utf8-posix.cpp@ 40651

Last change on this file since 40651 was 40651, checked in by vboxsync, 13 years ago
Runtime/r3/posix, RDP/client: Solaris 11 iconv change. Avoid referencing xpg5_iconv symbol by circumventing pragma redefine_extname.
Property svn:eol-style set to `native` Property svn:keywords set to `Id`
File size: 16.4 KB

Line
1	/* $Id: utf8-posix.cpp 40651 2012-03-26 16:29:58Z vboxsync $ */
2	/** @file
3	* IPRT - UTF-8 helpers, POSIX.
4	*/
5
6	/*
7	* Copyright (C) 2006-2010 Oracle Corporation
8	*
9	* This file is part of VirtualBox Open Source Edition (OSE), as
10	* available from http://www.215389.xyz. This file is free software;
11	* you can redistribute it and/or modify it under the terms of the GNU
12	* General Public License (GPL) as published by the Free Software
13	* Foundation, in version 2 as it comes in the "COPYING" file of the
14	* VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15	* hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16	*
17	* The contents of this file may alternatively be used under the terms
18	* of the Common Development and Distribution License Version 1.0
19	* (CDDL) only, as it comes in the "COPYING.CDDL" file of the
20	* VirtualBox OSE distribution, in which case the provisions of the
21	* CDDL are applicable instead of those of the GPL.
22	*
23	* You may elect to license modified versions of this file under the
24	* terms and conditions of either the GPL or the CDDL or both.
25	*/
26
27
28	/*******************************************************************************
29	* Header Files *
30	*******************************************************************************/
31	#include <iprt/string.h>
32	#include "internal/iprt.h"
33
34	#include <iprt/alloc.h>
35	#include <iprt/assert.h>
36	#include <iprt/err.h>
37	#include <iprt/string.h>
38
39	#include <errno.h>
40	#include <locale.h>
41
42	#if defined(RT_OS_SOLARIS)
43	# if !defined(_XPG6)
44	# define VBOX_XPG6_TMP_DEF
45	# define _XPG6
46	# endif
47	# if defined(__USE_LEGACY_PROTOTYPES__)
48	# define VBOX_LEGACY_PROTO_TMP_DEF
49	# undef __USE_LEGACY_PROTOTYPES__
50	# endif
51	#endif /* RT_OS_SOLARIS */
52
53	# include <iconv.h>
54
55	#if defined(RT_OS_SOLARIS)
56	# if defined(VBOX_XPG6_TMP_DEF)
57	# undef _XPG6
58	# undef VBOX_XPG6_TMP_DEF
59	# endif
60	# if defined(VBOX_LEGACY_PROTO_TMP_DEF)
61	# define __USE_LEGACY_PROTOTYPES__
62	# undef VBOX_LEGACY_PROTO_TMP_DEF
63	# endif
64	#endif /* RT_OS_SOLARIS */
65
66	#include <wctype.h>
67
68	#include <langinfo.h>
69
70	#include "internal/alignmentchecks.h"
71	#include "internal/string.h"
72	#ifdef RT_WITH_ICONV_CACHE
73	# include "internal/thread.h"
74	AssertCompile(sizeof(iconv_t) <= sizeof(void *));
75	#endif
76
77
78	/**
79	* Gets the codeset of the current locale (LC_CTYPE).
80	*
81	* @returns Pointer to read-only string with the codeset name.
82	*/
83	DECLHIDDEN(const char *) rtStrGetLocaleCodeset(void)
84	{
85	return nl_langinfo(CODESET);
86	}
87
88
89	#ifdef RT_WITH_ICONV_CACHE
90
91	/**
92	* Initializes the iconv handle cache associated with a thread.
93	*
94	* @param pThread The thread in question.
