunk_skipc (chunk, p);
  2000. }
  2001. return true;
  2002. }
  2003. /**
  2004. * Parse macro arguments as UCL object
  2005. * @param parser parser structure
  2006. * @param chunk the current data chunk
  2007. * @return
  2008. */
  2009. static ucl_object_t *
  2010. ucl_parse_macro_arguments (struct ucl_parser *parser,
  2011. struct ucl_chunk *chunk)
  2012. {
  2013. ucl_object_t *res = NULL;
  2014. struct ucl_parser *params_parser;
  2015. int obraces = 1, ebraces = 0, state = 0;
  2016. const unsigned char *p, *c;
  2017. size_t args_len = 0;
  2018. struct ucl_parser_saved_state saved;
  2019. saved.column = chunk->column;
  2020. saved.line = chunk->line;
  2021. saved.pos = chunk->pos;
  2022. saved.remain = chunk->remain;
  2023. p = chunk->pos;
  2024. if (*p != '(' || chunk->remain < 2) {
  2025. return NULL;
  2026. }
  2027. /* Set begin and start */
  2028. ucl_chunk_skipc (chunk, p);
  2029. c = p;
  2030. while ((p) < (chunk)->end) {
  2031. switch (state) {
  2032. case 0:
  2033. /* Parse symbols and check for '(', ')' and '"' */
  2034. if (*p == '(') {
  2035. obraces ++;
  2036. }
  2037. else if (*p == ')') {
  2038. ebraces ++;
  2039. }
  2040. else if (*p == '"') {
  2041. state = 1;
  2042. }
  2043. /* Check pairing */
  2044. if (obraces == ebraces) {
  2045. state = 99;
  2046. }
  2047. else {
  2048. args_len ++;
  2049. }
  2050. /* Check overflow */
  2051. if (chunk->remain == 0) {
  2052. goto restore_chunk;
  2053. }
  2054. ucl_chunk_skipc (chunk, p);
  2055. break;
  2056. case 1:
  2057. /* We have quote character, so skip all but quotes */
  2058. if (*p == '"' && *(p - 1) != '\\') {
  2059. state = 0;
  2060. }
  2061. if (chunk->remain == 0) {
  2062. goto restore_chunk;
  2063. }
  2064. args_len ++;
  2065. ucl_chunk_skipc (chunk, p);
  2066. break;
  2067. case 99:
  2068. /*
  2069. * We have read the full body of arguments, so we need to parse and set
  2070. * object from that
  2071. */
  2072. params_parser = ucl_parser_new (parser->flags);
  2073. if (!ucl_parser_add_chunk (params_parser, c, args_len)) {
  2074. ucl_set_err (parser, UCL_ESYNTAX, "macro arguments parsing error",
  2075. &parser->err);
  2076. }
  2077. else {
  2078. res = ucl_parser_get_object (params_parser);
  2079. }
  2080. ucl_parser_free (params_parser);
  2081. return res;
  2082. break;
  2083. }
  2084. }
  2085. return res;
  2086. restore_chunk:
  2087. chunk->column = saved.column;
  2088. chunk->line = saved.line;
  2089. chunk->pos = saved.pos;
  2090. chunk->remain = saved.remain;
  2091. return NULL;
  2092. }
  2093. #define SKIP_SPACES_COMMENTS(parser, chunk, p) do { \
  2094. while ((p) < (chunk)->end) { \
  2095. if (!ucl_test_character (*(p), UCL_CHARACTER_WHITESPACE_UNSAFE)) { \
  2096. if ((chunk)->remain >= 2 && ucl_lex_is_comment ((p)[0], (p)[1])) { \
  2097. if (!ucl_skip_comments (parser)) { \
  2098. return false; \
  2099. } \
  2100. p = (chunk)->pos; \
  2101. } \
  2102. break; \
  2103. } \
  2104. ucl_chunk_skipc (chunk, p); \
  2105. } \
  2106. } while(0)
  2107. /**
  2108. * Handle the main states of rcl parser
  2109. * @param parser parser structure
  2110. * @return true if chunk has been parsed and false in case of error
  2111. */
