Fossil SCM

This check-in has been fuzz tested against varying Markdown inputs.

A 5-days test in single-threaded mode on x86_64 machine with Linux didn't found any issues.
This test constituted of three parallel executions:

1. Fast fuzz testing (with -fsanitize=fuzzer) was started with test/markdown_test3.md, src/markdown.md and www/grep.md as seed inputs.

   After 5 days it had 264697 seeds occupying 1135 MB and its output showed:

   #285802794 REDUCE cov: 6015 ft: 42892 corp: 8354/7018Kb lim: 8110 exec/s: 672 rss: 107Mb L: 1077/8110 ...

2. "Slow" fuzz testing (with -fsanitize=fuzzer,undefined,address) was started with the same seed inputs.

   After 5 days it had 102401 seeds occupying 445 MB and its output showed:

   #83632513 REDUCE cov: 17483 ft: 124381 corp: 6222/7629Kb lim: 8110 exec/s: 199 rss: 959Mb L: 2984/8088 ...

3. "Very slow" fuzz testing (with -fsanitize=fuzzer,undefined,address) was started with all Markdown files from the project's source tree.

   After 5 days it had 76499 seeds occupying 799 MB and its output showed:

   #37437954 REDUCE cov: 17884 ft: 128434 corp: 5472/39Mb lim: 47603 exec/s: 89 rss: 1019Mb L: 18603/47474 ...

Fuzz testing with initial seeds taken from the final state of the first item was continued in "slow mode".
After 6.5 days it had 298196 seeds occupying 1278 MB and its output showed:

 #74365450  REDUCE cov: 17784 ft: 128605 corp: 7689/5253Kb lim: 8110 exec/s: 132 rss: 1043Mb L: 1257/8047 ...

Also a multi-threaded fuzz-testing has been running for a few days on a Raspberry Pi 4 machine. This test also succeeded.