95	*/
96	DECLHIDDEN(void) rtStrIconvCacheInit(PRTTHREADINT pThread)
97	{
98	for (size_t i = 0; i < RT_ELEMENTS(pThread->ahIconvs); i++)
99	pThread->ahIconvs[i] = (iconv_t)-1;
100	}
101
102	/**
103	* Destroys the iconv handle cache associated with a thread.
104	*
105	* @param pThread The thread in question.
106	*/
107	DECLHIDDEN(void) rtStrIconvCacheDestroy(PRTTHREADINT pThread)
108	{
109	for (size_t i = 0; i < RT_ELEMENTS(pThread->ahIconvs); i++)
110	{
111	iconv_t hIconv = (iconv_t)pThread->ahIconvs[i];
112	pThread->ahIconvs[i] = (iconv_t)-1;
113	if (hIconv != (iconv_t)-1)
114	iconv_close(hIconv);
115	}
116	}
117
118
119	/**
120	* Converts a string from one charset to another.
121	*
122	* @returns iprt status code.
123	* @param pvInput Pointer to intput string.
124	* @param cbInput Size (in bytes) of input string. Excludes any terminators.
125	* @param pszInputCS Codeset of the input string.
126	* @param ppvOutput Pointer to pointer to output buffer if cbOutput > 0.
127	* If cbOutput is 0 this is where the pointer to the allocated
128	* buffer is stored.
129	* @param cbOutput Size of the passed in buffer.
130	* @param pszOutputCS Codeset of the input string.
131	* @param cFactor Input vs. output size factor.
132	* @param phIconv Pointer to the cache entry.
133	*/
134	static int rtstrConvertCached(const void pvInput, size_t cbInput, const char pszInputCS,
135	void *ppvOutput, size_t cbOutput, const char pszOutputCS,
136	unsigned cFactor, iconv_t *phIconv)
137	{
138	/*
139	* Allocate buffer
140	*/
141	bool fUcs2Term;
142	void *pvOutput;
143	size_t cbOutput2;
144	if (!cbOutput)
145	{
146	cbOutput2 = cbInput * cFactor;
147	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
148	if (!pvOutput)
149	return VERR_NO_TMP_MEMORY;
150	fUcs2Term = true;
151	}
152	else
153	{
154	pvOutput = *ppvOutput;
155	fUcs2Term = !strcmp(pszOutputCS, "UCS-2")
156	\|\| !strcmp(pszOutputCS, "UTF-16")
157	\|\| !strcmp(pszOutputCS, "ucs-2")
158	\|\| !strcmp(pszOutputCS, "utf-16");
159	cbOutput2 = cbOutput - (fUcs2Term ? sizeof(RTUTF16) : 1);
160	if (cbOutput2 > cbOutput)
161	return VERR_BUFFER_OVERFLOW;
162	}
163
164	/*
165	* Use a loop here to retry with bigger buffers.
166	*/
167	for (unsigned cTries = 10; cTries > 0; cTries--)
168	{
169	/*
170	* Create conversion object if necessary.
171	*/
172	iconv_t hIconv = (iconv_t)*phIconv;
173	if (hIconv == (iconv_t)-1)
174	{
175	#ifdef RT_OS_SOLARIS
176	/* Solaris doesn't grok empty codeset strings, so help it find the current codeset. */
177	if (!*pszInputCS)
178	pszInputCS = rtStrGetLocaleCodeset();
179	if (!*pszOutputCS)
180	pszOutputCS = rtStrGetLocaleCodeset();
181	#endif
182	IPRT_ALIGNMENT_CHECKS_DISABLE(); /* glibc causes trouble */
183	*phIconv = hIconv = iconv_open(pszOutputCS, pszInputCS);
184	IPRT_ALIGNMENT_CHECKS_ENABLE();
185	}
186	if (hIconv != (iconv_t)-1)
187	{
188	/*
189	* Do the conversion.