  2112. static bool
  2113. ucl_state_machine (struct ucl_parser *parser)
  2114. {
  2115. ucl_object_t *obj, *macro_args;
  2116. struct ucl_chunk *chunk = parser->chunks;
  2117. const unsigned char *p, *c = NULL, *macro_start = NULL;
  2118. unsigned char *macro_escaped;
  2119. size_t macro_len = 0;
  2120. struct ucl_macro *macro = NULL;
  2121. bool next_key = false, end_of_object = false, ret;
  2122. if (parser->top_obj == NULL) {
  2123. parser->state = UCL_STATE_INIT;
  2124. }
  2125. p = chunk->pos;
  2126. while (chunk->pos < chunk->end) {
  2127. switch (parser->state) {
  2128. case UCL_STATE_INIT:
  2129. /*
  2130. * At the init state we can either go to the parse array or object
  2131. * if we got [ or { correspondingly or can just treat new data as
  2132. * a key of newly created object
  2133. */
  2134. if (!ucl_skip_comments (parser)) {
  2135. parser->prev_state = parser->state;
  2136. parser->state = UCL_STATE_ERROR;
  2137. return false;
  2138. }
  2139. else {
  2140. bool seen_obrace = false;
  2141. /* Skip any spaces */
  2142. while (p < chunk->end && ucl_test_character (*p,
  2143. UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  2144. ucl_chunk_skipc (chunk, p);
  2145. }
  2146. p = chunk->pos;
  2147. if (p < chunk->end) {
  2148. if (*p == '[') {
  2149. parser->state = UCL_STATE_VALUE;
  2150. ucl_chunk_skipc (chunk, p);
  2151. seen_obrace = true;
  2152. }
  2153. else {
  2154. if (*p == '{') {
  2155. ucl_chunk_skipc (chunk, p);
  2156. parser->state = UCL_STATE_KEY_OBRACE;
  2157. seen_obrace = true;
  2158. }
  2159. else {
  2160. parser->state = UCL_STATE_KEY;
  2161. }
  2162. }
  2163. }
  2164. if (parser->top_obj == NULL) {
  2165. if (parser->state == UCL_STATE_VALUE) {
  2166. obj = ucl_parser_add_container (NULL, parser, true, 0,
  2167. seen_obrace);
  2168. }
  2169. else {
  2170. obj = ucl_parser_add_container (NULL, parser, false, 0,
  2171. seen_obrace);
  2172. }
  2173. if (obj == NULL) {
  2174. return false;
  2175. }
  2176. parser->top_obj = obj;
  2177. parser->cur_obj = obj;
  2178. }
  2179. }
  2180. break;
  2181. case UCL_STATE_KEY:
  2182. case UCL_STATE_KEY_OBRACE:
  2183. /* Skip any spaces */
  2184. while (p < chunk->end && ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE)) {
  2185. ucl_chunk_skipc (chunk, p);
  2186. }
  2187. if (p == chunk->end || *p == '}') {
  2188. /* We have the end of an object */
  2189. parser->state = UCL_STATE_AFTER_VALUE;
  2190. continue;
  2191. }
  2192. if (parser->stack == NULL) {
  2193. /* No objects are on stack, but we want to parse a key */
  2194. ucl_set_err (parser, UCL_ESYNTAX, "top object is finished but the parser "
  2195. "expects a key", &parser->err);
  2196. parser->prev_state = parser->state;
  2197. parser->state = UCL_STATE_ERROR;
  2198. return false;
  2199. }
  2200. if (!ucl_parse_key (parser, chunk, &next_key, &end_of_object)) {
  2201. parser->prev_state = parser->state;
  2202. parser->state = UCL_STATE_ERROR;
  2203. return false;
  2204. }
  2205. if (end_of_object) {
  2206. p = chunk->pos;
  2207. parser->state = UCL_STATE_AFTER_VALUE;
  2208. continue;
  2209. }
  2210. else if (parser->state != UCL_STATE_MACRO_NAME) {
