44ccfacbe3a3d4b300f80d48651c77837a4b571e
galt
  Tue Apr 26 11:12:02 2022 -0700
SQL INJECTION Prevention Version 2 - this improves our methods by making subclauses of SQL that get passed around be both easy and correct to use. The way that was achieved was by getting rid of the obscure and not well used functions sqlSafefFrag and sqlDyStringPrintfFrag and replacing them with the plain versions of those functions, since these are not needed anymore. The new version checks for NOSQLINJ in unquoted %-s which is used to include SQL clauses, and will give an error the NOSQLINJ clause is not present, and this will automatically require the correct behavior by developers. sqlDyStringPrint is a very useful function, however because it was not enforced, users could use various other dyString functions and they operated without any awareness or checking for SQL correct use. Now those dyString functions are prohibited and it will produce an error if you try to use a dyString function on a SQL string, which is simply detected by the presence of the NOSQLINJ prefix.

diff --git src/hg/visiGene/knownToVisiGene/knownToVisiGene.c src/hg/visiGene/knownToVisiGene/knownToVisiGene.c
index be292ae..ba176e0 100644
--- src/hg/visiGene/knownToVisiGene/knownToVisiGene.c
+++ src/hg/visiGene/knownToVisiGene/knownToVisiGene.c
@@ -105,31 +105,33 @@
 	    if (secondary)
 		continue;
 	    }
 	else
 	    hashAdd(uniqHash, row[1], NULL);
 	}
     hashAdd(hash, row[0], cloneString(row[1]));
     }
 sqlFreeResult(&sr);
 }
 
 struct hash *keepersForChroms(struct sqlConnection *conn)
 /* Create hash of binKeepers keyed by chromosome */
 {
 struct hash *keeperHash = hashNew(0);
-struct sqlResult *sr = sqlGetResult(conn, NOSQLINJ "select chrom,size from chromInfo");
+char query[1024];
+sqlSafef(query, sizeof query, "select chrom,size from chromInfo");
+struct sqlResult *sr = sqlGetResult(conn, query);
 char **row;
 while ((row = sqlNextRow(sr)) != NULL)
     {
     char *chrom = row[0];
     int size = sqlUnsigned(row[1]);
     struct binKeeper *bk = binKeeperNew(0, size);
     hashAdd(keeperHash, chrom, bk);
     }
 sqlFreeResult(&sr);
 return keeperHash;
 }
 
 void bestProbeOverlap(struct sqlConnection *conn, char *probeTable, 
 	struct genePred *gpList, struct hash *gpToProbeHash)
 /* Create hash of most overlapping probe if any for each gene. Require
@@ -216,73 +218,75 @@
 char **row;
 struct hash *geneImageHash = newHash(18);
 struct hash *locusLinkImageHash = newHash(18);
 struct hash *refSeqImageHash = newHash(18);
 struct hash *genbankImageHash = newHash(18);
 struct hash *probeImageHash = newHash(18);
 struct hash *knownToLocusLinkHash = newHash(18);
 struct hash *knownToRefSeqHash = newHash(18);
 struct hash *knownToGeneHash = newHash(18);
 struct hash *favorHugoHash = newHash(18);
 struct hash *knownToProbeHash = newHash(18);
 struct hash *knownToAllProbeHash = newHash(18);
 struct genePred *knownList = NULL, *known;
 struct hash *dupeHash = newHash(17);
 
-
+char query[1024];
 probesDb  = optionVal("probesDb", database);
 struct sqlConnection *probesConn = sqlConnect(probesDb);
 vgProbes = sqlTableExists(probesConn,"vgProbes");
 vgAllProbes = sqlTableExists(probesConn,"vgAllProbes");
 
 /* Go through and make up hashes of images keyed by various fields. */
-sr = sqlGetResult(iConn,
-        NOSQLINJ "select image.id,imageFile.priority,gene.name,gene.locusLink,gene.refSeq,gene.genbank"
+sqlSafef(query, sizeof query, 
+        "select image.id,imageFile.priority,gene.name,gene.locusLink,gene.refSeq,gene.genbank"
 	",probe.id,submissionSet.privateUser,vgPrbMap.vgPrb,gene.id"
 	" from image,imageFile,imageProbe,probe,gene,submissionSet,vgPrbMap"
 	" where image.imageFile = imageFile.id"
 	" and image.id = imageProbe.image"
 	" and imageProbe.probe = probe.id"
 	" and probe.gene = gene.id"
 	" and image.submissionSet=submissionSet.id"
 	" and vgPrbMap.probe = probe.id");
+sr = sqlGetResult(iConn, query);
 
 while ((row = sqlNextRow(sr)) != NULL)
     {
     int id = sqlUnsigned(row[0]);
     float priority = atof(row[1]);
     int privateUser = sqlSigned(row[7]);
     char vgPrb_Id[256];
     safef(vgPrb_Id, sizeof(vgPrb_Id), "vgPrb_%s",row[8]);
     int geneId = sqlUnsigned(row[9]);
     if (privateUser == 0)
 	{
 	addPrioritizedImage(probeImageHash, id, priority, geneId, vgPrb_Id);
 	addPrioritizedImage(geneImageHash, id, priority, geneId, row[2]);
 	addPrioritizedImage(locusLinkImageHash, id, priority, geneId, row[3]);
 	addPrioritizedImage(refSeqImageHash, id, priority, geneId, row[4]);
 	addPrioritizedImage(genbankImageHash, id, priority, geneId, row[5]);
 	}
     }
 verbose(2, "Made hashes of image: geneImageHash %d, locusLinkImageHash %d, refSeqImageHash %d"
            ", genbankImageHash %d probeImageHash %d\n", 
             geneImageHash->elCount, locusLinkImageHash->elCount, refSeqImageHash->elCount, 
 	    genbankImageHash->elCount, probeImageHash->elCount);
 sqlFreeResult(&sr);
 
 /* Build up list of known genes. */
-sr = sqlGetResult(hConn, NOSQLINJ "select * from knownGene");
+sqlSafef(query, sizeof query, "select * from knownGene");
+sr = sqlGetResult(hConn, query);
 while ((row = sqlNextRow(sr)) != NULL)
     {
     struct genePred *known = genePredLoad(row);
     if (!hashLookup(dupeHash, known->name))
         {
 	hashAdd(dupeHash, known->name, NULL);
 	slAddHead(&knownList, known);
 	}
     }
 slReverse(&knownList);
 sqlFreeResult(&sr);
 verbose(2, "Got %d known genes\n", slCount(knownList));
 
 /* Build up hashes from knownGene to other things. */
 if (vgProbes)