fix: cache inline image payloads

[kawacukennedy]

What?

Inline cid: images never render. CachedAttachment stored only attachment metadata, so the decoded payload of inline attachments was dropped on cache write and never restored on cache read. inlineImagesFromAttachments skips attachments with empty Data, so the CID lookup always resolved against an empty map.

This PR adds a Data []byte field to CachedAttachment and:

• persists the decoded payload for inline attachments in both cache-write paths (PreviewBodyFetchedMsg, EmailBodyFetchedMsg),
• restores it in both cache-read paths,
• counts it in calculateEmailBodySize so LRU accounting stays accurate.

Why?

Fixes #1702. With the payload cached, the CID lookup in ProcessBodyWithInline resolves and inline images render via the Kitty graphics protocol like remote ones.

[floatpanebot]

Benchmark report — no significant change

Metrics worse: 0 · better: 0 (threshold: ±3%).

benchstat output

goos: linux
goarch: amd64
pkg: github.com/floatpane/matcha/backend
cpu: AMD EPYC 9V74 80-Core Processor                
                           │   old.txt    │               new.txt               │
                           │    sec/op    │    sec/op     vs base               │
ParseSearchQuery_Simple-4    2.559µ ± 28%   2.435µ ± 82%        ~ (p=0.818 n=6)
ParseSearchQuery_Complex-4   8.173µ ± 44%   5.982µ ± 39%        ~ (p=0.065 n=6)
TokenizeSearchQuery-4        4.170µ ± 19%   3.886µ ± 18%        ~ (p=0.900 n=6)
geomean                      4.434µ         3.840µ        -13.41%

                           │  old.txt   │              new.txt               │
                           │    B/op    │    B/op     vs base                │
ParseSearchQuery_Simple-4    26.00 ± 0%   26.00 ± 0%       ~ (p=1.000 n=6) ¹
ParseSearchQuery_Complex-4   762.0 ± 0%   762.0 ± 0%       ~ (p=1.000 n=6) ¹
TokenizeSearchQuery-4        176.0 ± 0%   176.0 ± 0%       ~ (p=1.000 n=6) ¹
geomean                      151.6        151.6       +0.00%
¹ all samples are equal

                           │  old.txt   │              new.txt               │
                           │ allocs/op  │ allocs/op   vs base                │
ParseSearchQuery_Simple-4    2.000 ± 0%   2.000 ± 0%       ~ (p=1.000 n=6) ¹
ParseSearchQuery_Complex-4   23.00 ± 0%   23.00 ± 0%       ~ (p=1.000 n=6) ¹
TokenizeSearchQuery-4        9.000 ± 0%   9.000 ± 0%       ~ (p=1.000 n=6) ¹
geomean                      7.453        7.453       +0.00%
¹ all samples are equal

pkg: github.com/floatpane/matcha/tui
                    │   old.txt    │               new.txt               │
                    │    sec/op    │    sec/op     vs base               │
LogPanelView-4        161.0µ ± 16%   158.6µ ± 14%        ~ (p=0.394 n=6)
SearchOverlayView-4   187.3µ ±  8%   168.5µ ±  8%  -10.05% (p=0.009 n=6)
InboxConstruction-4   895.7µ ± 16%   925.0µ ± 20%        ~ (p=0.589 n=6)
geomean               300.0µ         291.3µ         -2.91%

                    │    old.txt    │               new.txt                │
                    │     B/op      │     B/op       vs base               │
LogPanelView-4        44.67Ki ± 51%   33.23Ki ± 34%        ~ (p=0.675 n=6)
SearchOverlayView-4   33.17Ki ± 69%   56.14Ki ± 41%  +69.25% (p=0.028 n=6)
InboxConstruction-4   874.3Ki ±  0%   874.2Ki ±  0%   -0.02% (p=0.013 n=6)
geomean               109.0Ki         117.7Ki         +7.97%

                    │   old.txt   │              new.txt              │
                    │  allocs/op  │  allocs/op   vs base              │
LogPanelView-4         714.0 ± 0%    713.0 ± 0%       ~ (p=1.000 n=6)
SearchOverlayView-4    923.5 ± 0%    926.0 ± 0%  +0.27% (p=0.041 n=6)
InboxConstruction-4   3.479k ± 0%   3.477k ± 0%  -0.06% (p=0.041 n=6)
geomean               1.319k        1.319k       +0.02%

_{auto-generated by benchmarks.yml}