  2211. if (next_key && parser->stack->obj->type == UCL_OBJECT) {
  2212. /* Parse more keys and nest objects accordingly */
  2213. obj = ucl_parser_add_container (parser->cur_obj,
  2214. parser,
  2215. false,
  2216. parser->stack->e.params.level + 1,
  2217. parser->state == UCL_STATE_KEY_OBRACE);
  2218. if (obj == NULL) {
  2219. return false;
  2220. }
  2221. }
  2222. else {
  2223. parser->state = UCL_STATE_VALUE;
  2224. }
  2225. }
  2226. else {
  2227. c = chunk->pos;
  2228. }
  2229. p = chunk->pos;
  2230. break;
  2231. case UCL_STATE_VALUE:
  2232. /* We need to check what we do have */
  2233. if (!parser->cur_obj || !ucl_parse_value (parser, chunk)) {
  2234. parser->prev_state = parser->state;
  2235. parser->state = UCL_STATE_ERROR;
  2236. return false;
  2237. }
  2238. /* State is set in ucl_parse_value call */
  2239. p = chunk->pos;
  2240. break;
  2241. case UCL_STATE_AFTER_VALUE:
  2242. if (!ucl_parse_after_value (parser, chunk)) {
  2243. parser->prev_state = parser->state;
  2244. parser->state = UCL_STATE_ERROR;
  2245. return false;
  2246. }
  2247. if (parser->stack != NULL) {
  2248. if (parser->stack->obj->type == UCL_OBJECT) {
  2249. parser->state = UCL_STATE_KEY;
  2250. }
  2251. else {
  2252. /* Array */
  2253. parser->state = UCL_STATE_VALUE;
  2254. }
  2255. }
  2256. else {
  2257. /* Skip everything at the end */
  2258. return true;
  2259. }
  2260. p = chunk->pos;
  2261. break;
  2262. case UCL_STATE_MACRO_NAME:
  2263. if (parser->flags & UCL_PARSER_DISABLE_MACRO) {
  2264. if (!ucl_skip_macro_as_comment (parser, chunk)) {
  2265. /* We have invalid macro */
  2266. ucl_create_err (&parser->err,
  2267. "error at %s:%d at column %d: invalid macro",
  2268. chunk->fname ? chunk->fname : "memory",
  2269. chunk->line,
  2270. chunk->column);
  2271. parser->state = UCL_STATE_ERROR;
  2272. return false;
  2273. }
  2274. else {
  2275. p = chunk->pos;
  2276. parser->state = parser->prev_state;
  2277. }
  2278. }
  2279. else {
  2280. if (!ucl_test_character (*p, UCL_CHARACTER_WHITESPACE_UNSAFE) &&
  2281. *p != '(') {
  2282. ucl_chunk_skipc (chunk, p);
  2283. }
  2284. else {
  2285. if (c != NULL && p - c > 0) {
  2286. /* We got macro name */
  2287. macro_len = (size_t) (p - c);
  2288. HASH_FIND (hh, parser->macroes, c, macro_len, macro);
  2289. if (macro == NULL) {
  2290. ucl_create_err (&parser->err,
  2291. "error at %s:%d at column %d: "
  2292. "unknown macro: '%.*s', character: '%c'",
  2293. chunk->fname ? chunk->fname : "memory",
  2294. chunk->line,
  2295. chunk->column,
  2296. (int) (p - c),
  2297. c,
  2298. *chunk->pos);
  2299. parser->state = UCL_STATE_ERROR;
  2300. return false;
  2301. }
  2302. /* Now we need to skip all spaces */
  2303. SKIP_SPACES_COMMENTS(parser, chunk, p);
  2304. parser->state = UCL_STATE_MACRO;
  2305. }
  2306. else {
  2307. /* We have invalid macro name */
  2308. ucl_create_err (&parser->err,
  2309. "error at %s:%d at column %d: invalid macro name",
  2310. chunk->fname ? chunk->fname : "memory",
  2311. chunk->line,
  2312. chunk->column);
  2313. parser->state = UCL_STATE_ERROR;
  2314. return false;
  2315. }
  2316. }
  2317. }
  2318. break;
  2319. case UCL_STATE_MACRO:
  2320. if (*chunk->pos == '(') {
