From d800f1323a545bfcf2bf8fe3f63b6e430bb5e75d Mon Sep 17 00:00:00 2001
From: pd3 <pd3@sanger.ac.uk>
Date: Tue, 14 Nov 2023 21:01:17 +0100
Subject: [PATCH 1/2] Temporary workaround when excessive memory is required by
 FORMAT fields

The BCF limit of ~2GB per VCF row causes the parser to return an error.
In this temporary solution we drop the excessive fields, typically these
will be PL and other Number=G tags. in future work the library would
automatically convert such tags into their localized alternatives (LPL).
---
 vcf.c | 35 ++++++++++++++++++++++++++++++-----
 1 file changed, 30 insertions(+), 5 deletions(-)

diff --git a/vcf.c b/vcf.c
index 301d818cf..1a93bc0e6 100644
--- a/vcf.c
+++ b/vcf.c
@@ -2202,7 +2202,7 @@ int bcf_write(htsFile *hfp, bcf_hdr_t *h, bcf1_t *v)
     if ( hfp->format.format == vcf || hfp->format.format == text_format )
         return vcf_write(hfp,h,v);
 
-    if ( v->errcode )
+    if ( v->errcode & ~BCF_ERR_LIMITS ) // todo: unsure about the other BCF_ERR_LIMITS branches in vcf_parse_format_alloc4()
     {
         // vcf_parse1() encountered a new contig or tag, undeclared in the
         // header.  At this point, the header must have been printed,
@@ -3004,9 +3004,12 @@ static int vcf_parse_format_alloc4(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
         // malformed VCF data is less likely to take excessive memory and/or
         // time.
         if ((uint64_t) mem->l + v->n_sample * (uint64_t)f->size > INT_MAX) {
-            hts_log_error("Excessive memory required by FORMAT fields at %s:%"PRIhts_pos, bcf_seqname_safe(h,v), v->pos+1);
+            static int warned = 0;
+            if ( !warned ) hts_log_warning("Excessive memory required by FORMAT fields at %s:%"PRIhts_pos, bcf_seqname_safe(h,v), v->pos+1);
+            warned = 1;
             v->errcode |= BCF_ERR_LIMITS;
-            return -1;
+            f->size = f->offset = 0;
+            continue;
         }
 
         f->offset = mem->l;
@@ -3065,7 +3068,12 @@ static int vcf_parse_format_fill5(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
                 return -1;
             }
 
-            if (htype == BCF_HT_STR) {
+            if ( !z->size )
+            {
+                // this field is to be ignored, it's too big
+                while ( *t != ':' && *t ) t++;
+            }
+            else if (htype == BCF_HT_STR) {
                 int l;
                 if (z->is_gt) {
                     // Genotypes.
@@ -3237,10 +3245,14 @@ static int vcf_parse_format_fill5(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
 static int vcf_parse_format_gt6(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
                                 const char *p, const char *q, fmt_aux_t *fmt) {
     kstring_t *str = &v->indiv;
-    int i;
+    int i, need_downsize = 0;
     if (v->n_sample > 0) {
         for (i = 0; i < v->n_fmt; ++i) {
             fmt_aux_t *z = &fmt[i];
+            if ( !z->size ) {
+                need_downsize = 1;
+                continue;
+            }
             bcf_enc_int1(str, z->key);
             if ((z->y>>4&0xf) == BCF_HT_STR && !z->is_gt) {
                 bcf_enc_size(str, z->size, BCF_BT_CHAR);
@@ -3257,6 +3269,19 @@ static int vcf_parse_format_gt6(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
                 }
             }
         }
+
+    }
+    if ( need_downsize ) {
+        i = 1;
+        while ( i < v->n_fmt ) {
+            if ( !fmt[i].size )
+            {
+                memmove(&fmt[i-1],&fmt[i],sizeof(*fmt));
+                v->n_fmt--;
+            }
+            else
+                i++;
+        }
     }
 
     return 0;

From f1b5470c13dd50300cd72bd70ff70f930067d22b Mon Sep 17 00:00:00 2001
From: Petr Danecek <pd3@sanger.ac.uk>
Date: Tue, 21 Nov 2023 14:45:58 +0000
Subject: [PATCH 2/2] VCF parsing fix

The removal of large tags by d800f1323a545bfcf2bf8fe3f63b6e430bb5e75d inadvertently
changed output for invalid rows, such as the one introduced by 8f782d120433eb0321a1c0db6ee49e6bc838563e.
This was caused by relying on fmt->size==0 never happening - such assumption is
correct, I think, but only for valid tags, not such that failed parsing
---
 vcf.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/vcf.c b/vcf.c
index 1a93bc0e6..7499ad8b9 100644
--- a/vcf.c
+++ b/vcf.c
@@ -3008,7 +3008,8 @@ static int vcf_parse_format_alloc4(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
             if ( !warned ) hts_log_warning("Excessive memory required by FORMAT fields at %s:%"PRIhts_pos, bcf_seqname_safe(h,v), v->pos+1);
             warned = 1;
             v->errcode |= BCF_ERR_LIMITS;
-            f->size = f->offset = 0;
+            f->size = -1;
+            f->offset = 0;
             continue;
         }
 
@@ -3068,7 +3069,7 @@ static int vcf_parse_format_fill5(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
                 return -1;
             }
 
-            if ( !z->size )
+            if ( z->size==-1 )
             {
                 // this field is to be ignored, it's too big
                 while ( *t != ':' && *t ) t++;
@@ -3249,7 +3250,7 @@ static int vcf_parse_format_gt6(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
     if (v->n_sample > 0) {
         for (i = 0; i < v->n_fmt; ++i) {
             fmt_aux_t *z = &fmt[i];
-            if ( !z->size ) {
+            if ( z->size==-1 ) {
                 need_downsize = 1;
                 continue;
             }
@@ -3274,7 +3275,7 @@ static int vcf_parse_format_gt6(kstring_t *s, const bcf_hdr_t *h, bcf1_t *v,
     if ( need_downsize ) {
         i = 1;
         while ( i < v->n_fmt ) {
-            if ( !fmt[i].size )
+            if ( fmt[i].size==-1 )
             {
                 memmove(&fmt[i-1],&fmt[i],sizeof(*fmt));
                 v->n_fmt--;