xref: /llvm-project/llvm/lib/Support/AutoConvert.cpp (revision 04379c98638ac3901257b5fa319f9ece828af767)
1 //===- AutoConvert.cpp - Auto conversion between ASCII/EBCDIC -------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file contains functions used for auto conversion between
10 // ASCII/EBCDIC codepages specific to z/OS.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifdef __MVS__
15 
16 #include "llvm/Support/AutoConvert.h"
17 #include "llvm/Support/Error.h"
18 #include <cassert>
19 #include <fcntl.h>
20 #include <sys/stat.h>
21 #include <unistd.h>
22 
23 using namespace llvm;
24 
25 static int savedStdHandleAutoConversionMode[3] = {-1, -1, -1};
26 
27 int disablezOSAutoConversion(int FD) {
28   static const struct f_cnvrt Convert = {
29       SETCVTOFF, // cvtcmd
30       0,         // pccsid
31       0,         // fccsid
32   };
33 
34   return fcntl(FD, F_CONTROL_CVT, &Convert);
35 }
36 
37 int restorezOSStdHandleAutoConversion(int FD) {
38   assert(FD == STDIN_FILENO || FD == STDOUT_FILENO || FD == STDERR_FILENO);
39   if (savedStdHandleAutoConversionMode[FD] == -1)
40     return 0;
41   struct f_cnvrt Cvt = {
42       savedStdHandleAutoConversionMode[FD], // cvtcmd
43       0,                                    // pccsid
44       0,                                    // fccsid
45   };
46   return (fcntl(FD, F_CONTROL_CVT, &Cvt));
47 }
48 
49 int enablezOSAutoConversion(int FD) {
50   struct f_cnvrt Query = {
51       QUERYCVT, // cvtcmd
52       0,        // pccsid
53       0,        // fccsid
54   };
55 
56   if (fcntl(FD, F_CONTROL_CVT, &Query) == -1)
57     return -1;
58 
59   // We don't need conversion for UTF-8 tagged files.
60   // TODO: Remove the assumption of ISO8859-1 = UTF-8 here when we fully resolve
61   // problems related to UTF-8 tagged source files.
62   // When the pccsid is not ISO8859-1, autoconversion is still needed.
63   if (Query.pccsid == CCSID_ISO8859_1 &&
64       (Query.fccsid == CCSID_UTF_8 || Query.fccsid == CCSID_ISO8859_1))
65     return 0;
66 
67   // Save the state of std handles before we make changes to it.
68   if ((FD == STDIN_FILENO || FD == STDOUT_FILENO || FD == STDERR_FILENO) &&
69       savedStdHandleAutoConversionMode[FD] == -1)
70     savedStdHandleAutoConversionMode[FD] = Query.cvtcmd;
71 
72   if (FD == STDOUT_FILENO || FD == STDERR_FILENO)
73     Query.cvtcmd = SETCVTON;
74   else
75     Query.cvtcmd = SETCVTALL;
76 
77   Query.pccsid =
78       (FD == STDIN_FILENO || FD == STDOUT_FILENO || FD == STDERR_FILENO)
79           ? 0
80           : CCSID_UTF_8;
81   // Assume untagged files to be IBM-1047 encoded.
82   Query.fccsid = (Query.fccsid == FT_UNTAGGED) ? CCSID_IBM_1047 : Query.fccsid;
83   return fcntl(FD, F_CONTROL_CVT, &Query);
84 }
85 
86 std::error_code llvm::disablezOSAutoConversion(int FD) {
87   if (::disablezOSAutoConversion(FD) == -1)
88     return errnoAsErrorCode();
89 
90   return std::error_code();
91 }
92 
93 std::error_code llvm::enablezOSAutoConversion(int FD) {
94   if (::enablezOSAutoConversion(FD) == -1)
95     return errnoAsErrorCode();
96 
97   return std::error_code();
98 }
99 
100 std::error_code llvm::restorezOSStdHandleAutoConversion(int FD) {
101   if (::restorezOSStdHandleAutoConversion(FD) == -1)
102     return errnoAsErrorCode();
103 
104   return std::error_code();
105 }
106 
107 std::error_code llvm::setzOSFileTag(int FD, int CCSID, bool Text) {
108   assert((!Text || (CCSID != FT_UNTAGGED && CCSID != FT_BINARY)) &&
109          "FT_UNTAGGED and FT_BINARY are not allowed for text files");
110   struct file_tag Tag;
111   Tag.ft_ccsid = CCSID;
112   Tag.ft_txtflag = Text;
113   Tag.ft_deferred = 0;
114   Tag.ft_rsvflags = 0;
115 
116   if (fcntl(FD, F_SETTAG, &Tag) == -1)
117     return errnoAsErrorCode();
118   return std::error_code();
119 }
120 
121 ErrorOr<__ccsid_t> llvm::getzOSFileTag(const char *FileName, const int FD) {
122   // If we have a file descriptor, use it to find out file tagging. Otherwise we
123   // need to use stat() with the file path.
124   if (FD != -1) {
125     struct f_cnvrt Query = {
126         QUERYCVT, // cvtcmd
127         0,        // pccsid
128         0,        // fccsid
129     };
130     if (fcntl(FD, F_CONTROL_CVT, &Query) == -1)
131       return std::error_code(errno, std::generic_category());
132     return Query.fccsid;
133   }
134   struct stat Attr;
135   if (stat(FileName, &Attr) == -1)
136     return std::error_code(errno, std::generic_category());
137   return Attr.st_tag.ft_ccsid;
138 }
139 
140 ErrorOr<bool> llvm::needzOSConversion(const char *FileName, const int FD) {
141   ErrorOr<__ccsid_t> Ccsid = getzOSFileTag(FileName, FD);
142   if (std::error_code EC = Ccsid.getError())
143     return EC;
144   // We don't need conversion for UTF-8 tagged files or binary files.
145   // TODO: Remove the assumption of ISO8859-1 = UTF-8 here when we fully resolve
146   // problems related to UTF-8 tagged source files.
147   switch (*Ccsid) {
148   case CCSID_UTF_8:
149   case CCSID_ISO8859_1:
150   case FT_BINARY:
151     return false;
152   default:
153     return true;
154   }
155 }
156 
157 #endif //__MVS__
158