  2321. macro_args = ucl_parse_macro_arguments (parser, chunk);
  2322. p = chunk->pos;
  2323. if (macro_args) {
  2324. SKIP_SPACES_COMMENTS(parser, chunk, p);
  2325. }
  2326. }
  2327. else {
  2328. macro_args = NULL;
  2329. }
  2330. if (!ucl_parse_macro_value (parser, chunk, macro,
  2331. &macro_start, &macro_len)) {
  2332. parser->prev_state = parser->state;
  2333. parser->state = UCL_STATE_ERROR;
  2334. return false;
  2335. }
  2336. macro_len = ucl_expand_variable (parser, &macro_escaped,
  2337. macro_start, macro_len);
  2338. parser->state = parser->prev_state;
  2339. if (macro_escaped == NULL && macro != NULL) {
  2340. if (macro->is_context) {
  2341. ret = macro->h.context_handler (macro_start, macro_len,
  2342. macro_args,
  2343. parser->top_obj,
  2344. macro->ud);
  2345. }
  2346. else {
  2347. ret = macro->h.handler (macro_start, macro_len, macro_args,
  2348. macro->ud);
  2349. }
  2350. }
  2351. else if (macro != NULL) {
  2352. if (macro->is_context) {
  2353. ret = macro->h.context_handler (macro_escaped, macro_len,
  2354. macro_args,
  2355. parser->top_obj,
  2356. macro->ud);
  2357. }
  2358. else {
  2359. ret = macro->h.handler (macro_escaped, macro_len, macro_args,
  2360. macro->ud);
  2361. }
  2362. UCL_FREE (macro_len + 1, macro_escaped);
  2363. }
  2364. else {
  2365. ret = false;
  2366. ucl_set_err (parser, UCL_EINTERNAL,
  2367. "internal error: parser has macro undefined", &parser->err);
  2368. }
  2369. /*
  2370. * Chunk can be modified within macro handler
  2371. */
  2372. chunk = parser->chunks;
  2373. p = chunk->pos;
  2374. if (macro_args) {
  2375. ucl_object_unref (macro_args);
  2376. }
  2377. if (!ret) {
  2378. return false;
  2379. }
  2380. break;
  2381. default:
  2382. ucl_set_err (parser, UCL_EINTERNAL,
  2383. "internal error: parser is in an unknown state", &parser->err);
  2384. parser->state = UCL_STATE_ERROR;
  2385. return false;
  2386. }
  2387. }
  2388. if (parser->last_comment) {
  2389. if (parser->cur_obj) {
  2390. ucl_attach_comment (parser, parser->cur_obj, true);
  2391. }
  2392. else if (parser->stack && parser->stack->obj) {
  2393. ucl_attach_comment (parser, parser->stack->obj, true);
  2394. }
  2395. else if (parser->top_obj) {
  2396. ucl_attach_comment (parser, parser->top_obj, true);
  2397. }
  2398. else {
  2399. ucl_object_unref (parser->last_comment);
  2400. }
  2401. }
  2402. if (parser->stack != NULL && parser->state != UCL_STATE_ERROR) {
  2403. struct ucl_stack *st;
  2404. bool has_error = false;
  2405. LL_FOREACH (parser->stack, st) {
  2406. if (st->chunk != parser->chunks) {
  2407. break; /* Not our chunk, give up */
  2408. }
  2409. if (st->e.params.flags & UCL_STACK_HAS_OBRACE) {
  2410. if (parser->err == NULL) {
  2411. utstring_new (parser->err);
  2412. }
  2413. utstring_printf (parser->err, "%s:%d unmatched open brace at %d; ",
  2414. chunk->fname ? chunk->fname : "memory",
  2415. parser->chunks->line,
  2416. st->e.params.line);
  2417. has_error = true;
  2418. }
  2419. }
  2420. if (has_error) {
  2421. parser->err_code = UCL_EUNPAIRED;
  2422. return false;
  2423. }
  2424. }
  2425. return true;
  2426. }
  2427. #define UPRM_SAFE(fn, a, b, c, el) do { \
  2428. if (!fn(a, b, c, a)) \