190	*/
191	size_t cbInLeft = cbInput;
192	size_t cbOutLeft = cbOutput2;
193	const void *pvInputLeft = pvInput;
194	void *pvOutputLeft = pvOutput;
195	#if defined(RT_OS_LINUX) \|\| (defined(RT_OS_SOLARIS) && !defined(VBOX_SOLARIS_11_LEGACY_ICONV)) \|\| (defined(RT_OS_DARWIN) && defined(_DARWIN_FEATURE_UNIX_CONFORMANCE)) /* there are different opinions about the constness of the input buffer. */
196	if (iconv(hIconv, (char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
197	#else
198	if (iconv(hIconv, (const char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
199	#endif
200	{
201	if (!cbInLeft)
202	{
203	/*
204	* We're done, just add the terminator and return.
205	* (Two terminators to support UCS-2 output, too.)
206	*/
207	((char *)pvOutputLeft)[0] = '\0';
208	if (fUcs2Term)
209	((char *)pvOutputLeft)[1] = '\0';
210	*ppvOutput = pvOutput;
211	return VINF_SUCCESS;
212	}
213	errno = E2BIG;
214	}
215
216	/*
217	* If we failed because of output buffer space we'll
218	* increase the output buffer size and retry.
219	*/
220	if (errno == E2BIG)
221	{
222	if (!cbOutput)
223	{
224	RTMemTmpFree(pvOutput);
225	cbOutput2 *= 2;
226	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
227	if (!pvOutput)
228	return VERR_NO_TMP_MEMORY;
229	continue;
230	}
231	return VERR_BUFFER_OVERFLOW;
232	}
233
234	/*
235	* Close the handle on all other errors to make sure we won't carry
236	* any bad state with us.
237	*/
238	*phIconv = (iconv_t)-1;
239	iconv_close(hIconv);
240	}
241	break;
242	}
243
244	/* failure */
245	if (!cbOutput)
246	RTMemTmpFree(pvOutput);
247	return VERR_NO_TRANSLATION;
248	}
249
250	#endif /* RT_WITH_ICONV_CACHE */
251
252	/**
253	* Converts a string from one charset to another without using the handle cache.
254	*
255	* @returns IPRT status code.
256	*
257	* @param pvInput Pointer to intput string.
258	* @param cbInput Size (in bytes) of input string. Excludes any terminators.
259	* @param pszInputCS Codeset of the input string.
260	* @param ppvOutput Pointer to pointer to output buffer if cbOutput > 0.
261	* If cbOutput is 0 this is where the pointer to the allocated
262	* buffer is stored.
263	* @param cbOutput Size of the passed in buffer.
264	* @param pszOutputCS Codeset of the input string.
265	* @param cFactor Input vs. output size factor.
266	*/
267	static int rtStrConvertUncached(const void pvInput, size_t cbInput, const char pszInputCS,
268	void *ppvOutput, size_t cbOutput, const char pszOutputCS,
269	unsigned cFactor)
270	{
271	/*
272	* Allocate buffer
273	*/
274	bool fUcs2Term;
275	void *pvOutput;
276	size_t cbOutput2;
277	if (!cbOutput)
278	{
279	cbOutput2 = cbInput * cFactor;
280	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
281	if (!pvOutput)
282	return VERR_NO_TMP_MEMORY;
283	fUcs2Term = true;
284	}
285	else
286	{
287	pvOutput = *ppvOutput;
288	fUcs2Term = !strcmp(pszOutputCS, "UCS-2");
289	cbOutput2 = cbOutput - (fUcs2Term ? sizeof(RTUTF16) : 1);
290	if (cbOutput2 > cbOutput)
291	return VERR_BUFFER_OVERFLOW;
292	}
293
294	/*
295	* Use a loop here to retry with bigger buffers.
296	*/
297	for (unsigned cTries = 10; cTries > 0; cTries--)
298	{
299	/*
300	* Create conversion object.
301	*/
302	#ifdef RT_OS_SOLARIS
303	/* Solaris doesn't grok empty codeset strings, so help it find the current codeset. */
304	if (!*pszInputCS)
305	pszInputCS = rtStrGetLocaleCodeset();
306	if (!*pszOutputCS)
307	pszOutputCS = rtStrGetLocaleCodeset();
308	#endif
309	IPRT_ALIGNMENT_CHECKS_DISABLE(); /* glibc causes trouble */
310	iconv_t icHandle = iconv_open(pszOutputCS, pszInputCS);
311	IPRT_ALIGNMENT_CHECKS_ENABLE();
312	if (icHandle != (iconv_t)-1)
313	{
314	/*
315	* Do the conversion.
316	*/
317	size_t cbInLeft = cbInput;
318	size_t cbOutLeft = cbOutput2;
319	const void *pvInputLeft = pvInput;
320	void *pvOutputLeft = pvOutput;
321	#if defined(RT_OS_LINUX) \|\| (defined(RT_OS_SOLARIS) && !defined(VBOX_SOLARIS_11_LEGACY_ICONV)) \|\| (defined(RT_OS_DARWIN) && defined(_DARWIN_FEATURE_UNIX_CONFORMANCE)) /* there are different opinions about the constness of the input buffer. */
322	if (iconv(icHandle, (char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
323	#else
324	if (iconv(icHandle, (const char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
325	#endif
326	{
327	if (!cbInLeft)
328	{
329	/*
330	* We're done, just add the terminator and return.
331	* (Two terminators to support UCS-2 output, too.)
332	*/
333	iconv_close(icHandle);
334	((char *)pvOutputLeft)[0] = '\0';
335	if (fUcs2Term)
336	((char *)pvOutputLeft)[1] = '\0';
337	*ppvOutput = pvOutput;
338	return VINF_SUCCESS;
339	}
340	errno = E2BIG;
341	}
342	iconv_close(icHandle);
343
344	/*
345	* If we failed because of output buffer space we'll
346	* increase the output buffer size and retry.
347	*/
348	if (errno == E2BIG)
349	{
350	if (!cbOutput)
351	{
352	RTMemTmpFree(pvOutput);
353	cbOutput2 *= 2;
354	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
355	if (!pvOutput)
356	return VERR_NO_TMP_MEMORY;
357	continue;
358	}
359	return VERR_BUFFER_OVERFLOW;
360	}
361	}
362	break;
363	}
364
365	/* failure */
366	if (!cbOutput)
367	RTMemTmpFree(pvOutput);
368	return VERR_NO_TRANSLATION;
369	}
370
371
372	/**
373	* Wrapper that selects rtStrConvertCached or rtStrConvertUncached.
374	*
375	* @returns IPRT status code.
376	*
377	* @param pszInput Pointer to intput string.
378	* @param cchInput Size (in bytes) of input string. Excludes any
379	* terminators.
380	* @param pszInputCS Codeset of the input string.
381	* @param ppszOutput Pointer to pointer to output buffer if cbOutput > 0.
382	* If cbOutput is 0 this is where the pointer to the
383	* allocated buffer is stored.
384	* @param cbOutput Size of the passed in buffer.
385	* @param pszOutputCS Codeset of the input string.
386	* @param cFactor Input vs. output size factor.
387	* @param enmCacheIdx The iconv cache index.
