From 5c354366b032fbd5d0c94300ff9e8d70690dd6a6 Mon Sep 17 00:00:00 2001
From: "plvines (corpora)" <paul.vines@baesystems.com>
Date: Thu, 27 Feb 2020 18:50:03 +0000
Subject: [PATCH] Relaxed xref spacing
---
pdf.c | 15 +++++++++++----
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/pdf.c b/pdf.c
index 7ee7759..c301021 100644
--- a/pdf.c
+++ b/pdf.c
@@ -555,12 +555,15 @@ act_viol(const HParseResult *p, void *viol)
else {
severity = severity_parse->ast->seq->elements[0]->uint;
}
- offset = p->ast->seq->elements[1]->uint / 8;
+ offset = p->ast->seq->elements[p->ast->seq->used-1]->uint / 8;
fprintf(stderr, "VIOLATION[%d]@%d (0x%x): %s\n", severity, offset, offset, (char *) viol);
if (strictness && severity > strictness) {
exit(1);
}
/* Just return the parse AST, drop the h_tell */
+ if (p->ast->seq->used == 1) {
+ return (HParsedToken *) NULL;
+ }
return (HParsedToken *) p->ast->seq->elements[0];
}
@@ -1112,7 +1115,11 @@ init_parser(struct Env *aux)
H_RULE(xrtyp, CHX(h_ch('n'), h_ch('f')));
H_ARULE(xroff, REP(digit, 10));
H_ARULE(xrgen, REP(digit, 5));
- H_ARULE(xrent, SEQ(xroff, IGN(sp), xrgen, IGN(sp), xrtyp, IGN(xreol)));
+ H_ARULE(xrent, SEQ(xroff, IGN(CHX(VIOL(SEQ(lwchar, h_many1(lwchar)), "Multi-WS in xref offset_gen entry (severity=1)"), sp)),
+ xrgen, IGN(CHX(VIOL(SEQ(lwchar, h_many1(lwchar)), "Multi-WS in xref gen_use entry (severity=1)"), sp)),
+ xrtyp, IGN(CHX(VIOL(SEQ(wchar, wchar, h_many1(wchar)), "Greater-than-2-byte WS at end of xref entry (severity=1)"),
+ xreol,
+ VIOL(SEQ(h_many1(wchar)), "Nonconformant WS at end of xref entry (severity=1)")))));
H_RULE(xrhead, SEQ(nat, IGN(sp), nat, nl));
H_RULE(xrsub, SEQ(xrhead, h_many(xrent)));
H_ARULE(xrefs, SEQ(KW("xref"), nl, h_many(xrsub)));
@@ -1148,8 +1155,8 @@ init_parser(struct Env *aux)
H_RULE(xr_td, SEQ(xrefs, KW("trailer"), ws, dict));
H_RULE(hdr_junk, CHX(comment,
- VIOL(h_many1(h_butnot(h_ch_range(0, 255), SEQ(npair, wel, KW("obj")))),
- "Uncommented junk after header (severity=1)")));
+ VIOL(h_many1(h_butnot(h_ch_range(0, 255), SEQ(npair, wel, KW("obj")))),
+ "Uncommented junk after header (severity=1)")));
H_RULE(tail, SEQ(body, CHX(SEQ(h_optional(xr_td), startxr),
VIOL(SEQ(xr_td, OPT(SEQ(nl, KW("startxref"), nl, lws, nat, nl)),
OPT(nl), OPT(LIT("%%EOF")), OPT(nl)),
--
GitLab