  2429. goto el; \
  2430. } while (0)
  2431. struct ucl_parser*
  2432. ucl_parser_new (int flags)
  2433. {
  2434. struct ucl_parser *parser;
  2435. parser = UCL_ALLOC (sizeof (struct ucl_parser));
  2436. if (parser == NULL) {
  2437. return NULL;
  2438. }
  2439. memset (parser, 0, sizeof (struct ucl_parser));
  2440. UPRM_SAFE(ucl_parser_register_macro, parser, "include", ucl_include_handler, e0);
  2441. UPRM_SAFE(ucl_parser_register_macro, parser, "try_include", ucl_try_include_handler, e0);
  2442. UPRM_SAFE(ucl_parser_register_macro, parser, "includes", ucl_includes_handler, e0);
  2443. UPRM_SAFE(ucl_parser_register_macro, parser, "priority", ucl_priority_handler, e0);
  2444. UPRM_SAFE(ucl_parser_register_macro, parser, "load", ucl_load_handler, e0);
  2445. UPRM_SAFE(ucl_parser_register_context_macro, parser, "inherit", ucl_inherit_handler, e0);
  2446. parser->flags = flags;
  2447. parser->includepaths = NULL;
  2448. if (flags & UCL_PARSER_SAVE_COMMENTS) {
  2449. parser->comments = ucl_object_typed_new (UCL_OBJECT);
  2450. }
  2451. if (!(flags & UCL_PARSER_NO_FILEVARS)) {
  2452. /* Initial assumption about filevars */
  2453. ucl_parser_set_filevars (parser, NULL, false);
  2454. }
  2455. return parser;
  2456. e0:
  2457. ucl_parser_free(parser);
  2458. return NULL;
  2459. }
  2460. bool
  2461. ucl_parser_set_default_priority (struct ucl_parser *parser, unsigned prio)
  2462. {
  2463. if (parser == NULL) {
  2464. return false;
  2465. }
  2466. parser->default_priority = prio;
  2467. return true;
  2468. }
  2469. int
  2470. ucl_parser_get_default_priority (struct ucl_parser *parser)
  2471. {
  2472. if (parser == NULL) {
  2473. return -1;
  2474. }
  2475. return parser->default_priority;
  2476. }
  2477. bool
  2478. ucl_parser_register_macro (struct ucl_parser *parser, const char *macro,
  2479. ucl_macro_handler handler, void* ud)
  2480. {
  2481. struct ucl_macro *new;
  2482. if (macro == NULL || handler == NULL) {
  2483. return false;
  2484. }
  2485. new = UCL_ALLOC (sizeof (struct ucl_macro));
  2486. if (new == NULL) {
  2487. return false;
  2488. }
  2489. memset (new, 0, sizeof (struct ucl_macro));
  2490. new->h.handler = handler;
  2491. new->name = strdup (macro);
  2492. if (new->name == NULL) {
  2493. UCL_FREE (sizeof (struct ucl_macro), new);
  2494. return false;
  2495. }
  2496. new->ud = ud;
  2497. HASH_ADD_KEYPTR (hh, parser->macroes, new->name, strlen (new->name), new);
  2498. return true;
  2499. }
  2500. bool
  2501. ucl_parser_register_context_macro (struct ucl_parser *parser, const char *macro,
  2502. ucl_context_macro_handler handler, void* ud)
  2503. {
  2504. struct ucl_macro *new;
  2505. if (macro == NULL || handler == NULL) {
  2506. return false;
  2507. }
  2508. new = UCL_ALLOC (sizeof (struct ucl_macro));
  2509. if (new == NULL) {
  2510. return false;
  2511. }
  2512. memset (new, 0, sizeof (struct ucl_macro));
  2513. new->h.context_handler = handler;
  2514. new->name = strdup (macro);
  2515. if (new->name == NULL) {
  2516. UCL_FREE (sizeof (struct ucl_macro), new);
  2517. return false;
  2518. }
  2519. new->ud = ud;
  2520. new->is_context = true;
  2521. HASH_ADD_KEYPTR (hh, parser->macroes, new->name, strlen (new->name), new);
