X-Git-Url: https://diplodocus.org/git/nmh/blobdiff_plain/f2753d64be0a74a22018e03ec9c04cc9ca45d3ab..ddf3a8574f657dcb8c53fc5908e7bebbde1994b5:/uip/mhfixmsg.c diff --git a/uip/mhfixmsg.c b/uip/mhfixmsg.c index 9d676cbe..f410ecb4 100644 --- a/uip/mhfixmsg.c +++ b/uip/mhfixmsg.c @@ -5,19 +5,53 @@ * distribution for complete copyright information. */ -#include -#include -#include -#include -#include -#include "../sbr/m_mktemp.h" +#include "h/mh.h" +#include "sbr/m_name.h" +#include "sbr/m_gmprot.h" +#include "sbr/m_getfld.h" +#include "sbr/getarguments.h" +#include "sbr/concat.h" +#include "sbr/seq_setprev.h" +#include "sbr/seq_setcur.h" +#include "sbr/seq_save.h" +#include "sbr/smatch.h" +#include "sbr/fmt_rfc2047.h" +#include "sbr/cpydata.h" +#include "sbr/trimcpy.h" +#include "sbr/m_convert.h" +#include "sbr/m_backup.h" +#include "sbr/getfolder.h" +#include "sbr/folder_read.h" +#include "sbr/context_save.h" +#include "sbr/context_replace.h" +#include "sbr/context_find.h" +#include "sbr/readconfig.h" +#include "sbr/ambigsw.h" +#include "sbr/path.h" +#include "sbr/print_version.h" +#include "sbr/print_help.h" +#include "sbr/error.h" +#include "h/fmt_scan.h" +#include "h/mime.h" +#include "h/mhparse.h" +#include "h/done.h" +#include "h/utils.h" +#include "h/signals.h" +#include "sbr/m_maildir.h" +#include "sbr/m_mktemp.h" +#include "sbr/mime_type.h" +#include "mhmisc.h" #include "mhfree.h" +#include "mhoutsbr.h" +#include "mhshowsbr.h" #include #define MHFIXMSG_SWITCHES \ X("decodetext 8bit|7bit|binary", 0, DECODETEXTSW) \ X("nodecodetext", 0, NDECODETEXTSW) \ X("decodetypes", 0, DECODETYPESW) \ + X("decodeheaderfieldbodies utf-8", 0, DECODEHEADERFIELDBODIESSW) \ + X("nodecodeheaderfieldbodies", 0, NDECODEHEADERFIELDBODIESSW) \ X("crlflinebreaks", 0, CRLFLINEBREAKSSW) \ X("nocrlflinebreaks", 0, NCRLFLINEBREAKSSW) \ X("textcharset", 0, TEXTCHARSETSW) \ @@ -30,6 +64,8 @@ X("nofixboundary", 0, NFIXBOUNDARYSW) \ X("fixcte", 0, FIXCOMPOSITECTESW) \ X("nofixcte", 0, NFIXCOMPOSITECTESW) \ + X("checkbase64", 0, CHECKBASE64SW) \ + X("nocheckbase64", 0, NCHECKBASE64SW) \ X("fixtype mimetype", 0, FIXTYPESW) \ X("file file", 0, FILESW) \ X("outfile file", 0, OUTFILESW) \ @@ -56,23 +92,10 @@ int debugsw; /* Needed by mhparse.c. */ #define quitser pipeser -/* mhparse.c */ -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 */ -/* flags to/from parse_header_attrs */ -extern int suppress_extraneous_trailing_semicolon_warning; - -/* mhoutsbr.c */ -int output_message_fp (CT, FILE *, char *); - -/* mhmisc.c */ -void flush_errors (void); - /* * static prototypes */ -typedef struct fix_transformations { +typedef struct { int fixboundary; int fixcompositecte; svector_t fixtypes; @@ -80,13 +103,15 @@ typedef struct fix_transformations { int replacetextplain; int decodetext; char *decodetypes; + char *decodeheaderfieldbodies; /* Either NULL or "utf-8". */ /* Whether to use CRLF linebreaks, per RFC 2046 Sec. 4.1.1, par.1. */ int lf_line_endings; char *textcharset; + bool checkbase64; } fix_transformations; -int mhfixmsgsbr (CT *, char *, const fix_transformations *, FILE **, char *, - FILE **); +static int mhfixmsgsbr (CT *, char *, const fix_transformations *, + FILE **, char *, FILE **); static int fix_boundary (CT *, int *); static int copy_input_to_output (const char *, FILE *, const char *, FILE *); static int get_multipart_boundary (CT, char **); @@ -101,11 +126,15 @@ static int find_textplain_sibling (CT, int, int *); static int insert_new_text_plain_part (CT, int, CT); static CT build_text_plain_part (CT); static int insert_into_new_mp_alt (CT *, int *); +static int insert_into_new_mp_mixed (CT *, const char *, int *); static CT divide_part (CT); static void copy_ctinfo (CI, CI); static int decode_part (CT); +static size_t get_valid_base64 (CT, char **); +static size_t find_invalid_base64_pos (const char *); +static int check_base64_encoding (CT *); static int reformat_part (CT, char *, char *, char *, int); -static CT build_multipart_alt (CT, CT, int, int); +static CT build_multipart (CT, CT, int, int); static int boundary_in_content (FILE **, char *, const char *); static void transfer_noncontent_headers (CT, CT); static int set_ct_type (CT, int type, int subtype, int encoding); @@ -114,21 +143,24 @@ static int should_decode(const char *, const char *, const char *); static int content_encoding (CT, const char **); static int strip_crs (CT, int *); static void update_cte (CT); -static int least_restrictive_encoding (CT); +static int least_restrictive_encoding (CT) PURE; static int less_restrictive (int, int); static int convert_charsets (CT, char *, int *); -static int fix_always (CT, int *); +static int fix_always (CT *, const fix_transformations *, int *); +static int decode_header_field_bodies (CT, int *); static int fix_filename_param (char *, char *, PM *, PM *); static int fix_filename_encoding (CT); static int write_content (CT, const char *, char *, FILE *, int, int); static void set_text_ctparams(CT, char *, int); static int remove_file (const char *); -static void report (char *, char *, char *, char *, ...); +static void report (char *, char *, char *, char *, ...) + CHECK_PRINTF(4, 5); static void pipeser (int); int -main (int argc, char **argv) { +main (int argc, char **argv) +{ int msgnum; char *cp, *file = NULL, *folder = NULL; char *maildir = NULL, buf[100], *outfile = NULL; @@ -137,8 +169,8 @@ main (int argc, char **argv) { struct msgs *mp = NULL; CT *ctp; FILE *fp, *infp = NULL, *outfp = NULL; - int using_stdin = 0; - int chgflag = 1; + bool using_stdin = false; + bool chgflag = true; int status = OK; fix_transformations fx; fx.reformat = fx.fixcompositecte = fx.fixboundary = 1; @@ -146,10 +178,12 @@ main (int argc, char **argv) { fx.replacetextplain = 0; fx.decodetext = CE_8BIT; fx.decodetypes = "text,application/ics"; /* Default, per