Bug in srfile 1.0.6

[pepi37]

I was found bug in srfile 1.0.6 , but it looks like trivial one.

Inside abcd file there in lines...

...
...
ABCD 4973*2^$a+1 [1448313]
ABCD 4975*2^$a+1 [1448184]
4
60
2
10
14
6
52
26
12
10
12
12
24
...
...

After removing some factors, those two lines are converted to one
..
ABCD 4973*2^$a+1ABCD 4975*2^$a+1 [1448184]
4
60
2
10
14
58
26
22
24
24
..
and I got warning message

d:\PGEN\ABCMAKE\MAKESIEVE>sr sr_2.abcd -k factors.txt -G
srfile 1.0.5 -- A file utility for srsieve.
ERROR: Line 389: Malformed line in ABCD format file `sr_2.abcd'.

Content of sieve file for candidate 4973 is

2000000:P:1:2:257
4973 1448313

[rogue]

That is likely happening because there are no values for the first sequence. How did that file get created?

[pepi37]

Originally Posted by rogue

That is likely happening because there are no values for the first sequence. How did that file get created?

There is (one) value ( as I write it)

2000000:P:1:2:257
4973 1448313

[rogue]

I'm confused. That is a newpgen format. What created the ABCD file?

[pepi37]

Originally Posted by rogue

I'm confused. That is a newpgen format. What created the ABCD file?

Ok
step by step
first I made initial sieve and got many npg files
Then I join them and got ABCD file.
I run sr2sieve, and got few factors
Then I remove those factors from abcd file and got error.

Read again, you say that
That is likely happening because there are no values for the first sequence.
Then I reply to you that is not true and give you content of "problematic K" so you can see there is one value in that sequence.

[Batalov]

Yes, that's a bug.

Apparenly srfile always prints "ABCD 4973*2^$a+1" (a quasi header, no endline), and then when it finds first value in the bitmap it prints " [%d]\n" and then differences ("%d\n") after that.

This misfires in this particular case (apparently Predrag ran "srfile ... -k factors.txt ..."), after "-k" the last value for that series was removed.

The patch is obvious.

*** files.c.orig        2011-11-22 08:16:42.000000000 -0800
--- files.c     2017-05-14 09:36:02.290922900 -0700
***************
*** 865,870 ****
--- 865,871 ----
    abcd_p = p;
    for (count = 0, seq = 0; seq < seq_count; seq++)
    {
+     if(SEQ[seq].ncount == 0) continue;
      fprintf(file, "ABCD %" PRIu64 "*%" PRIu32 "^$a%+" PRId64,
              SEQ[seq].k, base, SEQ[seq].c);
      abcd_n = UINT32_MAX;

[rogue]

Fixed. I hope to release next week with the other changes.