  2522. return true;
  2523. }
  2524. void
  2525. ucl_parser_register_variable (struct ucl_parser *parser, const char *var,
  2526. const char *value)
  2527. {
  2528. struct ucl_variable *new = NULL, *cur;
  2529. if (var == NULL) {
  2530. return;
  2531. }
  2532. /* Find whether a variable already exists */
  2533. LL_FOREACH (parser->variables, cur) {
  2534. if (strcmp (cur->var, var) == 0) {
  2535. new = cur;
  2536. break;
  2537. }
  2538. }
  2539. if (value == NULL) {
  2540. if (new != NULL) {
  2541. /* Remove variable */
  2542. DL_DELETE (parser->variables, new);
  2543. free (new->var);
  2544. free (new->value);
  2545. UCL_FREE (sizeof (struct ucl_variable), new);
  2546. }
  2547. else {
  2548. /* Do nothing */
  2549. return;
  2550. }
  2551. }
  2552. else {
  2553. if (new == NULL) {
  2554. new = UCL_ALLOC (sizeof (struct ucl_variable));
  2555. if (new == NULL) {
  2556. return;
  2557. }
  2558. memset (new, 0, sizeof (struct ucl_variable));
  2559. new->var = strdup (var);
  2560. new->var_len = strlen (var);
  2561. new->value = strdup (value);
  2562. new->value_len = strlen (value);
  2563. DL_APPEND (parser->variables, new);
  2564. }
  2565. else {
  2566. free (new->value);
  2567. new->value = strdup (value);
  2568. new->value_len = strlen (value);
  2569. }
  2570. }
  2571. }
  2572. void
  2573. ucl_parser_set_variables_handler (struct ucl_parser *parser,
  2574. ucl_variable_handler handler, void *ud)
  2575. {
  2576. parser->var_handler = handler;
  2577. parser->var_data = ud;
  2578. }
  2579. bool
  2580. ucl_parser_add_chunk_full (struct ucl_parser *parser, const unsigned char *data,
  2581. size_t len, unsigned priority, enum ucl_duplicate_strategy strat,
  2582. enum ucl_parse_type parse_type)
  2583. {
  2584. struct ucl_chunk *chunk;
  2585. struct ucl_parser_special_handler *special_handler;
  2586. if (parser == NULL) {
  2587. return false;
  2588. }
  2589. if (data == NULL && len != 0) {
  2590. ucl_create_err (&parser->err, "invalid chunk added");
  2591. return false;
  2592. }
  2593. if (parser->state != UCL_STATE_ERROR) {
  2594. chunk = UCL_ALLOC (sizeof (struct ucl_chunk));
  2595. if (chunk == NULL) {
  2596. ucl_create_err (&parser->err, "cannot allocate chunk structure");
  2597. return false;
  2598. }
  2599. memset (chunk, 0, sizeof (*chunk));
  2600. /* Apply all matching handlers from the first to the last */
  2601. LL_FOREACH (parser->special_handlers, special_handler) {
  2602. if ((special_handler->flags & UCL_SPECIAL_HANDLER_PREPROCESS_ALL) ||
  2603. (len >= special_handler->magic_len &&
  2604. memcmp (data, special_handler->magic, special_handler->magic_len) == 0)) {
  2605. unsigned char *ndata = NULL;
  2606. size_t nlen = 0;
  2607. if (!special_handler->handler (parser, data, len, &ndata, &nlen,
  2608. special_handler->user_data)) {
  2609. ucl_create_err (&parser->err, "call for external handler failed");
  2610. return false;
  2611. }
  2612. struct ucl_parser_special_handler_chain *nchain;
  2613. nchain = UCL_ALLOC (sizeof (*nchain));
  2614. nchain->begin = ndata;
  2615. nchain->len = nlen;
  2616. nchain->special_handler = special_handler;
  2617. /* Free order is reversed */