man page. */ + fx.decodeheaderfieldbodies = NULL; fx.lf_line_endings = 0; fx.textcharset = NULL; + fx.checkbase64 = true; - if (nmh_init(argv[0], 2)) { return 1; } + if (nmh_init(argv[0], true, false)) { return 1; } arguments = getarguments (invo_name, argc, argv, 1); argp = arguments; @@ -164,7 +198,7 @@ main (int argc, char **argv) { ambigsw (cp, switches); done (1); case UNKWNSW: - adios (NULL, "-%s unknown", cp); + die("-%s unknown", cp); case HELPSW: snprintf (buf, sizeof buf, "%s [+folder] [msgs] [switches]", @@ -177,7 +211,7 @@ main (int argc, char **argv) { case DECODETEXTSW: if (! (cp = *argp++) || *cp == '-') { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } if (! strcasecmp (cp, "8bit")) { fx.decodetext = CE_8BIT; @@ -186,7 +220,7 @@ main (int argc, char **argv) { } else if (! strcasecmp (cp, "binary")) { fx.decodetext = CE_BINARY; } else { - adios (NULL, "invalid argument to %s", argp[-2]); + die("invalid argument to %s", argp[-2]); } continue; case NDECODETEXTSW: @@ -194,10 +228,25 @@ main (int argc, char **argv) { continue; case DECODETYPESW: if (! (cp = *argp++) || *cp == '-') { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } fx.decodetypes = cp; continue; + case DECODEHEADERFIELDBODIESSW: + if (! (cp = *argp++) || *cp == '-') { + die("missing argument to %s", argp[-2]); + } + fx.decodeheaderfieldbodies = cp; + if (strcasecmp (cp, "utf-8") && strcasecmp (cp, "utf8")) { + /* Because UTF-8 strings can't have embedded nulls. Other + encodings support that, too, but we won't bother to + enumerate them. */ + die("-decodeheaderfieldbodies only supports utf-8"); + } + continue; + case NDECODEHEADERFIELDBODIESSW: + fx.decodeheaderfieldbodies = NULL; + continue; case CRLFLINEBREAKSSW: fx.lf_line_endings = 0; continue; @@ -206,13 +255,19 @@ main (int argc, char **argv) { continue; case TEXTCHARSETSW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } fx.textcharset = cp; continue; case NTEXTCHARSETSW: fx.textcharset = 0; continue; + case CHECKBASE64SW: + fx.checkbase64 = true; + continue; + case NCHECKBASE64SW: + fx.checkbase64 = false; + continue; case FIXBOUNDARYSW: fx.fixboundary = 1; continue; @@ -227,14 +282,13 @@ main (int argc, char **argv) { continue; case FIXTYPESW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } if (! strncasecmp (cp, "multipart/", 10) || - ! strncasecmp (cp, "message/", 8)) { - adios (NULL, "-fixtype %s not allowed", cp); - } else if (! strchr (cp, '/')) { - adios (NULL, "-fixtype requires type/subtype"); - } + ! strncasecmp (cp, "message/", 8)) + die("-fixtype %s not allowed", cp); + if (! strchr (cp, '/')) + die("-fixtype requires type/subtype"); if (fx.fixtypes == NULL) { fx.fixtypes = svector_create (10); } svector_push_back (fx.fixtypes, cp); continue; @@ -252,29 +306,29 @@ main (int argc, char **argv) { continue; case FILESW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } file = *cp == '-' ? mh_xstrdup (cp) : path (cp, TFILE); continue; case OUTFILESW: if (! (cp = *argp++) || (*cp == '-' && cp[1])) { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } outfile = *cp == '-' ? mh_xstrdup (cp) : path (cp, TFILE); continue; case RPROCSW: if (!(rmmproc = *argp++) || *rmmproc == '-') { - adios (NULL, "missing argument to %s", argp[-2]); + die("missing argument to %s", argp[-2]); } continue; case NRPRCSW: rmmproc = NULL; continue; case CHGSW: - chgflag = 1; + chgflag = true; continue; case NCHGSW: - chgflag = 0; + chgflag = false; continue; case VERBSW: verbosw = 1; @@ -285,11 +339,9 @@ main (int argc, char **argv) { } } if (*cp == '+' || *cp == '@') { - if (folder) { - adios (NULL, "only one folder at a time!"); - } else { - folder = pluspath (cp); - } + if (folder) + die("only one folder at a time!"); + folder = pluspath (cp); } else { if (*cp == '/') { /* Interpret a full path as a filename, not a message. */ @@ -307,19 +359,19 @@ main (int argc, char **argv) { * Read the standard profile setup */ if ((fp = fopen (cp = etcpath ("mhn.defaults"), "r"))) { - readconfig ((struct node **) 0, fp, cp, 0); + readconfig(NULL, fp, cp, 0); fclose (fp); } - suppress_bogus_mp_content_warning = skip_mp_cte_check = 1; - suppress_extraneous_trailing_semicolon_warning = 1; + suppress_bogus_mp_content_warning = skip_mp_cte_check = true; + suppress_extraneous_trailing_semicolon_warning = true; if (! context_find ("path")) { free (path ("./", TFOLDER)); } if (file && msgs.size) { - adios (NULL, "cannot specify msg and file at same time!"); + die("cannot specify msg and file at same time!"); } if (outfile) { @@ -346,10 +398,10 @@ main (int argc, char **argv) { int fd; char *cp; - using_stdin = 1; + using_stdin = true; if ((cp = m_mktemp2 (NULL, invo_name, &fd, NULL)) == NULL) { - adios (NULL, "unable to create temporary file in %s", + die("unable to create temporary file in %s", get_temp_dir()); } else { free (file); @@ -359,7 +411,7 @@ main (int argc, char **argv) { if (close (fd)) { (void) m_unlink (file); - adios (NULL, "failed to write temporary file"); + die("failed to write temporary file"); } } @@ -414,12 +466,12 @@ main (int argc, char **argv) { /* read folder and create message structure */ if (! (mp = folder_read (folder, 1))) { - adios (NULL, "unable to read folder %s", folder); + die("unable to read folder %s", folder); } /* check for empty folder */ if (mp->nummsg == 0) { - adios (NULL, "no messages in %s", folder); + die("no messages in %s", folder); } /* parse all the message ranges/sequences and set SELECTED */ @@ -485,7 +537,10 @@ main (int argc, char **argv) { if (*cts) { for (ctp = cts; *ctp; ++ctp) { - status += mhfixmsgsbr (ctp, maildir, &fx, &infp, outfile, &outfp); + status = + mhfixmsgsbr (ctp, maildir, &fx, &infp, outfile, &outfp) == OK + ? 0 + : 1; free_content (*ctp); if (using_stdin) { @@ -501,7 +556,7 @@ main (int argc, char **argv) { status = 1; } - mh_xfree(maildir); + free(maildir); free (cts); if (fx.fixtypes != NULL) { svector_free (fx.fixtypes); } @@ -512,7 +567,7 @@ main (int argc, char **argv) { free (folder); free (arguments); - done (status); + done (status == OK ? 0 : 1); return NOTOK; } @@ -520,15 +575,16 @@ main (int argc, char **argv) { /* * Apply transformations to one message. */ -int +static int mhfixmsgsbr (CT *ctp, char *maildir, const fix_transformations *fx, - FILE **infp, char *outfile, FILE **outfp) { + FILE **infp, char *outfile, FILE **outfp) +{ /* Store input filename in case one of the transformations, i.e., fix_boundary(), rewrites to a tmp file. */ char *input_filename = maildir ? concat (maildir, "/", (*ctp)->c_file, NULL) : mh_xstrdup ((*ctp)->c_file); - int modify_inplace = 0; + bool modify_inplace = false; int message_mods = 0; int status = OK; @@ -540,24 +596,24 @@ mhfixmsgsbr (CT *ctp, char *maildir, const fix_transformations *fx, } if (outfile == NULL) { - modify_inplace = 1; + modify_inplace = true; if ((*ctp)->c_file) { char *tempfile; /* outfp will be closed by the caller */ if ((tempfile = m_mktemp2 (NULL, invo_name, NULL, outfp)) == NULL) { - adios (NULL, "unable to create temporary file in %s", + die("unable to create temporary file in %s", get_temp_dir()); } outfile = mh_xstrdup (tempfile); } else { - adios (NULL, "missing both input and output filenames\n"); + die("missing both input and output filenames\n"); } } /* else *outfp was defined by caller */ reverse_alternative_parts (*ctp); - status = fix_always (*ctp, &message_mods); + status = fix_always (ctp, fx, &message_mods); if (status == OK && fx->fixboundary) { status = fix_boundary (ctp, &message_mods); } @@ -576,6 +632,9 @@ mhfixmsgsbr (CT *ctp, char *maildir, const fix_transformations *fx, &message_mods); update_cte (*ctp); } + if (status == OK && fx->decodeheaderfieldbodies) { + status = decode_header_field_bodies(*ctp, &message_mods); + } if (status == OK && fx->textcharset != NULL) { status = convert_charsets (*ctp, fx->textcharset, &message_mods); } @@ -629,7 +688,8 @@ mhfixmsgsbr (CT *ctp, char *maildir, const fix_transformations *fx, */ static int copy_input_to_output (const char *input_filename, FILE *infp, - const char *output_filename, FILE *outfp) { + const char *output_filename, FILE *outfp) +{ int in = fileno (infp); int out = fileno (outfp); int status = OK; @@ -648,7 +708,8 @@ copy_input_to_output (const char *input_filename, FILE *infp, * Fix mismatched outer level boundary. */ static int -fix_boundary (CT *ct, int *message_mods) { +fix_boundary (CT *ct, int *message_mods) +{ struct multipart *mp; int status = OK; @@ -724,7 +785,8 @@ fix_boundary (CT *ct, int *message_mods) { * Find boundary at end of multipart. */ static int -get_multipart_boundary (CT ct, char **part_boundary) { +get_multipart_boundary (CT ct, char **part_boundary) +{ char buffer[NMH_BUFSIZ]; char *end_boundary = NULL; off_t begin = (off_t) ct->c_end > (off_t) (ct->c_begin + sizeof buffer) @@ -811,12 +873,13 @@ get_multipart_boundary (CT ct, char **part_boundary) { * Open and copy ct->c_file to file, replacing the multipart boundary. */ static int -replace_boundary (CT ct, char *file, char *boundary) { +replace_boundary (CT ct, char *file, char *boundary) +{ FILE *fpin, *fpout; int compnum, state; char buf[NMH_BUFSIZ], name[NAMESZ]; char *np, *vp; - m_getfld_state_t gstate = 0; + m_getfld_state_t gstate; int status = OK; if (ct->c_file == NULL) { @@ -835,10 +898,11 @@ replace_boundary (CT ct, char *file, char *boundary) { return NOTOK; } + gstate = m_getfld_state_init(fpin); for (compnum = 1;;) { int bufsz = (int) sizeof buf; - switch (state = m_getfld (&gstate, name, buf, &bufsz, fpin)) { + switch (state = m_getfld2(&gstate, name, buf, &bufsz)) { case FLD: case FLDPLUS: compnum++; @@ -850,28 +914,28 @@ replace_boundary (CT ct, char *file, char *boundary) { /* if necessary, get rest of field */ while (state == FLDPLUS) { bufsz = sizeof buf; - state = m_getfld (&gstate, name, buf, &bufsz, fpin); + state = m_getfld2(&gstate, name, buf, &bufsz); vp = add (buf, vp); /* add to previous value */ } if (strcasecmp (TYPE_FIELD, np)) { fprintf (fpout, "%s:%s", np, vp); } else { - char *new_ctline, *new_params; + char *new_ctline, *new_params; - replace_param(&ct->c_ctinfo.ci_first_pm, - &ct->c_ctinfo.ci_last_pm, "boundary", - boundary, 0); + replace_param(&ct->c_ctinfo.ci_first_pm, + &ct->c_ctinfo.ci_last_pm, "boundary", + boundary, 0); - new_ctline = concat(" ", ct->c_ctinfo.ci_type, "/", - ct->c_ctinfo.ci_subtype, NULL); - new_params = output_params(strlen(TYPE_FIELD) + - strlen(new_ctline) + 1, - ct->c_ctinfo.ci_first_pm, NULL, 0); + new_ctline = concat(" ", ct->c_ctinfo.ci_type, "/", + ct->c_ctinfo.ci_subtype, NULL); + new_params = output_params(LEN(TYPE_FIELD) + + strlen(new_ctline) + 1, + ct->c_ctinfo.ci_first_pm, NULL, 0); fprintf (fpout, "%s:%s%s\n", np, new_ctline, - FENDNULL(new_params)); - free(new_ctline); - mh_xfree(new_params); + FENDNULL(new_params)); + free(new_ctline); + free(new_params); } free (vp); @@ -917,7 +981,8 @@ replace_boundary (CT ct, char *file, char *boundary) { * Fix Content-Type header to reflect the content of its part. */ static int -fix_types (CT ct, svector_t fixtypes, int *message_mods) { +fix_types (CT ct, svector_t fixtypes, int *message_mods) +{ int status = OK; switch (ct->c_type) { @@ -1031,7 +1096,8 @@ fix_types (CT ct, svector_t fixtypes, int *message_mods) { * Replace a substring, allocating space to hold the new one. */ char * -replace_substring (char **str, const char *old, const char *new) { +replace_substring (char **str, const char *old, const char *new) +{ char *cp; if ((cp = strstr (*str, old))) { @@ -1060,7 +1126,8 @@ replace_substring (char **str, const char *old, const char *new) { * Remove a name=value parameter, given just its name, from a header value. */ char * -remove_parameter (char *str, const char *name) { +remove_parameter (char *str, const char *name) +{ /* It looks to me, based on the BNF in RFC 2045, than there can't be whitespace between the parameter name and the "=", or between the "=" and the parameter value. */ @@ -1109,7 +1176,8 @@ remove_parameter (char *str, const char *name) { * 8 bit. */ static int -fix_composite_cte (CT ct, int *message_mods) { +fix_composite_cte (CT ct, int *message_mods) +{ int status = OK; if (ct->c_type == CT_MESSAGE || ct->c_type == CT_MULTIPART) { @@ -1119,18 +1187,17 @@ fix_composite_cte (CT ct, int *message_mods) { for (hf = ct->c_first_hf; hf; hf = hf->next) { char *name = hf->name; - for (; *name && isspace ((unsigned char) *name); ++name) { + for (; isspace((unsigned char)*name); ++name) { continue; } if (! strncasecmp (name, ENCODING_FIELD, - strlen (ENCODING_FIELD))) { + LEN(ENCODING_FIELD))) { char *prefix = "Nmh-REPLACED-INVALID-"; HF h; NEW(h); h->name = mh_xstrdup (hf->name); - h->hf_encoding = hf->hf_encoding; h->next = hf->next; hf->next = h; @@ -1181,13 +1248,14 @@ fix_composite_cte (CT ct, int *message_mods) { * Set content encoding. */ static int -set_ce (CT ct, int encoding) { +set_ce (CT ct, int encoding) +{ const char *ce = ce_str (encoding); const struct str2init *ctinit = get_ce_method (ce); if (ctinit) { char *cte = concat (" ", ce, "\n", NULL); - int found_cte = 0; + bool found_cte = false; HF hf; /* Decoded contents might be in ct->c_cefile.ce_file, if the caller is decode_text_parts (). Save because we'll @@ -1214,7 +1282,7 @@ set_ce (CT ct, int encoding) { /* Update/add Content-Transfer-Encoding header field. */ for (hf = ct->c_first_hf; hf; hf = hf->next) { if (! strcasecmp (ENCODING_FIELD, hf->name)) { - found_cte = 1; + found_cte = true; free (hf->value); hf->value = cte; } @@ -1238,7 +1306,8 @@ 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, int replacetextplain) { +ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain) +{ int status = OK; switch ((*ct)->c_type) { @@ -1399,10 +1468,11 @@ ensure_text_plain (CT *ct, CT parent, int *message_mods, int replacetextplain) { */ static int find_textplain_sibling (CT parent, int replacetextplain, - int *new_subpart_number) { + int *new_subpart_number) +{ struct multipart *mp = (struct multipart *) parent->c_ctparams; struct part *part, *prev; - int has_text_plain = 0; + bool has_text_plain = false; for (prev = part = mp->mp_parts; part; part = part->mp_next) { ++*new_subpart_number; @@ -1425,7 +1495,7 @@ find_textplain_sibling (CT parent, int replacetextplain, free_content (old_part->mp_part); free (old_part); } else { - has_text_plain = 1; + has_text_plain = true; } break; } @@ -1437,10 +1507,11 @@ find_textplain_sibling (CT parent, int replacetextplain, /* - * Insert a new text/plain part. + * Insert a new text/plain part in a multipart part. */ static int -insert_new_text_plain_part (CT ct, int new_subpart_number, CT parent) { +insert_new_text_plain_part (CT ct, int new_subpart_number, CT parent) +{ struct multipart *mp = (struct multipart *) parent->c_ctparams; struct part *new_part; @@ -1469,7 +1540,8 @@ insert_new_text_plain_part (CT ct, int new_subpart_number, CT parent) { * Create a text/plain part to go along with non-plain sibling part. */ static CT -build_text_plain_part (CT encoded_part) { +build_text_plain_part (CT encoded_part) +{ CT tp_part = divide_part (encoded_part); char *tmp_plain_file = NULL; @@ -1507,13 +1579,15 @@ build_text_plain_part (CT encoded_part) { * Slip new text/plain part into a new multipart/alternative. */ static int -insert_into_new_mp_alt (CT *ct, int *message_mods) { +insert_into_new_mp_alt (CT *ct, int *message_mods) +{ + /* The following will call decode_part(). */ CT tp_part = build_text_plain_part (*ct); int status = OK; if (tp_part) { - CT mp_alt = build_multipart_alt (*ct, tp_part, CT_MULTIPART, - MULTI_ALTERNATE); + CT mp_alt = build_multipart (*ct, tp_part, CT_MULTIPART, + MULTI_ALTERNATE); if (mp_alt) { struct multipart *mp = (struct multipart *) mp_alt->c_ctparams; @@ -1543,11 +1617,85 @@ insert_into_new_mp_alt (CT *ct, int *message_mods) { } +/* + * Slip new text/plain part into a new multipart/mixed. + */ +static int +insert_into_new_mp_mixed (CT *ct, const char *content, int *message_mods) +{ + CT main_part = divide_part (*ct); + const char *reason = NULL; + const int encoding = content_encoding (main_part, &reason); + int status = OK; + + if (set_ct_type(main_part, (*ct)->c_type, (*ct)->c_subtype, + main_part->c_encoding) != OK) { + inform("failed to set Content-Type of main part"); + return NOTOK; + } + if (set_ct_type(*ct, (*ct)->c_type, (*ct)->c_subtype, encoding) != OK) { + inform("failed to set Content-Type of new part"); + return NOTOK; + } + + if (main_part) { + /* Load remainder into the new part. */ + CE cefile = &(*ct)->c_cefile; + CT mp_alt; + + cefile->ce_file = + mh_xstrdup(m_mktemp2 (NULL, invo_name, NULL, &cefile->ce_fp)); + if (cefile->ce_file == NULL) { + die("unable to create temporary file in %s", get_temp_dir()); + } + cefile->ce_unlink = 1; + fprintf (cefile->ce_fp, "%s", content); + + /* Put both parts into a new multipart. */ + mp_alt = build_multipart (*ct, main_part, CT_MULTIPART, MULTI_MIXED); + if (mp_alt) { + struct multipart *mp = (struct multipart *) mp_alt->c_ctparams; + + /* So fix_composite_cte doesn't try to overwrite the encoding. If + the content needs to be decoded, c_encoding will be properly + set. */ + mp_alt->c_encoding = encoding; + + if (mp && mp->mp_parts) { + mp->mp_parts->mp_part = main_part; + /* Make the new multipart/alternative the parent. */ + *ct = mp_alt; + + ++*message_mods; + if (verbosw) { + report (NULL, (*ct)->c_partno, (*ct)->c_file, + "insert text/plain part"); + } + } else { + free_content (main_part); + free_content (mp_alt); + status = NOTOK; + } + } else { + inform("failed to build multipart/alternate"); + status = NOTOK; + } + } else { + /* Should never happen. */ + inform("failed to insert new text part into multipart/related"); + status = NOTOK; + } + + return status; +} + + /* * Clone a MIME part. */ static CT -divide_part (CT ct) { +divide_part (CT ct) +{ CT new_part; NEW0(new_part); @@ -1577,20 +1725,21 @@ divide_part (CT ct) { * Copy the content info from one part to another. */ static void -copy_ctinfo (CI dest, CI src) { +copy_ctinfo (CI dest, CI src) +{ PM s_pm, d_pm; dest->ci_type = src->ci_type ? mh_xstrdup (src->ci_type) : NULL; dest->ci_subtype = src->ci_subtype ? mh_xstrdup (src->ci_subtype) : NULL; 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, 0); - if (s_pm->pm_charset) { - d_pm->pm_charset = mh_xstrdup(s_pm->pm_charset); + d_pm = add_param(&dest->ci_first_pm, &dest->ci_last_pm, s_pm->pm_name, + s_pm->pm_value, 0); + if (s_pm->pm_charset) { + d_pm->pm_charset = mh_xstrdup(s_pm->pm_charset); } - if (s_pm->pm_lang) { - d_pm->pm_lang = mh_xstrdup(s_pm->pm_lang); + if (s_pm->pm_lang) { + d_pm->pm_lang = mh_xstrdup(s_pm->pm_lang); } } @@ -1603,14 +1752,14 @@ copy_ctinfo (CI dest, CI src) { * Decode content. */ static int -decode_part (CT ct) { - char *tmp_decoded; - int status; +decode_part (CT ct) +{ + char *tempfile, *tmp_decoded; FILE *file; - char *tempfile; + int status; if ((tempfile = m_mktemp2 (NULL, invo_name, NULL, &file)) == NULL) { - adios (NULL, "unable to create temporary file in %s", get_temp_dir()); + die("unable to create temporary file in %s", get_temp_dir()); } tmp_decoded = mh_xstrdup (tempfile); /* The following call will load ct->c_cefile.ce_file with the tmp @@ -1627,13 +1776,142 @@ decode_part (CT ct) { } +/* + * If base64-encoded content has a text trailer, return the location, relative + * to c->c_begin, where the valid base64 ends. And return the trailer in the + * addresses pointed to by remainderp. The caller is responsible for + * deallocating that. If no text trailer, return ct->c_end - ct->c_begin and + * leave remainderp unchanged. + */ +static size_t +get_valid_base64 (CT ct, char **remainderp) { + const size_t len = ct->c_end - ct->c_begin; + char *buf, format[16]; + size_t pos; + int fd; + + if (! ct->c_fp && ((ct->c_fp = fopen (ct->c_file, "r")) == NULL)) { + advise (ct->c_file, "unable to open for reading"); + return NOTOK; + } + if ((fd = fileno (ct->c_fp)) == -1 || + lseek (fd, ct->c_begin, SEEK_SET) == (off_t) -1) { + advise (ct->c_file, "unable to seek in"); + return NOTOK; + } + buf = mh_xmalloc(len + 1); + snprintf(format, sizeof format, "%%%luc", (unsigned long) len); + if (fscanf(ct->c_fp, format, buf) == EOF) { + advise (ct->c_file, "unable to read"); + return NOTOK; + } + buf[len] = '\0'; + + pos = find_invalid_base64_pos(buf); + + if (ct->c_begin + pos < (size_t) ct->c_end) { + *remainderp = mh_xstrdup(&buf[pos]); + } else { + pos = ct->c_end - ct->c_begin; + } + free(buf); + + return pos; +} + + +/* + * Find position in byte string of invalid base64 code. Skip individual + * invalid characters because RFC 2045 Sec 6.8 says they should be ignored. + * The motivating use case is a text footer that was mistakenly applied to + * base64 content. Therefore, if any of these is found, return the position + * of: + * 1. The byte (or end) after one or two consecutive pad ('=') bytes. + * 2. The first of a pair of invalid base64 bytes. + * + * If the base64 code is valid, return the position of the null terminator. + * + * encoded - the base64-encoded string + */ +static size_t +find_invalid_base64_pos (const char *encoded) { + const char *cp; + size_t pos; + bool found_pad = false; + unsigned int found_invalid = 0; + + for (cp = encoded, pos = 0; + *cp && ! found_pad && found_invalid < 2; + ++cp, ++pos) { + if (isspace ((unsigned char) *cp) || + isalnum ((unsigned char) *cp) || + *cp == '+' || *cp == '/' || *cp == '=') { + /* Valid base64 byte. */ + if (*cp == '=') { + /* "evidence that the end of the data has been reached" + according to RFC 2045 */ + found_pad = true; + } + /* Require consecutive invalid bytes. Let decodeBase64() handle + individual ones. */ + found_invalid = 0; + } else { + ++found_invalid; + } + } + + if (found_pad && *cp && *cp == '=') { + /* Skip over last in pair of ==. */ + ++cp, ++pos; + } else if (found_invalid == 2) { + /* If a pair of consecutive invalid bytes, back up to first one. */ + --cp, --pos; + --cp, --pos; + } + + /* Skip over any trailing whitespace. */ + while (*cp && isspace((unsigned char) *cp)) { + ++cp, ++pos; + } + + return pos; +} + + +/* + * Check for valid base64 encoding, and "fix" if invalid. + */ +static int +check_base64_encoding (CT *ctp) +{ + char *remainder = NULL; + int status = OK; + + /* If there's a footer after base64 content, set c_end to before it, and + store the footer in remainder. */ + (*ctp)->c_end = (*ctp)->c_begin + get_valid_base64(*ctp, &remainder); + + if (remainder != NULL) { + /* Move ct to a subpart of a new multipart/related, and add the + remainder as a new text/plain subpart of it. */ + int ignore_message_mods = 0; + + status = insert_into_new_mp_mixed(ctp, remainder, &ignore_message_mods); + free(remainder); + } + + return status; +} + + /* * Reformat content as plain text. * Some of the arguments aren't really needed now, but maybe will * be in the future for other than text types. */ static int -reformat_part (CT ct, char *file, char *type, char *subtype, int c_type) { +reformat_part (CT ct, char *file, char *type, char *subtype, int c_type) +{ int output_subtype, output_encoding; const char *reason = NULL; char *cp, *cf; @@ -1696,12 +1974,15 @@ reformat_part (CT ct, char *file, char *type, char *subtype, int c_type) { /* - * Fill in a multipart/alternative part. + * Create and fill in a multipart part. */ static CT -build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { +build_multipart (CT first_part, CT new_part, int type, int subtype) +{ char *boundary_prefix = "----=_nmh-multipart"; - char *boundary = concat (boundary_prefix, first_alt->c_partno, NULL); + static unsigned int bp_uses = 0; + char bp_uses_buf[16]; + char *boundary; char *boundary_indicator = "; boundary="; char *typename, *subtypename, *name; CT ct; @@ -1711,7 +1992,15 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { NEW0(ct); - /* Set up the multipart/alternative part. These fields of *ct were + if (bp_uses > 0) { + snprintf(bp_uses_buf, sizeof bp_uses_buf - 1, "-%d", bp_uses++); + } else { + bp_uses_buf[0] = '\0'; + } + boundary = + concat (boundary_prefix, bp_uses_buf, first_part->c_partno, NULL); + + /* Set up the multipart part. These fields of *ct were initialized to 0 by mh_xcalloc(): c_fp, c_unlink, c_begin, c_end, c_vrsn, c_ctline, c_celine, @@ -1724,7 +2013,7 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { c_showproc, c_termproc, c_storeproc, c_storage, c_folder */ - ct->c_file = mh_xstrdup (first_alt->c_file); + ct->c_file = mh_xstrdup (first_part->c_file); ct->c_type = type; ct->c_subtype = subtype; @@ -1747,8 +2036,7 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { boundary_in_content (&new_part->c_cefile.ce_fp, new_part->c_cefile.ce_file, boundary)) == NOTOK) { - free_content (ct); - return NULL; + goto return_null; } } @@ -1759,8 +2047,7 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { boundary_in_content (&new_part->c_fp, new_part->c_file, boundary)) == NOTOK) { - free_content (ct); - return NULL; + goto return_null; } } @@ -1773,15 +2060,14 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { snprintf (buffer2, sizeof buffer2, "%d", serial); boundary = concat (boundary_prefix, - FENDNULL(first_alt->c_partno), + FENDNULL(first_part->c_partno), "-", buffer2, NULL); } } if (found_boundary) { inform("giving up trying to find a unique boundary"); - free_content (ct); - return NULL; + goto return_null; } } @@ -1789,18 +2075,18 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { boundary, "\"", NULL); /* Load c_first_hf and c_last_hf. */ - transfer_noncontent_headers (first_alt, ct); + transfer_noncontent_headers (first_part, ct); add_header (ct, mh_xstrdup (TYPE_FIELD), concat (name, "\n", NULL)); free (name); /* Load c_partno. */ - if (first_alt->c_partno) { - ct->c_partno = mh_xstrdup (first_alt->c_partno); - free (first_alt->c_partno); - first_alt->c_partno = concat (ct->c_partno, ".1", NULL); + if (first_part->c_partno) { + ct->c_partno = mh_xstrdup (first_part->c_partno); + free (first_part->c_partno); + first_part->c_partno = concat (ct->c_partno, ".1", NULL); new_part->c_partno = concat (ct->c_partno, ".2", NULL); } else { - first_alt->c_partno = mh_xstrdup ("1"); + first_part->c_partno = mh_xstrdup ("1"); new_part->c_partno = mh_xstrdup ("2"); } @@ -1815,7 +2101,7 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { NEW(p); NEW(p->mp_next); p->mp_next->mp_next = NULL; - p->mp_next->mp_part = first_alt; + p->mp_next->mp_part = first_part; NEW0(m); m->mp_start = concat (boundary, "\n", NULL); @@ -1826,6 +2112,11 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { free (boundary); return ct; + +return_null: + free_content(ct); + free(boundary); + return NULL; } @@ -1833,10 +2124,11 @@ build_multipart_alt (CT first_alt, CT new_part, int type, int subtype) { * Check that the boundary does not appear in the content. */ static int -boundary_in_content (FILE **fp, char *file, const char *boundary) { +boundary_in_content (FILE **fp, char *file, const char *boundary) +{ char buffer[NMH_BUFSIZ]; size_t bytes_read; - int found_boundary = 0; + bool found_boundary = false; /* free_content() will close *fp if we fopen it here. */ if (! *fp && (*fp = fopen (file, "r")) == NULL) { @@ -1847,7 +2139,7 @@ boundary_in_content (FILE **fp, char *file, const char *boundary) { fseeko (*fp, 0L, SEEK_SET); while ((bytes_read = fread (buffer, 1, sizeof buffer, *fp)) > 0) { if (find_str (buffer, bytes_read, boundary)) { - found_boundary = 1; + found_boundary = true; break; } } @@ -1860,14 +2152,15 @@ boundary_in_content (FILE **fp, char *file, const char *boundary) { * Remove all non-Content headers. */ static void -transfer_noncontent_headers (CT old, CT new) { +transfer_noncontent_headers (CT old, CT new) +{ HF hp, hp_prev; hp_prev = hp = old->c_first_hf; while (hp) { HF next = hp->next; - if (strncasecmp (XXX_FIELD_PRF, hp->name, strlen (XXX_FIELD_PRF))) { + if (strncasecmp (XXX_FIELD_PRF, hp->name, LEN(XXX_FIELD_PRF))) { if (hp == old->c_last_hf) { if (hp == old->c_first_hf) { old->c_last_hf = old->c_first_hf = NULL; @@ -1904,14 +2197,15 @@ transfer_noncontent_headers (CT old, CT new) { * Set content type. */ static int -set_ct_type (CT ct, int type, int subtype, int encoding) { +set_ct_type (CT ct, int type, int subtype, int encoding) +{ char *typename = ct_type_str (type); char *subtypename = ct_subtype_str (type, subtype); /* E.g, " text/plain" */ char *type_subtypename = concat (" ", typename, "/", subtypename, NULL); /* E.g, " text/plain\n" */ char *name_plus_nl = concat (type_subtypename, "\n", NULL); - int found_content_type = 0; + bool found_content_type = false; HF hf; const char *cp = NULL; char *ctline; @@ -1920,7 +2214,7 @@ set_ct_type (CT ct, int type, int subtype, int encoding) { /* Update/add Content-Type header field. */ for (hf = ct->c_first_hf; hf; hf = hf->next) { if (! strcasecmp (TYPE_FIELD, hf->name)) { - found_content_type = 1; + found_content_type = true; free (hf->value); hf->value = (cp = strchr (ct->c_ctline, ';')) ? concat (type_subtypename, cp, "\n", NULL) @@ -1965,7 +2259,8 @@ set_ct_type (CT ct, int type, int subtype, int encoding) { */ static int decode_text_parts (CT ct, int encoding, const char *decodetypes, - int *message_mods) { + int *message_mods) +{ int status = OK; int lf_line_endings = 0; @@ -2085,23 +2380,24 @@ decode_text_parts (CT ct, int encoding, const char *decodetypes, * decodetypes (which came from the -decodetypes switch). */ static int -should_decode(const char *decodetypes, const char *type, const char *subtype) { +should_decode(const char *decodetypes, const char *type, const char *subtype) +{ /* Quick search for matching type[/subtype] in decodetypes: bracket decodetypes with commas, then search for ,type, and ,type/subtype, in it. */ - int found_match = 0; + bool found_match = false; char *delimited_decodetypes = concat(",", decodetypes, ",", NULL); char *delimited_type = concat(",", type, ",", NULL); if (nmh_strcasestr(delimited_decodetypes, delimited_type)) { - found_match = 1; + found_match = true; } else if (subtype != NULL) { char *delimited_type_subtype = concat(",", type, "/", subtype, ",", NULL); if (nmh_strcasestr(delimited_decodetypes, delimited_type_subtype)) { - found_match = 1; + found_match = true; } free(delimited_type_subtype); } @@ -2120,7 +2416,8 @@ should_decode(const char *decodetypes, const char *type, const char *subtype) { * to a string explaining why. */ static int -content_encoding (CT ct, const char **reason) { +content_encoding (CT ct, const char **reason) +{ CE ce = &ct->c_cefile; int encoding = CE_7BIT; @@ -2179,7 +2476,8 @@ content_encoding (CT ct, const char **reason) { * Strip carriage returns from content. */ static int -strip_crs (CT ct, int *message_mods) { +strip_crs (CT ct, int *message_mods) +{ char *charset = content_charset (ct); int status = OK; @@ -2196,8 +2494,8 @@ strip_crs (CT ct, int *message_mods) { FILE **fp = NULL; size_t begin; size_t end; - int has_crs = 0; - int opened_input_file = 0; + bool has_crs = false; + bool opened_input_file = false; if (ct->c_cefile.ce_file) { file = &ct->c_cefile.ce_file; @@ -2216,7 +2514,7 @@ strip_crs (CT ct, int *message_mods) { advise (*file, "unable to open for reading"); status = NOTOK; } else { - opened_input_file = 1; + opened_input_file = true; } } } @@ -2236,13 +2534,13 @@ strip_crs (CT ct, int *message_mods) { modify the content. */ char *cp; size_t i; - int last_char_was_cr = 0; + bool last_char_was_cr = false; if (end > 0) { bytes_to_read -= bytes_read; } for (i = 0, cp = buffer; i < bytes_read; ++i, ++cp) { if (*cp == '\n' && last_char_was_cr) { - has_crs = 1; + has_crs = true; break; } @@ -2256,7 +2554,7 @@ strip_crs (CT ct, int *message_mods) { char *tempfile = m_mktemp2 (NULL, invo_name, &fd, NULL); if (tempfile == NULL) { - adios (NULL, "unable to create temporary file in %s", + die("unable to create temporary file in %s", get_temp_dir()); } stripped_content_file = mh_xstrdup (tempfile); @@ -2267,11 +2565,11 @@ strip_crs (CT ct, int *message_mods) { 0) { char *cp; size_t i; - int last_char_was_cr = 0; + bool last_char_was_cr = false; for (i = 0, cp = buffer; i < bytes_read; ++i, ++cp) { if (*cp == '\r') { - last_char_was_cr = 1; + last_char_was_cr = true; } else if (last_char_was_cr) { if (*cp != '\n') { if (write (fd, "\r", 1) < 0) { @@ -2281,12 +2579,12 @@ strip_crs (CT ct, int *message_mods) { if (write (fd, cp, 1) < 0) { advise (tempfile, "write"); } - last_char_was_cr = 0; + last_char_was_cr = false; } else { if (write (fd, cp, 1) < 0) { advise (tempfile, "write"); } - last_char_was_cr = 0; + last_char_was_cr = false; } } } @@ -2295,13 +2593,14 @@ strip_crs (CT ct, int *message_mods) { inform("unable to write temporary file %s, continuing...", stripped_content_file); (void) m_unlink (stripped_content_file); + free(stripped_content_file); status = NOTOK; } else { /* Replace the decoded file with the converted one. */ if (ct->c_cefile.ce_file && ct->c_cefile.ce_unlink) (void) m_unlink (ct->c_cefile.ce_file); - mh_xfree(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; @@ -2332,19 +2631,20 @@ strip_crs (CT ct, int *message_mods) { * of the part C-T-E's. */ static void -update_cte (CT ct) { +update_cte (CT ct) +{ const int least_restrictive_enc = least_restrictive_encoding (ct); if (least_restrictive_enc != CE_UNKNOWN && least_restrictive_enc != CE_7BIT) { char *cte = concat (" ", ce_str (least_restrictive_enc), "\n", NULL); HF hf; - int found_cte = 0; + bool found_cte = false; /* Update/add Content-Transfer-Encoding header field. */ for (hf = ct->c_first_hf; hf; hf = hf->next) { if (! strcasecmp (ENCODING_FIELD, hf->name)) { - found_cte = 1; + found_cte = true; free (hf->value); hf->value = cte; } @@ -2361,7 +2661,8 @@ update_cte (CT ct) { * within a message. */ static int -least_restrictive_encoding (CT ct) { +least_restrictive_encoding (CT ct) +{ int encoding = CE_UNKNOWN; switch (ct->c_type) { @@ -2409,7 +2710,8 @@ least_restrictive_encoding (CT ct) { * CE_8BIT is less restrictive than CE_7BIT. */ static int -less_restrictive (int encoding, int second_encoding) { +less_restrictive (int encoding, int second_encoding) +{ switch (second_encoding) { case CE_BINARY: return encoding != CE_BINARY; @@ -2428,7 +2730,8 @@ less_restrictive (int encoding, int second_encoding) { * Convert character set of each part. */ static int -convert_charsets (CT ct, char *dest_charset, int *message_mods) { +convert_charsets (CT ct, char *dest_charset, int *message_mods) +{ int status = OK; switch (ct->c_type) { @@ -2493,36 +2796,37 @@ convert_charsets (CT ct, char *dest_charset, int *message_mods) { * headers, respectively. */ static int -fix_always (CT ct, int *message_mods) { +fix_always (CT *ctp, const fix_transformations *fx, int *message_mods) +{ int status = OK; - switch (ct->c_type) { + switch ((*ctp)->c_type) { case CT_MULTIPART: { - struct multipart *m = (struct multipart *) ct->c_ctparams; + struct multipart *m = (struct multipart *) (*ctp)->c_ctparams; struct part *part; for (part = m->mp_parts; status == OK && part; part = part->mp_next) { - status = fix_always (part->mp_part, message_mods); + status = fix_always (&part->mp_part, fx, message_mods); } break; } case CT_MESSAGE: - if (ct->c_subtype == MESSAGE_EXTERNAL) { - struct exbody *e = (struct exbody *) ct->c_ctparams; + if ((*ctp)->c_subtype == MESSAGE_EXTERNAL) { + struct exbody *e = (struct exbody *) (*ctp)->c_ctparams; - status = fix_always (e->eb_content, message_mods); + status = fix_always (&e->eb_content, fx, message_mods); } break; default: { HF hf; - if (ct->c_first_hf) { - fix_filename_encoding (ct); + if ((*ctp)->c_first_hf) { + fix_filename_encoding (*ctp); } - for (hf = ct->c_first_hf; hf; hf = hf->next) { + for (hf = (*ctp)->c_first_hf; hf; hf = hf->next) { size_t len = strlen (hf->value); if (strcasecmp (hf->name, TYPE_FIELD) != 0 && @@ -2546,37 +2850,102 @@ fix_always (CT ct, int *message_mods) { hf->value[len - 1] = '\0'; /* Also, if Content-Type parameter, remove trailing ';' - from ct->c_ctline. This probably isn't necessary + from (*ctp)->c_ctline. This probably isn't necessary but can't hurt. */ - if (strcasecmp(hf->name, TYPE_FIELD) == 0 && ct->c_ctline) { - size_t l = strlen(ct->c_ctline) - 1; - while (isspace((unsigned char)(ct->c_ctline[l])) || - ct->c_ctline[l] == ';') { - ct->c_ctline[l--] = '\0'; + if (strcasecmp(hf->name, TYPE_FIELD) == 0 && (*ctp)->c_ctline) { + size_t l = strlen((*ctp)->c_ctline) - 1; + while (isspace((unsigned char)((*ctp)->c_ctline[l])) || + (*ctp)->c_ctline[l] == ';') { + (*ctp)->c_ctline[l--] = '\0'; if (l == 0) { break; } } } ++*message_mods; if (verbosw) { - report (NULL, ct->c_partno, ct->c_file, + report (NULL, (*ctp)->c_partno, (*ctp)->c_file, "remove trailing ; from %s parameter value", hf->name); } } } + + if (fx->checkbase64 && (*ctp)->c_encoding == CE_BASE64) { + status = check_base64_encoding (ctp); + } }} return status; } +/* + * Decodes UTF-8 encoded header values. Similar to fix_filename_param(), but + * does not modify any MIME parameter values. + */ +static int +decode_header_field_bodies (CT ct, int *message_mods) +{ + int status = OK; + + switch (ct->c_type) { + case CT_MULTIPART: { + struct multipart *m = (struct multipart *) ct->c_ctparams; + struct part *part; + + for (part = m->mp_parts; status == OK && part; part = part->mp_next) { + status = decode_header_field_bodies (part->mp_part, message_mods); + } + break; + } + + case CT_MESSAGE: + if (ct->c_subtype == MESSAGE_EXTERNAL) { + struct exbody *e = (struct exbody *) ct->c_ctparams; + + status = decode_header_field_bodies (e->eb_content, message_mods); + } + break; + } + + HF hf; + + for (hf = ct->c_first_hf; hf; hf = hf->next) { + /* Only decode UTF-8 values. */ + if (hf->value && has_suffix(hf->value, "?=\n") && + (! strncasecmp (hf->value, " =?utf8?", 8) || + ! strncasecmp (hf->value, " =?utf-8?", 9))) { + /* Looks like an RFC 2047 encoded parameter. */ + char decoded[PATH_MAX + 1]; + + if (decode_rfc2047 (hf->value, decoded, sizeof decoded)) { + const size_t len = strlen(decoded); + + /* decode_rfc2047() could truncate if the buffer fills up. + Detect and discard if that happened. */ + if (len < sizeof(decoded) - 1 && strcmp(hf->value, decoded)) { + hf->value = mh_xrealloc (hf->value, len + 1); + strncpy (hf->value, decoded, len + 1); + ++*message_mods; + } + } else { + inform("failed to decode %s parameter %s", hf->name, hf->value); + status = NOTOK; + } + } + } + + return status; +} + + /* * Factor out common code for loops in fix_filename_encoding(). */ static int -fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm) { - int fixed = 0; +fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm) +{ + bool fixed = false; if (has_prefix(value, "=?") && has_suffix(value, "?=")) { /* Looks like an RFC 2047 encoded parameter. */ @@ -2585,7 +2954,7 @@ fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm) { if (decode_rfc2047 (value, decoded, sizeof decoded)) { /* Encode using RFC 2231. */ replace_param (first_pm, last_pm, name, decoded, 0); - fixed = 1; + fixed = true; } else { inform("failed to decode %s parameter %s", name, value); } @@ -2601,7 +2970,8 @@ fix_filename_param (char *name, char *value, PM *first_pm, PM *last_pm) { * headers, respectively. */ static int -fix_filename_encoding (CT ct) { +fix_filename_encoding (CT ct) +{ PM pm; HF hf; int fixed = 0; @@ -2667,7 +3037,8 @@ fix_filename_encoding (CT ct) { */ static int write_content (CT ct, const char *input_filename, char *outfile, FILE *outfp, - int modify_inplace, int message_mods) { + int modify_inplace, int message_mods) +{ int status = OK; if (modify_inplace) { @@ -2715,7 +3086,7 @@ write_content (CT ct, const char *input_filename, char *outfile, FILE *outfp, } } else { inform("unable to remove input file %s, " - "not modifying it, continuing...", infile); + "not modifying it, continuing...", infile); (void) m_unlink (outfile); status = NOTOK; } @@ -2744,7 +3115,8 @@ write_content (CT ct, const char *input_filename, char *outfile, FILE *outfp, * function to do it. It touches the parts the decodetypes identifies. */ static void -set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings) { +set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings) +{ switch (ct->c_type) { case CT_MULTIPART: { struct multipart *m = (struct multipart *) ct->c_ctparams; @@ -2780,7 +3152,8 @@ set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings) { * use the standard MH backup file. */ static int -remove_file (const char *file) { +remove_file (const char *file) +{ if (rmmproc) { char *rmm_command = concat (rmmproc, " ", file, NULL); int status = system (rmm_command); @@ -2799,7 +3172,8 @@ remove_file (const char *file) { * Output formatted message to user. */ static void -report (char *what, char *partno, char *filename, char *message, ...) { +report (char *what, char *partno, char *filename, char *message, ...) +{ va_list args; char *fmt;