X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/91baa2e0868d9d2e5cf664d63f4fbba25c4470cf..e38cc16e06bbbd80a0fc02585e736fbc3dee7221:/uip/mhfixmsg.c?ds=inline diff --git a/uip/mhfixmsg.c b/uip/mhfixmsg.c index 1c3a632e..9f317acd 100644 --- a/uip/mhfixmsg.c +++ b/uip/mhfixmsg.c @@ -23,6 +23,8 @@ X("notextcodeset", 0, NTEXTCODESETSW) \ X("reformat", 0, REFORMATSW) \ X("noreformat", 0, NREFORMATSW) \ + X("replacetextplain", 0, REPLACETEXTPLAINSW) \ + X("noreplacetextplain", 0, NREPLACETEXTPLAINSW) \ X("fixboundary", 0, FIXBOUNDARYSW) \ X("nofixboundary", 0, NFIXBOUNDARYSW) \ X("fixcte", 0, FIXCTESW) \ @@ -51,7 +53,6 @@ int debugsw; /* Needed by mhparse.c. */ #define quitser pipeser /* mhparse.c */ -extern char *tmp; /* directory to place tmp files */ extern int skip_mp_cte_check; /* flag to InitMultiPart */ extern int suppress_bogus_mp_content_warning; /* flag to InitMultiPart */ extern int bogus_mp_content; /* flag from InitMultiPart */ @@ -78,6 +79,7 @@ typedef struct fix_transformations { int fixboundary; int fixcte; int reformat; + int replacetextplain; int decodetext; char *textcodeset; } fix_transformations; @@ -90,7 +92,7 @@ static int replace_boundary (CT, char *, const char *); static char *update_attr (char *, const char *, const char *e); static int fix_multipart_cte (CT, int *); static int set_ce (CT, int); -static int ensure_text_plain (CT *, CT, int *); +static int ensure_text_plain (CT *, CT, int *, int); static CT build_text_plain_part (CT); static CT divide_part (CT); static void copy_ctinfo (CI, CI); @@ -103,7 +105,7 @@ static void transfer_noncontent_headers (CT, CT); static int set_ct_type (CT, int type, int subtype, int encoding); static int decode_text_parts (CT, int, int *); static int content_encoding (CT); -static int strip_crs (CT); +static int strip_crs (CT, int *); static int convert_codesets (CT, char *, int *); static int convert_codeset (CT, char *, int *); static char *content_codeset (CT); @@ -128,18 +130,13 @@ main (int argc, char **argv) { int status = OK; fix_transformations fx; fx.reformat = fx.fixcte = fx.fixboundary = 1; + fx.replacetextplain = 0; fx.decodetext = CE_8BIT; fx.textcodeset = NULL; - done = freects_done; - -#ifdef LOCALE - setlocale(LC_ALL, ""); -#endif - invo_name = r1bindex (argv[0], '/'); + if (nmh_init(argv[0], 1)) { return 1; } - /* read user profile/context */ - context_read(); + done = freects_done; arguments = getarguments (invo_name, argc, argv, 1); argp = arguments; @@ -205,19 +202,22 @@ main (int argc, char **argv) { case NREFORMATSW: fx.reformat = 0; continue; - + case REPLACETEXTPLAINSW: + fx.replacetextplain = 1; + continue; + case NREPLACETEXTPLAINSW: + fx.replacetextplain = 0; + continue; case FILESW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) adios (NULL, "missing argument to %s", argp[-2]); file = *cp == '-' ? add (cp, NULL) : path (cp, TFILE); continue; - case OUTFILESW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) adios (NULL, "missing argument to %s", argp[-2]); outfile = *cp == '-' ? add (cp, NULL) : path (cp, TFILE); continue; - case RPROCSW: if (!(rmmproc = *argp++) || *rmmproc == '-') adios (NULL, "missing argument to %s", argp[-2]); @@ -225,7 +225,6 @@ main (int argc, char **argv) { case NRPRCSW: rmmproc = NULL; continue; - case VERBSW: verbosw = 1; continue; @@ -239,8 +238,14 @@ main (int argc, char **argv) { adios (NULL, "only one folder at a time!"); else folder = pluspath (cp); - } else - app_msgarg(&msgs, cp); + } else { + if (*cp == '/') { + /* Interpret a full path as a filename, not a message. */ + file = add (cp, NULL); + } else { + app_msgarg (&msgs, cp); + } + } } SIGNAL (SIGQUIT, quitser); @@ -254,16 +259,6 @@ main (int argc, char **argv) { fclose (fp); } - /* - * Check for storage directory. If specified, - * then