388	*/
389	DECLINLINE(int) rtStrConvertWrapper(const char pchInput, size_t cchInput, const char pszInputCS,
390	char *ppszOutput, size_t cbOutput, const char pszOutputCS,
391	unsigned cFactor, RTSTRICONV enmCacheIdx)
392	{
393	#ifdef RT_WITH_ICONV_CACHE
394	RTTHREAD hSelf = RTThreadSelf();
395	if (hSelf != NIL_RTTHREAD)
396	{
397	PRTTHREADINT pThread = rtThreadGet(hSelf);
398	if (pThread)
399	{
400	if ((pThread->fIntFlags & (RTTHREADINT_FLAGS_ALIEN \| RTTHREADINT_FLAGS_MAIN)) != RTTHREADINT_FLAGS_ALIEN)
401	{
402	int rc = rtstrConvertCached(pchInput, cchInput, pszInputCS,
403	(void **)ppszOutput, cbOutput, pszOutputCS,
404	cFactor, (iconv_t *)&pThread->ahIconvs[enmCacheIdx]);
405	rtThreadRelease(pThread);
406	return rc;
407	}
408	rtThreadRelease(pThread);
409	}
410	}
411	#endif
412	return rtStrConvertUncached(pchInput, cchInput, pszInputCS,
413	(void **)ppszOutput, cbOutput, pszOutputCS,
414	cFactor);
415	}
416
417
418	/**
419	* Internal API for use by the path conversion code.
420	*
421	* @returns IPRT status code.
422	*
423	* @param pszInput Pointer to intput string.
424	* @param cchInput Size (in bytes) of input string. Excludes any
425	* terminators.
426	* @param pszInputCS Codeset of the input string.
427	* @param ppszOutput Pointer to pointer to output buffer if cbOutput > 0.
428	* If cbOutput is 0 this is where the pointer to the
429	* allocated buffer is stored.
430	* @param cbOutput Size of the passed in buffer.
431	* @param pszOutputCS Codeset of the input string.
432	* @param cFactor Input vs. output size factor.
433	* @param enmCacheIdx The iconv cache index.
434	*/
435	DECLHIDDEN(int) rtStrConvert(const char pchInput, size_t cchInput, const char pszInputCS,
436	char *ppszOutput, size_t cbOutput, const char pszOutputCS,
437	unsigned cFactor, RTSTRICONV enmCacheIdx)
438	{
439	Assert(enmCacheIdx >= 0 && enmCacheIdx < RTSTRICONV_END);
440	return rtStrConvertWrapper(pchInput, cchInput, pszInputCS,
441	ppszOutput, cbOutput, pszOutputCS,
442	cFactor, enmCacheIdx);
443	}
444
445
446	RTR3DECL(int) RTStrUtf8ToCurrentCPTag(char *ppszString, const char pszString, const char *pszTag)
447	{
448	Assert(ppszString);
449	Assert(pszString);
450	*ppszString = NULL;
451
452	/*
453	* Assume result string length is not longer than UTF-8 string.
454	*/
455	size_t cch = strlen(pszString);
456	if (cch <= 0)
457	{
458	/* zero length string passed. */
459	ppszString = (char )RTMemTmpAllocZTag(sizeof(char), pszTag);
460	if (*ppszString)
461	return VINF_SUCCESS;
462	return VERR_NO_TMP_MEMORY;
463	}
464	return rtStrConvertWrapper(pszString, cch, "UTF-8", ppszString, 0, "", 1, RTSTRICONV_UTF8_TO_LOCALE);
465	}
466
467
468	RTR3DECL(int) RTStrCurrentCPToUtf8Tag(char *ppszString, const char pszString, const char *pszTag)
469	{
470	Assert(ppszString);
471	Assert(pszString);
472	*ppszString = NULL;
473
474	/*
475	* Attempt with UTF-8 length of 2x the native length.
476	*/
477	size_t cch = strlen(pszString);
478	if (cch <= 0)
479	{
480	/* zero length string passed. */
481	ppszString = (char )RTMemTmpAllocZTag(sizeof(char), pszTag);
482	if (*ppszString)
483	return VINF_SUCCESS;
484	return VERR_NO_TMP_MEMORY;
485	}
486	return rtStrConvertWrapper(pszString, cch, "", ppszString, 0, "UTF-8", 2, RTSTRICONV_LOCALE_TO_UTF8);
487	}
488

Note: See TracBrowser for help on using the repository browser.

source: vbox/trunk/src/VBox/Runtime/r3/posix/utf8-posix.cpp@ 40651

Download in other formats: