mbishop@crc.ac.uk (Martin Bishop) (01/26/91)
I came across a problem with fasta when running against small datasets when it is possible that there are no scores higher than the cutoff. Its in fffasta.c in routine showbest() when bbp is a pointer to NULL and then the next bit of code tries to print out the results. I cured this as below: /* This is the place where trouble occurs if no scores greater than cutoff! */ else $ if (bbp == NULL) return; /*fprintf(stderr,"Heading for trouble\n"); fprintf(stderr,"(*bbp).score %d hcutoff %d\n",(*bbp).score,hcutoff); fflush(stderr); */ if ((bbp->score > (2*hcutoff)) || (bbp->score > hcutoff && bbp->gscore > (4*bbp->score/3))) $ istart=istop; nshow += 10; if (istart<mshow) goto l1;
wrp@biochsn.acc.Virginia.EDU (William R. Pearson) (01/28/91)
In article <12746.9101261447@crc.ac.uk> mbishop@crc.ac.uk (Martin Bishop) writes: >I came across a problem with fasta when running against small datasets >when it is possible that there are no scores higher than the cutoff. >Its in fffasta.c in routine showbest() when bbp is a pointer to >NULL and then the next bit of code tries to print out the results. >I cured this as below: > This problem should not occur with fasta version 1.5, since there is no "cutoff". It will be fixed in version 1.4. Thanks for the report. If any of you know about other problems with fasta, please tell me. It seems silly for people to know about bugs for months or years and not tell the author of the software. Bill Pearson