fuzz: improve CI iteration strategy, add corpus minimization and summary reporting

[joostjager]

Fuzzing has become increasingly important with the recent wave of changes: async persist, channel manager refactors, splicing, and zero-fee channels. These are complex state machine changes where the fuzzer is one of our best tools for catching edge cases.

This PR gives the fuzz CI some overdue attention. The main goals are visibility into what the fuzzer is actually doing, and making the iteration budget more meaningful:

• The old fixed 30-second wall clock limit meant fast targets wasted time while slow targets barely got through their corpus. Iteration counts now scale with corpus size, so every target replays its full corpus with room for new mutations.
• On main, corpus minimization runs after each target to prune redundant inputs and keep the cache from growing indefinitely. PRs can opt in via the fuzz-minimize label.
• A summary table is printed at the end of each run, giving visibility into per-target corpus sizes, deltas, and timings:

  ==== Fuzz Summary ====
  Target                                     Iters  Corpus     Fuzzed       Fuzz time    Minimized     Min. time
  ------                                     -----  ------  --------------- ---------  --------------- ---------
  base32_target                              27216    3277   3286 (+9)           0m4s    346 (-2940)        0m4s
  bech32_parse_target                        91160   11270  11285 (+15)          0m4s    959 (-10326)       0m8s
  bolt11_deser_target                        99144   12268  12273 (+5)           0m5s   1288 (-10985)      0m26s
  chanmon_consistency_target                 11904    1363   2531 (+1168)       6m56s   1908 (-623)       36m44s
  chanmon_deser_target                      150112   18639  18675 (+36)          0m8s   1250 (-17425)      0m15s
  ======================
  

• The fuzz sanity check (cargo test on fuzz targets) is split into its own parallel job since it doesn't depend on the corpus.

Builds on #4496, so the fuzzer actually passes.

[ldk-reviews-bot]

👋 Thanks for assigning @TheBlueMatt as a reviewer!
I'll wait for their review and will help manage the review process.
Once they submit their review, I'll check if a second reviewer would be helpful.

[codecov]

Codecov Report

❌ Patch coverage is 83.66013% with 25 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.21%. Comparing base (7d82e14) to head (b7fd027).
⚠️ Report is 6 commits behind head on main.

Files with missing lines	Patch %	Lines
lightning/src/ln/channel.rs	83.14%	14 Missing and 1 partial ⚠️
lightning/src/ln/channelmanager.rs	80.00%	6 Missing and 4 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4497      +/-   ##
==========================================
+ Coverage   86.19%   86.21%   +0.01%     
==========================================
  Files         161      160       -1     
  Lines      107459   107514      +55     
  Branches   107459   107514      +55     
==========================================
+ Hits        92621    92688      +67     
+ Misses      12219    12208      -11     
+ Partials     2619     2618       -1     

Flag	Coverage Δ
tests	86.21% <83.66%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.