store temporary files there. Else we - * store them in standard nmh directory. - */ - if ((cp = context_find (nmhstorage)) && *cp) - tmp = concat (cp, "/", invo_name, NULL); - else - tmp = add (m_maildir (invo_name), NULL); - suppress_bogus_mp_content_warning = skip_mp_cte_check = 1; if (! context_find ("path")) @@ -288,17 +283,17 @@ main (int argc, char **argv) { using_stdin = 1; - if ((cp = m_mktemp2 (tmp, invo_name, &fd, NULL)) == NULL) { - adios (NULL, "unable to create temporary file"); + if ((cp = m_mktemp2 (NULL, invo_name, &fd, NULL)) == NULL) { + adios (NULL, "unable to create temporary file in %s", + get_temp_dir()); } else { free (file); file = add (cp, NULL); - chmod (file, 0600); cpydata (STDIN_FILENO, fd, "-", file); } if (close (fd)) { - unlink (file); + (void) m_unlink (file); adios (NULL, "failed to write temporary file"); } } @@ -361,7 +356,7 @@ main (int argc, char **argv) { status += mhfixmsgsbr (ctp, &fx, outfile); if (using_stdin) { - unlink (file); + (void) m_unlink (file); if (! outfile) { /* Just calling m_backup() unlinks the backup file. */ @@ -374,7 +369,6 @@ main (int argc, char **argv) { } free (outfile); - free (tmp); free (file); /* done is freects_done, which will clean up all of cts. */ @@ -396,7 +390,12 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) { modify_inplace = 1; if ((*ctp)->c_file) { - outfile = add (m_mktemp2 (tmp, invo_name, NULL, NULL), NULL); + char *tempfile; + if ((tempfile = m_mktemp2 (NULL, invo_name, NULL, NULL)) == NULL) { + adios (NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + outfile = add (tempfile, NULL); } else { adios (NULL, "missing both input and output filenames\n"); } @@ -410,7 +409,8 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) { status = fix_multipart_cte (*ctp, &message_mods); } if (status == OK && fx->reformat) { - status = ensure_text_plain (ctp, NULL, &message_mods); + status = + ensure_text_plain (ctp, NULL, &message_mods, fx->replacetextplain); } if (status == OK && fx->decodetext) { status = decode_text_parts (*ctp, fx->decodetext, &message_mods); @@ -457,7 +457,7 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) { } if (modify_inplace) { - if (status != OK) unlink (outfile); + if (status != OK) (void) m_unlink (outfile); free (outfile); outfile = NULL; } @@ -513,7 +513,7 @@ fix_boundary (CT *ct, int *message_mods) { if (get_multipart_boundary (*ct, &part_boundary) == OK) { char *fixed; - if ((fixed = m_mktemp2 (tmp, invo_name, NULL, &(*ct)->c_fp))) { + if ((fixed = m_mktemp2 (NULL, invo_name, NULL, &(*ct)->c_fp))) { if (replace_boundary (*ct, fixed, part_boundary) == OK) { char *filename = add ((*ct)->c_file, NULL); @@ -533,7 +533,8 @@ fix_boundary (CT *ct, int *message_mods) { status = NOTOK; } } else { - advise (NULL, "unable to create temporary file"); + advise (NULL, "unable to create temporary file in %s", + get_temp_dir()); status = NOTOK; } @@ -873,7 +874,7 @@ set_ce (CT ct, int encoding) { /* Make sure each text part has a corresponding text/plain part. */ static int -ensure_text_plain (CT *ct, CT parent, int *message_mods) { +ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain) { int status = OK; switch ((*ct)->c_type) { @@ -886,17 +887,36 @@ ensure_text_plain (CT *ct, CT parent, int *message_mods) { if (parent && parent->c_type == CT_MULTIPART && parent->c_subtype == MULTI_ALTERNATE) { struct multipart *mp = (struct multipart *) parent->c_ctparams; - struct part *part; + struct part *part, *prev; int new_subpart_number = 1; /* See if there is a sibling text/plain. */ - for (part = mp->mp_parts; part; part = part->mp_next) { + for (prev = part = mp->mp_parts; part; part = part->mp_next) { ++new_subpart_number; if (part->mp_part->c_type == CT_TEXT && part->mp_part->c_subtype == TEXT_PLAIN) { - has_text_plain = 1; + if (replacetextplain) { + struct part *old_part; + if (part == mp->mp_parts) { + old_part = mp->mp_parts; + mp->mp_parts = part->mp_next; + } else { + old_part = prev->mp_next; + prev->mp_next = part->mp_next; + } + if (verbosw) { + report (parent->c_partno, parent->c_file, + "remove text/plain part %s", + old_part->mp_part->c_partno); + } + free_content (old_part->mp_part); + free (old_part); + } else { + has_text_plain = 1; + } break; } + prev = part; } if (! has_text_plain) { @@ -967,7 +987,8 @@ ensure_text_plain (CT *ct, CT parent, int *message_mods) { for (part = mp->mp_parts; status == OK && part; part = part->mp_next) { if ((*ct)->c_type == CT_MULTIPART) { - status = ensure_text_plain (&part->mp_part, *ct, message_mods); + status = ensure_text_plain (&part->mp_part, *ct, message_mods, + replacetextplain); } } break; @@ -978,7 +999,8 @@ ensure_text_plain (CT *ct, CT parent, int *message_mods) { struct exbody *e; e = (struct exbody *) (*ct)->c_ctparams; - status = ensure_text_plain (&e->eb_content, *ct, message_mods); + status = ensure_text_plain (&e->eb_content, *ct, message_mods, + replacetextplain); } break; } @@ -997,7 +1019,13 @@ build_text_plain_part (CT encoded_part) { contains the decoded contents. And the decoding function, such as openQuoted, will have set ...->ce_unlink to 1 so that it will be unlinked by free_content (). */ - tmp_plain_file = add (m_mktemp2 (tmp, invo_name, NULL, NULL), NULL); + char *tempfile; + + if ((tempfile = m_mktemp2 (NULL, invo_name, NULL, NULL)) == NULL) { + advise (NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + tmp_plain_file = add (tempfile, NULL); if (reformat_part (tp_part, tmp_plain_file, tp_part->c_ctinfo.ci_type, tp_part->c_ctinfo.ci_subtype, @@ -1007,7 +1035,7 @@ build_text_plain_part (CT encoded_part) { } free_content (tp_part); - unlink (tmp_plain_file); + (void) m_unlink (tmp_plain_file); free (tmp_plain_file); return NULL; @@ -1045,19 +1073,19 @@ divide_part (CT ct) { static void copy_ctinfo (CI dest, CI src) { - char **s_ap, **d_ap, **s_vp, **d_vp; + PM s_pm, d_pm; dest->ci_type = src->ci_type ? add (src->ci_type, NULL) : NULL; dest->ci_subtype = src->ci_subtype ? add (src->ci_subtype, NULL) : NULL; - for (s_ap = src->ci_attrs, d_ap = dest->ci_attrs, - s_vp = src->ci_values, d_vp = dest->ci_values; - *s_ap; - ++s_ap, ++d_ap, ++s_vp, ++d_vp) { - *d_ap = add (*s_ap, NULL); - *d_vp = *s_vp; + for (s_pm = src->ci_first_pm; s_pm; s_pm = s_pm->pm_next) { + d_pm = add_param(&dest->ci_first_pm, &dest->ci_last_pm, s_pm->pm_name, + s_pm->pm_value); + if (s_pm->pm_charset) + d_pm->pm_charset = getcpy(s_pm->pm_charset); + if (s_pm->pm_lang) + d_pm->pm_lang = getcpy(s_pm->pm_lang); } - *d_ap = NULL; dest->ci_comment = src->ci_comment ? add (src->ci_comment, NULL) : NULL; dest->ci_magic = src->ci_magic ? add (src->ci_magic, NULL) : NULL; @@ -1068,13 +1096,17 @@ static int decode_part (CT ct) { char *tmp_decoded; int status; + char *tempfile; - tmp_decoded = add (m_mktemp2 (tmp, invo_name, NULL, NULL), NULL); + if ((tempfile = m_mktemp2 (NULL, invo_name, NULL, NULL)) == NULL) { + adios (NULL, "unable to create temporary file in %s", get_temp_dir()); + } + tmp_decoded = add (tempfile, NULL); /* The following call will load ct->c_cefile.ce_file with the tmp filename of the decoded content. tmp_decoded will contain the encoded output, get rid of that. */ status = output_message (ct, tmp_decoded); - unlink (tmp_decoded); + (void) m_unlink (tmp_decoded); free (tmp_decoded); return status; @@ -1134,7 +1166,7 @@ reformat_part (CT ct, char *file, char *type, char *subtype, int c_type) { /* Unlink decoded content tmp file and free its filename to avoid leaks. The file stream should already have been closed. */ if (ct->c_cefile.ce_unlink) { - unlink (ct->c_cefile.ce_file); + (void) m_unlink (ct->c_cefile.ce_file); free (ct->c_cefile.ce_file); ct->c_cefile.ce_file = NULL; ct->c_cefile.ce_unlink = 0; @@ -1182,7 +1214,6 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { CT ct; struct part *p; struct multipart *m; - char *cp; const struct str2init *ctinit; if ((ct = (CT) calloc (1, sizeof *ct)) == NULL) @@ -1282,14 +1313,8 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { ct->c_ctinfo.ci_subtype = add (subtypename, NULL); } - name = concat (" ", typename, "/", subtypename, boundary_indicator, - boundary, NULL); - if ((cp = strstr (name, boundary_indicator))) { - ct->c_ctinfo.ci_attrs[0] = name; - ct->c_ctinfo.ci_attrs[1] = NULL; - /* ci_values don't get free'd, so point into ci_attrs. */ - ct->c_ctinfo.ci_values[0] = cp + strlen (boundary_indicator); - } + add_param(&ct->c_ctinfo.ci_first_pm, &ct->c_ctinfo.ci_last_pm, + "boundary", boundary); p = (struct part *) mh_xmalloc (sizeof *p); p->mp_next = (struct part *) mh_xmalloc (sizeof *p->mp_next); @@ -1454,7 +1479,7 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { : ct->c_ctline ? ct->c_ctline : ""); } - unlink (ct->c_cefile.ce_file); + (void) m_unlink (ct->c_cefile.ce_file); free (ct->c_cefile.ce_file); ct->c_cefile.ce_file = NULL; } else if (ct->c_encoding == CE_QUOTED && @@ -1468,7 +1493,7 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { : ct->c_ctline ? ct->c_ctline : ""); } - unlink (ct->c_cefile.ce_file); + (void) m_unlink (ct->c_cefile.ce_file); free (ct->c_cefile.ce_file); ct->c_cefile.ce_file = NULL; } else { @@ -1485,7 +1510,7 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { report (ct->c_partno, ct->c_file, "decode%s", ct->c_ctline ? ct->c_ctline : ""); } - strip_crs (ct); + strip_crs (ct, message_mods); } else { status = NOTOK; } @@ -1497,7 +1522,7 @@ decode_text_parts (CT ct, int encoding, int *message_mods) { } case CE_8BIT: case CE_7BIT: - strip_crs (ct); + strip_crs (ct, message_mods); break; default: break; @@ -1543,6 +1568,7 @@ content_encoding (CT ct) { int encoding = CE_7BIT; if (ce->ce_file) { + size_t line_len = 0; char buffer[BUFSIZ]; size_t inbytes; @@ -1556,7 +1582,6 @@ content_encoding (CT ct) { (inbytes = fread (buffer, 1, sizeof buffer, ce->ce_fp)) > 0) { char *cp; size_t i; - size_t line_len = 0; int last_char_was_cr = 0; for (i = 0, cp = buffer; i < inbytes; ++i, ++cp) { @@ -1583,13 +1608,20 @@ content_encoding (CT ct) { static int -strip_crs (CT ct) { +strip_crs (CT ct, int *message_mods) { /* norm_charmap() is case sensitive. */ char *codeset = upcase (content_codeset (ct)); int status = OK; - /* Only strip carriage returns if content is ASCII. */ - if (! strcmp (norm_charmap (codeset), "US-ASCII")) { + /* Only strip carriage returns if content is ASCII or another + codeset that has the same readily recognizable CR followed by a + LF. We can include UTF-8 here because if the high-order bit of + a UTF-8 byte is 0, then it must be a single-byte ASCII + character. */ + if (! strcmp (norm_charmap (codeset), "US-ASCII") || + ! strncmp (norm_charmap (codeset), "ISO-8859-", 9) || + ! strncmp (norm_charmap (codeset), "UTF-8", 5) || + ! strncmp (norm_charmap (codeset), "WINDOWS-12", 10)) { char **file = NULL; FILE **fp = NULL; size_t begin; @@ -1650,8 +1682,14 @@ strip_crs (CT ct) { if (has_crs) { int fd; - char *stripped_content_file = - add (m_mktemp2 (tmp, invo_name, &fd, NULL), NULL); + char *stripped_content_file; + char *tempfile = m_mktemp2 (NULL, invo_name, &fd, NULL); + + if (tempfile == NULL) { + adios (NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + stripped_content_file = add (tempfile, NULL); /* Strip each CR before a LF from the content. */ fseeko (*fp, begin, SEEK_SET); @@ -1672,25 +1710,31 @@ strip_crs (CT ct) { write (fd, cp, 1); last_char_was_cr = 0; } - } } if (close (fd)) { - admonish (NULL, "unable to write temporaty file %s", + admonish (NULL, "unable to write temporary file %s", stripped_content_file); - unlink (stripped_content_file); + (void) m_unlink (stripped_content_file); status = NOTOK; } else { /* Replace the decoded file with the converted one. */ if (ct->c_cefile.ce_file) { if (ct->c_cefile.ce_unlink) { - unlink (ct->c_cefile.ce_file); + (void) m_unlink (ct->c_cefile.ce_file); } free (ct->c_cefile.ce_file); } ct->c_cefile.ce_file = stripped_content_file; ct->c_cefile.ce_unlink = 1; + + ++*message_mods; + if (verbosw) { + report (ct->c_partno, + begin == 0 && end == 0 ? "" : *file, + "stripped CRs"); + } } } @@ -1711,19 +1755,22 @@ content_codeset (CT ct) { const char *const charset = "charset"; char *default_codeset = NULL; CI ctinfo = &ct->c_ctinfo; - char **ap, **vp; char **src_codeset = NULL; + PM pm; - for (ap = ctinfo->ci_attrs, vp = ctinfo->ci_values; *ap; ++ap, ++vp) { - if (! strcasecmp (*ap, charset)) { - src_codeset = vp; + for (pm = ctinfo->ci_first_pm; pm; pm = pm->pm_next) { + if (! strcasecmp (pm->pm_name, charset)) { + src_codeset = &pm->pm_value; break; } } /* RFC 2045, Sec. 5.2: default to us-ascii. */ - if (src_codeset == NULL) src_codeset = &default_codeset; - if (*src_codeset == NULL) *src_codeset = "US-ASCII"; + if (src_codeset == NULL) { + src_codeset = &default_codeset; + if (*src_codeset == NULL) + *src_codeset = "US-ASCII"; + } return *src_codeset; } @@ -1797,6 +1844,7 @@ convert_codeset (CT ct, char *dest_codeset, int *message_mods) { int opened_input_file = 0; char src_buffer[BUFSIZ]; HF hf; + char *tempfile; if ((conv_desc = iconv_open (dest_codeset, src_codeset)) == (iconv_t) -1) { @@ -1804,7 +1852,11 @@ convert_codeset (CT ct, char *dest_codeset, int *message_mods) { return -1; } - dest = add (m_mktemp2 (tmp, invo_name, &fd, NULL), NULL); + if ((tempfile = m_mktemp2 (NULL, invo_name, &fd, NULL)) == NULL) { + adios (NULL, "unable to create temporary file in %s", + get_temp_dir()); + } + dest = add (tempfile, NULL); if (ct->c_cefile.ce_file) { file = &ct->c_cefile.ce_file; @@ -1838,7 +1890,8 @@ convert_codeset (CT ct, char *dest_codeset, int *message_mods) { min (bytes_to_read, sizeof src_buffer), *fp)) > 0) { char dest_buffer[BUFSIZ]; - char *ib = src_buffer, *ob = dest_buffer; + ICONV_CONST char *ib = src_buffer; + char *ob = dest_buffer; size_t outbytes = sizeof dest_buffer; size_t outbytes_before = outbytes; @@ -1866,7 +1919,7 @@ convert_codeset (CT ct, char *dest_codeset, int *message_mods) { /* Replace the decoded file with the converted one. */ if (ct->c_cefile.ce_file) { if (ct->c_cefile.ce_unlink) { - unlink (ct->c_cefile.ce_file); + (void) m_unlink (ct->c_cefile.ce_file); } free (ct->c_cefile.ce_file); } @@ -1905,7 +1958,7 @@ convert_codeset (CT ct, char *dest_codeset, int *message_mods) { } } } else { - unlink (dest); + (void) m_unlink (dest); } #else /* ! HAVE_ICONV */ NMH_UNUSED (message_mods); @@ -1955,7 +2008,7 @@ write_content (CT ct, char *input_filename, char *outfile, int modify_inplace, } if (new != -1) close (new); if (old != -1) close (old); - unlink (outfile); + (void) m_unlink (outfile); if (i < 0) { /* The -file argument processing used path() to @@ -1971,7 +2024,7 @@ write_content (CT ct, char *input_filename, char *outfile, int modify_inplace, } else { admonish (NULL, "unable to remove input file %s, " "not modifying it", infile); - unlink (outfile); + (void) m_unlink (outfile); status = NOTOK; } @@ -1981,7 +2034,7 @@ write_content (CT ct, char *input_filename, char *outfile, int modify_inplace, } } else { /* No modifications and didn't need the tmp outfile. */ - unlink (outfile); + (void) m_unlink (outfile); } } else { /* Output is going to some file. Produce it whether or not