Small output readability improvements

[sim642]

Our messages contain various redundant information which makes the messages long and harder to read. Based on some option we could try to produce simpler output for some things:

• (0,[0,0]) → 0 — constant integers don't need the constant repeated in every domain.
• (Unknown int([0,8]),[0,255]) → [0,255] — intervals are clearer than equivalent def_exc ranges.
• (Unknown int([-63,63]),[-9223372036854775808,9223372036854775807]) → ⊤ — top intervals are easier recognized as such.
• (alloc@sid:82773@tid:main(#top)) → (alloc@sid:82773@tid:main) — disabled malloc uniqueness counter is excessive.
• (alloc@sid:488@tid:Top Threads(#top)) → (alloc@sid:488) — top thread ID is excessive.
• (alloc@sid:82843@tid:main(#top))[def_exc:0; intervals:[0,0]] → (alloc@sid:82843@tid:main)[0] — covered by above.
• (alloc@sid:82832@tid:main(#top))[def_exc:Unknown int([-63,63]); intervals:[-9223372036854775808,9223372036854775807]] → (alloc@sid:82832@tid:main)[⊤] — covered by above.
• mhp:{tid=POOL_thread; created=All Threads} → mhp:{tid=POOL_thread} — top created threads set cannot exclude anything anyway?
• mhp:{tid=main; created=All Threads}, thread:main → thread:main — thread ID repeated by threadid and mhp analyses.
• mhp components should be displayed separately because they are queried from other analyses anyway (Small output readability improvements #1312 (review)).
• PR Small output readability improvements #1312 didn't do it for IntDomain to_yojson, so it looks uglier in abstract debugger.