utf8-posix.cpp@ 30294

Last change on this file since 30294 was 30294, checked in by vboxsync, 14 years ago
build fix
Property svn:eol-style set to `native` Property svn:keywords set to `Id`
File size: 16.2 KB

Line
1	/* $Id: utf8-posix.cpp 30294 2010-06-17 21:51:43Z 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.virtualbox.org. 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	#include <iconv.h>
42	#include <wctype.h>
43
44	#include <langinfo.h>
45
46	#include "internal/alignmentchecks.h"
47	#include "internal/string.h"
48	#ifdef RT_WITH_ICONV_CACHE
49	# include "internal/thread.h"
50	AssertCompile(sizeof(iconv_t) <= sizeof(void *));
51	#endif
52
53
54	/**
55	* Gets the codeset of the current locale (LC_CTYPE).
56	*
57	* @returns Pointer to read-only string with the codeset name.
58	*/
59	const char *rtStrGetLocaleCodeset(void)
60	{
61	return nl_langinfo(CODESET);
62	}
63
64
65	#ifdef RT_WITH_ICONV_CACHE
66
67	/**
68	* Initializes the iconv handle cache associated with a thread.
69	*
70	* @param pThread The thread in question.
71	*/
72	void rtStrIconvCacheInit(PRTTHREADINT pThread)
73	{
74	for (size_t i = 0; i < RT_ELEMENTS(pThread->ahIconvs); i++)
75	pThread->ahIconvs[i] = (iconv_t)-1;
76	}
77
78	/**
79	* Destroys the iconv handle cache associated with a thread.
80	*
81	* @param pThread The thread in question.
82	*/
83	void rtStrIconvCacheDestroy(PRTTHREADINT pThread)
84	{
85	for (size_t i = 0; i < RT_ELEMENTS(pThread->ahIconvs); i++)
86	{
87	iconv_t hIconv = (iconv_t)pThread->ahIconvs[i];
88	pThread->ahIconvs[i] = (iconv_t)-1;
89	if (hIconv != (iconv_t)-1)
90	iconv_close(hIconv);
91	}
92	}
93
94
95	/**
96	* Converts a string from one charset to another.
97	*
98	* @returns iprt status code.
99	* @param pvInput Pointer to intput string.
100	* @param cbInput Size (in bytes) of input string. Excludes any terminators.
101	* @param pszInputCS Codeset of the input string.
102	* @param ppvOutput Pointer to pointer to output buffer if cbOutput > 0.
103	* If cbOutput is 0 this is where the pointer to the allocated
104	* buffer is stored.
105	* @param cbOutput Size of the passed in buffer.
106	* @param pszOutputCS Codeset of the input string.
107	* @param cFactor Input vs. output size factor.
108	* @param phIconv Pointer to the cache entry.
109	*/
110	static int rtstrConvertCached(const void pvInput, size_t cbInput, const char pszInputCS,
111	void *ppvOutput, size_t cbOutput, const char pszOutputCS,
112	unsigned cFactor, iconv_t *phIconv)
113	{
114	/*
115	* Allocate buffer
116	*/
117	bool fUcs2Term;
118	void *pvOutput;
119	size_t cbOutput2;
120	if (!cbOutput)
121	{
122	cbOutput2 = cbInput * cFactor;
123	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
124	if (!pvOutput)
125	return VERR_NO_TMP_MEMORY;
126	fUcs2Term = true;
127	}
128	else
129	{
130	pvOutput = *ppvOutput;
131	fUcs2Term = !strcmp(pszOutputCS, "UCS-2")
132	\|\| !strcmp(pszOutputCS, "UTF-16")
133	\|\| !strcmp(pszOutputCS, "ucs-2")
134	\|\| !strcmp(pszOutputCS, "utf-16");
135	cbOutput2 = cbOutput - (fUcs2Term ? sizeof(RTUTF16) : 1);
136	if (cbOutput2 > cbOutput)
137	return VERR_BUFFER_OVERFLOW;
138	}
139
140	/*
141	* Use a loop here to retry with bigger buffers.
142	*/
143	for (unsigned cTries = 10; cTries > 0; cTries--)
144	{
145	/*
146	* Create conversion object if necessary.
147	*/
148	iconv_t hIconv = (iconv_t)*phIconv;
149	if (hIconv == (iconv_t)-1)
150	{
151	#ifdef RT_OS_SOLARIS
152	/* Solaris doesn't grok empty codeset strings, so help it find the current codeset. */
153	if (!*pszInputCS)
154	pszInputCS = rtStrGetLocaleCodeset();
155	if (!*pszOutputCS)
156	pszOutputCS = rtStrGetLocaleCodeset();
157	#endif
158	IPRT_ALIGNMENT_CHECKS_DISABLE(); /* glibc causes trouble */
159	*phIconv = hIconv = iconv_open(pszOutputCS, pszInputCS);
160	IPRT_ALIGNMENT_CHECKS_ENABLE();
161	}
162	if (hIconv != (iconv_t)-1)
163	{
164	/*
165	* Do the conversion.
166	*/
167	size_t cbInLeft = cbInput;
168	size_t cbOutLeft = cbOutput2;
169	const void *pvInputLeft = pvInput;
170	void *pvOutputLeft = pvOutput;
171	#if defined(RT_OS_LINUX) \|\| (defined(RT_OS_DARWIN) && defined(_DARWIN_FEATURE_UNIX_CONFORMANCE)) /* there are different opinions about the constness of the input buffer. */
172	if (iconv(hIconv, (char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
173	#else
174	if (iconv(hIconv, (const char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
175	#endif
176	{
177	if (!cbInLeft)
178	{
179	/*
180	* We're done, just add the terminator and return.
181	* (Two terminators to support UCS-2 output, too.)
182	*/
183	((char *)pvOutputLeft)[0] = '\0';
184	if (fUcs2Term)
185	((char *)pvOutputLeft)[1] = '\0';
186	*ppvOutput = pvOutput;
187	return VINF_SUCCESS;
188	}
189	errno = E2BIG;
190	}
191
192	/*
193	* If we failed because of output buffer space we'll
194	* increase the output buffer size and retry.
195	*/
196	if (errno == E2BIG)
197	{
198	if (!cbOutput)
199	{
200	RTMemTmpFree(pvOutput);
201	cbOutput2 *= 2;
202	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
203	if (!pvOutput)
204	return VERR_NO_TMP_MEMORY;
205	continue;
206	}
207	return VERR_BUFFER_OVERFLOW;
208	}
209
210	/*
211	* Close the handle on all other errors to make sure we won't carry
212	* any bad state with us.
213	*/
214	*phIconv = (iconv_t)-1;
215	iconv_close(hIconv);
216	}
217	break;
218	}
219
220	/* failure */
221	if (!cbOutput)
222	RTMemTmpFree(pvOutput);
223	return VERR_NO_TRANSLATION;
224	}
225
226	#endif /* RT_WITH_ICONV_CACHE */
227
228	/**
229	* Converts a string from one charset to another without using the handle cache.
230	*
231	* @returns IPRT status code.
232	*
233	* @param pvInput Pointer to intput string.
234	* @param cbInput Size (in bytes) of input string. Excludes any terminators.
235	* @param pszInputCS Codeset of the input string.
236	* @param ppvOutput Pointer to pointer to output buffer if cbOutput > 0.
237	* If cbOutput is 0 this is where the pointer to the allocated
238	* buffer is stored.
239	* @param cbOutput Size of the passed in buffer.
240	* @param pszOutputCS Codeset of the input string.
241	* @param cFactor Input vs. output size factor.
242	*/
243	static int rtStrConvertUncached(const void pvInput, size_t cbInput, const char pszInputCS,
244	void *ppvOutput, size_t cbOutput, const char pszOutputCS,
245	unsigned cFactor)
246	{
247	/*
248	* Allocate buffer
249	*/
250	bool fUcs2Term;
251	void *pvOutput;
252	size_t cbOutput2;
253	if (!cbOutput)
254	{
255	cbOutput2 = cbInput * cFactor;
256	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
257	if (!pvOutput)
258	return VERR_NO_TMP_MEMORY;
259	fUcs2Term = true;
260	}
261	else
262	{
263	pvOutput = *ppvOutput;
264	fUcs2Term = !strcmp(pszOutputCS, "UCS-2");
265	cbOutput2 = cbOutput - (fUcs2Term ? sizeof(RTUTF16) : 1);
266	if (cbOutput2 > cbOutput)
267	return VERR_BUFFER_OVERFLOW;
268	}
269
270	/*
271	* Use a loop here to retry with bigger buffers.
