X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fdata-io%2Fdata-writer.c;h=bf9505e7dd82878270a59ab13372c2a0137720c4;hb=cf5c60f6f2c5d7d6dcad09bc30162b70ad01c181;hp=b1b955c85478c5f5be2407f7e43b9b33690409ed;hpb=707848060e414fe93458834446dd7cdbf800667f;p=pspp diff --git a/src/language/data-io/data-writer.c b/src/language/data-io/data-writer.c index b1b955c854..bf9505e7dd 100644 --- a/src/language/data-io/data-writer.c +++ b/src/language/data-io/data-writer.c @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 1997-2004, 2006 Free Software Foundation, Inc. + Copyright (C) 1997-2004, 2006, 2010, 2011, 2012, 2013 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -16,25 +16,30 @@ #include -#include +#include "language/data-io/data-writer.h" #include #include +#include #include #include -#include -#include -#include -#include -#include -#include +#include "data/file-name.h" +#include "data/make-file.h" +#include "language/data-io/file-handle.h" +#include "libpspp/assertion.h" +#include "libpspp/encoding-guesser.h" +#include "libpspp/integer-format.h" +#include "libpspp/i18n.h" +#include "libpspp/message.h" +#include "libpspp/str.h" -#include "minmax.h" -#include "xalloc.h" +#include "gl/minmax.h" +#include "gl/xalloc.h" #include "gettext.h" #define _(msgid) gettext (msgid) +#define N_(msgid) (msgid) /* Data file writer. */ struct dfm_writer @@ -43,16 +48,35 @@ struct dfm_writer struct fh_lock *lock; /* Exclusive access to file. */ FILE *file; /* Associated file. */ struct replace_file *rf; /* Atomic file replacement support. */ + char *encoding; /* Encoding. */ + enum fh_line_ends line_ends; /* Line ends for text files. */ + + int unit; /* Unit width, in bytes. */ + char cr[MAX_UNIT]; /* \r in encoding, 'unit' bytes long. */ + char lf[MAX_UNIT]; /* \n in encoding, 'unit' bytes long. */ + char spaces[32]; /* 32 bytes worth of ' ' in encoding. */ }; -/* Opens a file handle for writing as a data file. */ +/* Opens a file handle for writing as a data file. + + The encoding of the file written is by default that of FH itself. If + ENCODING is nonnull, then it overrides the default encoding. + + *However*: ENCODING directly affects only text strings written by the data + writer code itself, that is, new-lines in FH_MODE_TEXT and space padding in + FH_MODE_FIXED mode. The client must do its own encoding translation for the + data that it writes. (This is unavoidable because sometimes the data + written includes binary data that reencoding would mangle.) The client can + obtain the encoding to re-encode into with dfm_writer_get_encoding(). */ struct dfm_writer * -dfm_open_writer (struct file_handle *fh) +dfm_open_writer (struct file_handle *fh, const char *encoding) { + struct encoding_info ei; struct dfm_writer *w; struct fh_lock *lock; + int ofs; - lock = fh_lock (fh, FH_REF_FILE, "data file", FH_ACC_WRITE, false); + lock = fh_lock (fh, FH_REF_FILE, N_("data file"), FH_ACC_WRITE, false); if (lock == NULL) return NULL; @@ -60,15 +84,26 @@ dfm_open_writer (struct file_handle *fh) if (w != NULL) return w; + encoding = encoding_guess_parse_encoding (encoding != NULL + ? encoding + : fh_get_encoding (fh)); + get_encoding_info (&ei, encoding); + w = xmalloc (sizeof *w); w->fh = fh_ref (fh); w->lock = lock; - w->rf = replace_file_start (fh_get_file_name (w->fh), "wb", - (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP - | S_IROTH | S_IWOTH), &w->file, NULL); + w->rf = replace_file_start (w->fh, "wb", 0666, &w->file); + w->encoding = xstrdup (encoding); + w->line_ends = fh_get_line_ends (fh); + w->unit = ei.unit; + memcpy (w->cr, ei.cr, sizeof w->cr); + memcpy (w->lf, ei.lf, sizeof w->lf); + for (ofs = 0; ofs + ei.unit <= sizeof w->spaces; ofs += ei.unit) + memcpy (&w->spaces[ofs], ei.space, ei.unit); + if (w->rf == NULL) { - msg (ME, _("An error occurred while opening \"%s\" for writing " + msg (ME, _("An error occurred while opening `%s' for writing " "as a data file: %s."), fh_get_file_name (w->fh), strerror (errno)); dfm_close_writer (w); @@ -79,7 +114,7 @@ dfm_open_writer (struct file_handle *fh) return w; } -/* Returns false if an I/O error occurred on WRITER, true otherwise. */ +/* Returns true if an I/O error occurred on WRITER, false otherwise. */ bool dfm_write_error (const struct dfm_writer *writer) { @@ -102,10 +137,12 @@ dfm_put_record (struct dfm_writer *w, const char *rec, size_t len) { case FH_MODE_TEXT: fwrite (rec, len, 1, w->file); - putc ('\n', w->file); + if (w->line_ends == FH_END_CRLF) + fwrite (w->cr, w->unit, 1, w->file); + fwrite (w->lf, w->unit, 1, w->file); break; - case FH_MODE_BINARY: + case FH_MODE_FIXED: { size_t record_width = fh_get_record_width (w->fh); size_t write_bytes = MIN (len, record_width); @@ -113,14 +150,52 @@ dfm_put_record (struct dfm_writer *w, const char *rec, size_t len) fwrite (rec, write_bytes, 1, w->file); while (pad_bytes > 0) { - static const char spaces[32] = " "; - size_t chunk = MIN (pad_bytes, sizeof spaces); - fwrite (spaces, chunk, 1, w->file); + size_t chunk = MIN (pad_bytes, sizeof w->spaces); + fwrite (w->spaces, chunk, 1, w->file); pad_bytes -= chunk; } } break; + case FH_MODE_VARIABLE: + { + uint32_t size = len; + integer_convert (INTEGER_NATIVE, &size, INTEGER_LSB_FIRST, &size, + sizeof size); + fwrite (&size, sizeof size, 1, w->file); + fwrite (rec, len, 1, w->file); + fwrite (&size, sizeof size, 1, w->file); + } + break; + + case FH_MODE_360_VARIABLE: + case FH_MODE_360_SPANNED: + { + size_t ofs = 0; + if (fh_get_mode (w->fh) == FH_MODE_360_VARIABLE) + len = MIN (65527, len); + while (ofs < len) + { + size_t chunk = MIN (65527, len - ofs); + uint32_t bdw = (chunk + 8) << 16; + int scc = (ofs == 0 && chunk == len ? 0 + : ofs == 0 ? 1 + : ofs + chunk == len ? 2 + : 3); + uint32_t rdw = ((chunk + 4) << 16) | (scc << 8); + + integer_convert (INTEGER_NATIVE, &bdw, INTEGER_MSB_FIRST, &bdw, + sizeof bdw); + integer_convert (INTEGER_NATIVE, &rdw, INTEGER_MSB_FIRST, &rdw, + sizeof rdw); + fwrite (&bdw, 1, sizeof bdw, w->file); + fwrite (&rdw, 1, sizeof rdw, w->file); + fwrite (rec + ofs, 1, chunk, w->file); + ofs += chunk; + } + } + break; + default: NOT_REACHED (); } @@ -143,16 +218,24 @@ dfm_close_writer (struct dfm_writer *w) if (w->file != NULL) { const char *file_name = fh_get_file_name (w->fh); - ok = !dfm_write_error (w) && !fn_close (file_name, w->file); + ok = !dfm_write_error (w) && !fn_close (w->fh, w->file); if (!ok) - msg (ME, _("I/O error occurred writing data file \"%s\"."), file_name); + msg (ME, _("I/O error occurred writing data file `%s'."), file_name); if (ok ? !replace_file_commit (w->rf) : !replace_file_abort (w->rf)) ok = false; } fh_unref (w->fh); + free (w->encoding); free (w); return ok; } + +/* Returns the encoding of data written to WRITER. */ +const char * +dfm_writer_get_encoding (const struct dfm_writer *writer) +{ + return writer->encoding; +}