  2618. LL_PREPEND (chunk->special_handlers, nchain);
  2619. data = ndata;
  2620. len = nlen;
  2621. }
  2622. }
  2623. if (parse_type == UCL_PARSE_AUTO && len > 0) {
  2624. /* We need to detect parse type by the first symbol */
  2625. if ((*data & 0x80) == 0x80 && (*data >= 0xdc && *data <= 0xdf)) {
  2626. parse_type = UCL_PARSE_MSGPACK;
  2627. }
  2628. else if (*data == '(') {
  2629. parse_type = UCL_PARSE_CSEXP;
  2630. }
  2631. else {
  2632. parse_type = UCL_PARSE_UCL;
  2633. }
  2634. }
  2635. chunk->begin = data;
  2636. chunk->remain = len;
  2637. chunk->pos = chunk->begin;
  2638. chunk->end = chunk->begin + len;
  2639. chunk->line = 1;
  2640. chunk->column = 0;
  2641. chunk->priority = priority;
  2642. chunk->strategy = strat;
  2643. chunk->parse_type = parse_type;
  2644. if (parser->cur_file) {
  2645. chunk->fname = strdup (parser->cur_file);
  2646. }
  2647. LL_PREPEND (parser->chunks, chunk);
  2648. parser->recursion ++;
  2649. if (parser->recursion > UCL_MAX_RECURSION) {
  2650. ucl_create_err (&parser->err, "maximum include nesting limit is reached: %d",
  2651. parser->recursion);
  2652. return false;
  2653. }
  2654. if (len > 0) {
  2655. /* Need to parse something */
  2656. switch (parse_type) {
  2657. default:
  2658. case UCL_PARSE_UCL:
  2659. return ucl_state_machine (parser);
  2660. case UCL_PARSE_MSGPACK:
  2661. return ucl_parse_msgpack (parser);
  2662. case UCL_PARSE_CSEXP:
  2663. return ucl_parse_csexp (parser);
  2664. }
  2665. }
  2666. else {
  2667. /* Just add empty chunk and go forward */
  2668. if (parser->top_obj == NULL) {
  2669. /*
  2670. * In case of empty object, create one to indicate that we've
  2671. * read something
  2672. */
  2673. parser->top_obj = ucl_object_new_full (UCL_OBJECT, priority);
  2674. }
  2675. return true;
  2676. }
  2677. }
  2678. ucl_create_err (&parser->err, "a parser is in an invalid state");
  2679. return false;
  2680. }
  2681. bool
  2682. ucl_parser_add_chunk_priority (struct ucl_parser *parser,
  2683. const unsigned char *data, size_t len, unsigned priority)
  2684. {
  2685. /* We dereference parser, so this check is essential */
  2686. if (parser == NULL) {
  2687. return false;
  2688. }
  2689. return ucl_parser_add_chunk_full (parser, data, len,
  2690. priority, UCL_DUPLICATE_APPEND, UCL_PARSE_UCL);
  2691. }
  2692. bool
  2693. ucl_parser_add_chunk (struct ucl_parser *parser, const unsigned char *data,
  2694. size_t len)
  2695. {
  2696. if (parser == NULL) {
  2697. return false;
  2698. }
  2699. return ucl_parser_add_chunk_full (parser, data, len,
  2700. parser->default_priority, UCL_DUPLICATE_APPEND, UCL_PARSE_UCL);
  2701. }
  2702. bool
  2703. ucl_parser_insert_chunk (struct ucl_parser *parser, const unsigned char *data,
  2704. size_t len)
  2705. {
  2706. if (parser == NULL || parser->top_obj == NULL) {
  2707. return false;
  2708. }
  2709. bool res;
  2710. struct ucl_chunk *chunk;
  2711. int state = parser->state;
  2712. parser->state = UCL_STATE_INIT;
  2713. /* Prevent inserted chunks from unintentionally closing the current object */
  2714. if (parser->stack != NULL && parser->stack->next != NULL) {
  2715. parser->stack->e.params.level = parser->stack->next->e.params.level;
  2716. }
  2717. res = ucl_parser_add_chunk_full (parser, data, len, parser->chunks->priority,
  2718. parser->chunks->strategy, parser->chunks->parse_type);
  2719. /* Remove chunk from the stack */
  2720. chunk = parser->chunks;
  2721. if (chunk != NULL) {
  2722. parser->chunks = chunk->next;
  2723. ucl_chunk_free (chunk);
  2724. parser->recursion --;
  2725. }
  2726. parser->state = state;
  2727. return res;
  2728. }
  2729. bool
  2730. ucl_parser_add_string_priority (struct ucl_parser *parser, const char *data,
  2731. size_t len, unsigned priority)
  2732. {
  2733. if (data == NULL) {
  2734. ucl_create_err (&parser->err, "invalid string added");
  2735. return false;
  2736. }
  2737. if (len == 0) {
  2738. len = strlen (data);
  2739. }
  2740. return ucl_parser_add_chunk_priority (parser,
  2741. (const unsigned char *)data, len, priority);
  2742. }
  2743. bool
  2744. ucl_parser_add_string (struct ucl_parser *parser, const char *data,
  2745. size_t len)
  2746. {
  2747. if (parser == NULL) {
  2748. return false;
  2749. }
  2750. return ucl_parser_add_string_priority (parser,
  2751. (const unsigned char *)data, len, parser->default_priority);
  2752. }
  2753. bool
  2754. ucl_set_include_path (struct ucl_parser *parser, ucl_object_t *paths)
  2755. {
  2756. if (parser == NULL || paths == NULL) {
  2757. return false;
  2758. }
  2759. if (parser->includepaths == NULL) {
  2760. parser->includepaths = ucl_object_copy (paths);
  2761. }
  2762. else {
  2763. ucl_object_unref (parser->includepaths);
  2764. parser->includepaths = ucl_object_copy (paths);
  2765. }
  2766. if (parser->includepaths == NULL) {
  2767. return false;
  2768. }
  2769. return true;
  2770. }
  2771. unsigned char ucl_parser_chunk_peek (struct ucl_parser *parser)
  2772. {
  2773. if (parser == NULL || parser->chunks == NULL || parser->chunks->pos == NULL || parser->chunks->end == NULL ||
  2774. parser->chunks->pos == parser->chunks->end) {
  2775. return 0;
  2776. }
  2777. return( *parser->chunks->pos );
  2778. }
  2779. bool ucl_parser_chunk_skip (struct ucl_parser *parser)
  2780. {
  2781. if (parser == NULL || parser->chunks == NULL || parser->chunks->pos == NULL || parser->chunks->end == NULL ||
  2782. parser->chunks->pos == parser->chunks->end) {
  2783. return false;
  2784. }
  2785. const unsigned char *p = parser->chunks->pos;
  2786. ucl_chunk_skipc( parser->chunks, p );
  2787. if( parser->chunks->pos != NULL ) return true;
  2788. return false;
  2789. }
  2790. ucl_object_t*
  2791. ucl_parser_get_current_stack_object (struct ucl_parser *parser, unsigned int depth)
  2792. {
  2793. ucl_object_t *obj;
  2794. if (parser == NULL || parser->stack == NULL) {
  2795. return NULL;
  2796. }
  2797. struct ucl_stack *stack = parser->stack;
  2798. if(stack == NULL || stack->obj == NULL || ucl_object_type (stack->obj) != UCL_OBJECT)
  2799. {
  2800. return NULL;
  2801. }
  2802. for( unsigned int i = 0; i < depth; ++i )
  2803. {
  2804. stack = stack->next;
  2805. if(stack == NULL || stack->obj == NULL || ucl_object_type (stack->obj) != UCL_OBJECT)
  2806. {
  2807. return NULL;
  2808. }
  2809. }
  2810. obj = ucl_object_ref (stack->obj);
  2811. return obj;
  2812. }