272	*/
273	for (unsigned cTries = 10; cTries > 0; cTries--)
274	{
275	/*
276	* Create conversion object.
277	*/
278	#ifdef RT_OS_SOLARIS
279	/* Solaris doesn't grok empty codeset strings, so help it find the current codeset. */
280	if (!*pszInputCS)
281	pszInputCS = rtStrGetLocaleCodeset();
282	if (!*pszOutputCS)
283	pszOutputCS = rtStrGetLocaleCodeset();
284	#endif
285	IPRT_ALIGNMENT_CHECKS_DISABLE(); /* glibc causes trouble */
286	iconv_t icHandle = iconv_open(pszOutputCS, pszInputCS);
287	IPRT_ALIGNMENT_CHECKS_ENABLE();
288	if (icHandle != (iconv_t)-1)
289	{
290	/*
291	* Do the conversion.
292	*/
293	size_t cbInLeft = cbInput;
294	size_t cbOutLeft = cbOutput2;
295	const void *pvInputLeft = pvInput;
296	void *pvOutputLeft = pvOutput;
297	#if defined(RT_OS_LINUX) \|\| (defined(RT_OS_DARWIN) && defined(_DARWIN_FEATURE_UNIX_CONFORMANCE)) /* there are different opinions about the constness of the input buffer. */
298	if (iconv(icHandle, (char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
299	#else
300	if (iconv(icHandle, (const char )&pvInputLeft, &cbInLeft, (char )&pvOutputLeft, &cbOutLeft) != (size_t)-1)
301	#endif
302	{
303	if (!cbInLeft)
304	{
305	/*
306	* We're done, just add the terminator and return.
307	* (Two terminators to support UCS-2 output, too.)
308	*/
309	iconv_close(icHandle);
310	((char *)pvOutputLeft)[0] = '\0';
311	if (fUcs2Term)
312	((char *)pvOutputLeft)[1] = '\0';
313	*ppvOutput = pvOutput;
314	return VINF_SUCCESS;
315	}
316	errno = E2BIG;
317	}
318	iconv_close(icHandle);
319
320	/*
321	* If we failed because of output buffer space we'll
322	* increase the output buffer size and retry.
323	*/
324	if (errno == E2BIG)
325	{
326	if (!cbOutput)
327	{
328	RTMemTmpFree(pvOutput);
329	cbOutput2 *= 2;
330	pvOutput = RTMemTmpAlloc(cbOutput2 + sizeof(RTUTF16));
331	if (!pvOutput)
332	return VERR_NO_TMP_MEMORY;
333	continue;
334	}
335	return VERR_BUFFER_OVERFLOW;
336	}
337	}
338	break;
339	}
340
341	/* failure */
342	if (!cbOutput)
343	RTMemTmpFree(pvOutput);
344	return VERR_NO_TRANSLATION;
345	}
346
347
348	/**
349	* Wrapper that selects rtStrConvertCached or rtStrConvertUncached.
350	*
351	* @returns IPRT status code.
352	*
353	* @param pszInput Pointer to intput string.
354	* @param cchInput Size (in bytes) of input string. Excludes any
355	* terminators.
356	* @param pszInputCS Codeset of the input string.
357	* @param ppszOutput Pointer to pointer to output buffer if cbOutput > 0.
358	* If cbOutput is 0 this is where the pointer to the
359	* allocated buffer is stored.
360	* @param cbOutput Size of the passed in buffer.
361	* @param pszOutputCS Codeset of the input string.
362	* @param cFactor Input vs. output size factor.
363	* @param enmCacheIdx The iconv cache index.
364	*/
365	DECLINLINE(int) rtStrConvertWrapper(const char pchInput, size_t cchInput, const char pszInputCS,
366	char *ppszOutput, size_t cbOutput, const char pszOutputCS,
367	unsigned cFactor, RTSTRICONV enmCacheIdx)
368	{
369	#ifdef RT_WITH_ICONV_CACHE
370	RTTHREAD hSelf = RTThreadSelf();
371	if (hSelf != NIL_RTTHREAD)
372	{
373	PRTTHREADINT pThread = rtThreadGet(hSelf);
374	if ( pThread
375	&& (pThread->fIntFlags & (RTTHREADINT_FLAGS_ALIEN \| RTTHREADINT_FLAGS_MAIN)) != RTTHREADINT_FLAGS_ALIEN)
376	return rtstrConvertCached(pchInput, cchInput, pszInputCS,
377	(void **)ppszOutput, cbOutput, pszOutputCS,
378	cFactor, (iconv_t *)&pThread->ahIconvs[enmCacheIdx]);
379	}
380	#endif
381	return rtStrConvertUncached(pchInput, cchInput, pszInputCS,
382	(void **)ppszOutput, cbOutput, pszOutputCS,
383	cFactor);
384	}
385
386
387	/**
388	* Internal API for use by the path conversion code.
389	*
390	* @returns IPRT status code.
391	*
392	* @param pszInput Pointer to intput string.
393	* @param cchInput Size (in bytes) of input string. Excludes any
394	* terminators.
395	* @param pszInputCS Codeset of the input string.
396	* @param ppszOutput Pointer to pointer to output buffer if cbOutput > 0.
397	* If cbOutput is 0 this is where the pointer to the
398	* allocated buffer is stored.
399	* @param cbOutput Size of the passed in buffer.
400	* @param pszOutputCS Codeset of the input string.
401	* @param cFactor Input vs. output size factor.
402	* @param enmCacheIdx The iconv cache index.
403	*/
404	int rtStrConvert(const char pchInput, size_t cchInput, const char pszInputCS,
405	char *ppszOutput, size_t cbOutput, const char pszOutputCS,
406	unsigned cFactor, RTSTRICONV enmCacheIdx)
407	{
408	Assert(enmCacheIdx >= 0 && enmCacheIdx < RTSTRICONV_END);
409	return rtStrConvertWrapper(pchInput, cchInput, pszInputCS,
410	ppszOutput, cbOutput, pszOutputCS,
411	cFactor, enmCacheIdx);
412	}
413
414
415	/**
416	* Allocates tmp buffer, translates pszString from UTF8 to current codepage.
417	*
418	* @returns iprt status code.
419	* @param ppszString Receives pointer of allocated native CP string.
420	* The returned pointer must be freed using RTStrFree().
421	* @param pszString UTF-8 string to convert.
422	*/
423	RTR3DECL(int) RTStrUtf8ToCurrentCP(char *ppszString, const char pszString)
424	{
425	Assert(ppszString);
426	Assert(pszString);
427	*ppszString = NULL;
428
429	/*
430	* Assume result string length is not longer than UTF-8 string.
431	*/
432	size_t cch = strlen(pszString);
433	if (cch <= 0)
434	{
435	/* zero length string passed. */
436	ppszString = (char )RTMemTmpAllocZ(sizeof(char));
437	if (*ppszString)
438	return VINF_SUCCESS;
439	return VERR_NO_TMP_MEMORY;
440	}
441	return rtStrConvertWrapper(pszString, cch, "UTF-8", ppszString, 0, "", 1, RTSTRICONV_UTF8_TO_LOCALE);
442	}
443
444
445	/**
446	* Allocates tmp buffer, translates pszString from current codepage to UTF-8.
447	*
448	* @returns iprt status code.
449	* @param ppszString Receives pointer of allocated UTF-8 string.
450	* The returned pointer must be freed using RTStrFree().
451	* @param pszString Native string to convert.
452	*/
453	RTR3DECL(int) RTStrCurrentCPToUtf8(char *ppszString, const char pszString)
454	{
455	Assert(ppszString);
456	Assert(pszString);
457	*ppszString = NULL;
458
459	/*
460	* Attempt with UTF-8 length of 2x the native length.
461	*/
462	size_t cch = strlen(pszString);
463	if (cch <= 0)
464	{
465	/* zero length string passed. */
466	ppszString = (char )RTMemTmpAllocZ(sizeof(char));
467	if (*ppszString)
468	return VINF_SUCCESS;
469	return VERR_NO_TMP_MEMORY;
470	}
471	return rtStrConvertWrapper(pszString, cch, "", ppszString, 0, "UTF-8", 2, RTSTRICONV_LOCALE_TO_UTF8);
472	}
473

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

source: vbox/trunk/src/VBox/Runtime/r3/posix/utf8-posix.cpp@ 30294

Download in other formats: