AUIS revival

Timeline
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

87 check-ins using file revival/doc/mail-quickstart.md version 1de82d9883

2026-09-28
15:13
WIP snapshot: Economist newsletter rendering (preheader chars, zero margins, wrapper/div colors, 600px column, link colors, column widths); not yet visually approved Leaf check-in: c894a84876 user: wdc tags: html-economist-wip
05:29
docs: roadmap item 10 resolved, new items 11-13; revival.md and design doc cover the table-layout work Leaf check-in: cebe27d9b8 user: wdc tags: html
05:28
htmlatk: valign/height/cellpadding, rowspan side columns, float runs, centered and shrink-to-fit tables, floated-image paragraphs, div line breaks, left-justified cells check-in: 36b73f69ce user: wdc tags: html
05:28
htmlpart: keep valign/height/cellpadding attributes and line-height/clear/float/margin styles check-in: ee0de0a0af user: wdc tags: html
05:27
lset/lsetv: leaf valign, halign, minheight, padding placement for HTML table cells (lset \V 8) check-in: a232e8403c user: wdc tags: html
03:17
docs: roadmap item 9 (lsetscrollview 4px clip) resolved check-in: 0b7b15c67d user: wdc tags: html
03:17
lsetscrlc/lsetscrlv: share one minwidth safety pad so a fitting tableau isn't clipped 4px on the right check-in: 99fd1423e0 user: wdc tags: html
03:03
httpimg.c/text822.c/htmlatk.c: remove always-on /tmp/httpimg-trace.log image tracing; scroll.c: mark elevator-drag note resolved check-in: 9d50f98dcb user: wdc tags: html
02:18
roadmap.md/revival.md/html-mail-rendering-design.md: log bookrack table-layout fixes (rowspan grid, empty columns, right-fixed cells), tableau row stacking, and remaining open items check-in: 1551773f72 user: wdc tags: html
01:07
htmlatk.c: honor rowspan column occupancy, zero-width empty columns, and trailing fixed-width cells; fixes bookrack Book Reviews crushed text, clipped starred cover, missing right margin check-in: 5d9b7447e7 user: wdc tags: html
01:06
lsetv.ch/lsetv.c: add lsetview_MakeHorzFixedRight, a split whose right child gets a fixed pixel width (lpair_HFixed) check-in: 9e9ff1ce73 user: wdc tags: html
2026-09-26
19:43
New htmlprv tool: standalone graphical driver for htmlatk.c, no mailbox/network needed check-in: a7315893f7 user: wdc tags: html
19:42
lsetscrlv.c/lsetscrlc.c: tableau tables no longer reflow wider than their own design width check-in: 9ef09eac4e user: wdc tags: html
19:15
revival/tests/bookrack.html: cache 2 more images missed by an uppercase SRC= attribute check-in: 1d06387b9e user: wdc tags: html
05:57
revival/tests: cache bookrack.html's images locally, stop re-fetching over the network check-in: 2d6e2660fd user: wdc tags: html
05:45
roadmap.md: close item 2's two remaining follow-ups (spacer-drop eagerness, cell tiling seams) with today's fixes check-in: 07f449a466 user: wdc tags: html
05:43
htmlatk.c: set noseam on every HTML-table split, closing the last 1-2px seam/border artifacts around bookrack.html's book-review cards check-in: a5cdfd562f user: wdc tags: html
05:42
matte: add noborder opt-in flag -- matte always reserves a 1-2px picture-frame margin and draws an outline around its embedded child; HTML-table content (detected by dataobject class) now skips both, matching a real browser's borderless table rendering check-in: fa1178afe1 user: wdc tags: html
05:42
lset/lpair: add noseam (lpair_NOSEAM) opt-in bit -- lpair always reserves a 1px seam between a split's two children even with nobar set; HTML-table splits now skip it so adjacent cells touch exactly, matching a real browser check-in: 82ea99e0e0 user: wdc tags: html
05:12
htmlatk.c: keep a blank spacer row when dropping it would fuse differently-colored or independently-bordered rows together (bookrack.html: FEATURED TITLES gap, book-review card divider); htmlpart.c: stop stripping border-bottom from style= check-in: 3823290ecd user: wdc tags: html
03:27
roadmap.md/porting-assessment.md: close item 2 (background-color, now fixed) with its two live follow-ups, and log a new possibly-latent NodeIsVisuallyEmpty/bgcolor gap. check-in: 68e5c881f3 user: wdc tags: html
03:21
lsetv.c: paint a leaf's resolved background color via SetBackgroundColor (after LinkTree, not before -- fixes a live null-colormap crash), and make split DesiredSize honor its own composed minwidth instead of only echoing children's sizes. check-in: dc27abd246 user: wdc tags: html
03:20
htmlatk.c: resolve+cascade a cell's background color (style/legacy attr, td->tr->table fallback) and propagate it recursively through spliced rows and embedded nested views; htmlatktest.c dumps it. check-in: 0d8067d500 user: wdc tags: html
03:20
htmlpart.c/lset.ch/lset.c: keep bgcolor (table/td/th/tr) and add a persisted lset.bgcolor field (\V6) to hold it. check-in: abf89473c8 user: wdc tags: html
2026-09-25
22:12
revival.md: document the scrollable-tableau HTML table feature and the xcolormap color-table capacity fix/open follow-up. check-in: fbe761d0e8 user: wdc tags: html
21:56
roadmap.md: log the DecodeWeight click/drag rescaling bug (h) found regression-testing click scroll. check-in: 901e339f9e user: wdc tags: html
21:55
textv.c: DecodeWeight was rescaling click positions meant for drag only, zeroing real click offsets inside tall embedded views; click/right-click scroll into a tableau now lands correctly instead of warping to its top. check-in: 1a05ac2fbe user: wdc tags: html
19:15
roadmap.md: log the lsetview.c DesiredSize guard and the observable short->int overflow crash found in tonight's regression testing. check-in: 7d17c3da49 user: wdc tags: html
19:14
observe.c/.ch: widen nObservers/maxObservers short->int, check malloc/realloc for failure; 16-bit counter overflowed live at 27310 colormap observers, crashing AddObserver. Full-tree rebuild. check-in: 32cbb86fb2 user: wdc tags: html
19:12
lsetv.c: guard lsetview__DesiredSize's leaf case against recursing into an unlinked child (view_GetIM NULL); same lazy-linking hazard already fixed in matte.c/lsetscrlc.c, found chasing a live click-scroll crash into Book Rack. check-in: 44c0c7b5ee user: wdc tags: html
17:33
html-scroll-plan.md: addendum pointing at the six bookrack-testing bugs found in Steps 2/3's own subsystems since this plan's 'done' status, per roadmap.md. check-in: 1c902a7348 user: wdc tags: html
17:28
roadmap.md: log the bookrack bug chain (LinkTree dictionary crash, nested-table double-scrollbar, matte.c drag-scroll crash, lsetview/lsetscrollcontent h-scroll redraw, captions FINESCROLL sibling-site collapse) and the .ch install-propagation gotcha found chasing the last one. check-in: 9936e0906a user: wdc tags: html
17:20
textv.c/textv.ch: EncodePosition/DecodePosition macromethods still hardcoded the pre-widening FINESCROLL shift (7); promoted to shared textview_FINESCROLL so subclasses (messages/captions) decode correctly again. Fixes captions scrollbar collapsing to whole-document. check-in: da17ba914b user: wdc tags: html
16:54
lsetv.c/lsetscrlc.c: fix Book Rack tableau horizontal-scrollbar redraw; lsetview__FullUpdate forwarded a possibly-0x0 rectangle straight to real content (lpair.c has an identical pre-existing guard lsetview never got), and lsetscrollcontent's own repaint idiom never cleared background first. Both fixed; horizontal and vertical scrollbars confirmed fully functional. check-in: 38b27c2d35 user: wdc tags: html
05:11
matte.c: guard DesiredSize against an unlinked child; recursing into real drawing code (textview__LineRedraw's color allocation) on a not-yet-linked matte child crashed during drag-scroll through deeply-nested HTML tables (Book Rack). check-in: 357be57f2c user: wdc tags: html
05:11
htmlatk.c: only the outermost table embedding routes to lsetscrollview; a nested table (rendered via BuildLsetCell's cell-ambient recursion) was independently wrapping itself too, producing nested double scrollbars in deeply-nested tableaux (Book Rack). check-in: 78b6b00b3d user: wdc tags: html
05:11
textv.c: fix LinkTree/FinalizeObject crash on textviews with recorded scroll weights; the per-object dictionary loop treated the ScrollWeightsKey blob as a viewref and dispatched view_LinkTree/Destroy on garbage. check-in: 6b77e50892 user: wdc tags: html
04:43
roadmap.md/html-scroll-plan.md: elevator-drag fully fixed, close out Step 3. check-in: 2d2be23208 user: wdc tags: html
04:41
scroll.c/textv.c: fix elevator-drag-into-endzone landing at wrong position; set_frame() int/long truncation, DoUpdate self->lines[-1] OOB read, setframe() off-reset-to-0 past document end. Elevator-drag scrolling now fully correct. check-in: 1f71b5b4c9 user: wdc tags: html
03:21
roadmap.md/html-scroll-plan.md: elevator-drag mostly fixed, one follow-up bug documented. check-in: 9e29ba62d2 user: wdc tags: html
03:21
textv.c/.ch: weight oversized containsView lines by real height for elevator-drag scrollbar position; fixes drag through national-grid.html's table. One follow-up bug still open, see html-scroll-plan.md. check-in: 9155a296e5 user: wdc tags: html
2026-09-24
04:21
roadmap.md: close national-grid.html whitespace bug (root cause + fix), open new item for the elevator-drag scrolling bug found immediately after, still in progress check-in: 7cc6444132 user: wdc tags: html
04:21
scroll.c: fix what_is_at()'s int/long truncation (LP64); WIP elevator-drag fix in HandleThumbing, not yet correct (see its own comment for full status/trail) -- fixes click-driven scrolling accuracy alongside textv.c's FINESCROLL widen, drag still needs work. Debug tracing kept in for continued investigation check-in: 0a5cb73005 user: wdc tags: html
04:21
textv.c: widen scrollbar elevator's fine-position encoding (FINESCROLL/FINEMASK 7->14 bits) so it doesn't saturate for a view many thousand pixels tall; fixes click/page landing accuracy for messages' new HTML-table tableau feature. Temporary debug tracing (dbglog->/tmp/elevator-debug.log) kept in for ongoing elevator-drag investigation, see scroll.c check-in: a4fc339ca1 user: wdc tags: html
04:20
htmlatk.c: stop icon-cell fixedpx from leaking into BuildLsetGrid's tableau-floor detection; fixes national-grid.html's large-blank-vertical-space bug (unrelated paragraph was being crushed to an 87px icon's own width) check-in: 5dd03fa919 user: wdc tags: html
2026-09-23
17:45
roadmap.md/html-scroll-plan.md: update table-reflow open item (1) with the correct fixed-width-plus-scroll diagnosis and current status; add LinkedIn fixture used to develop/verify it check-in: feeb5a2c2c user: wdc tags: html
17:43
messages: new lsetscrollview/lsetscrollcontent classes give a fixed-width HTML table its own self-contained horizontal scroll (Thunderbird-style), routed from htmlatk.c wherever a table declares a real pixel width; lpair.c/textv.c/drawtxtv.c untouched. Fixes a lazy-linking DesiredSize crash and a chrome-flicker/double-frame bug found live; wdc-confirmed working. National-grid whitespace rendering bug still open (roadmap.md item 1) check-in: 1c31299679 user: wdc tags: html
17:40
lset.ch/.c: add minwidth field, computed bottom-up by htmlatk.c, to carry a table's declared pixel width through nested lset/lpair splits check-in: 4d9de95d28 user: wdc tags: html
17:40
scroll.ch/.c: add per-instance SetDrawBorder/GetDrawBorder, immune to InitPrefs's ScrollDrawBorders profile default via new drawborderoverridden flag; needed by messages' new lsetscrollview class check-in: 8ee69bf028 user: wdc tags: html
2026-09-18
03:03
cid: embedded-image resolution: Content-ID parsing in mimepart.c, sibling-part scoping in text822.c's resolver; confirmed against a real inbox message check-in: 7e4863587c user: wdc tags: html
2026-09-16
03:05
Link-hover cursor: htmllinkview posts Cursor_Gunsight over table-cell links, gated by new htmlatk_TextHasLink() so non-link cells are unaffected check-in: 1d646a5d64 user: wdc tags: html
02:35
roadmap.md: scope the cursor-on-hover open item with the PostCursor mechanism found live (no core-ATK change needed for images; text-link rectangle geometry still unscoped) check-in: 8956e81a55 user: wdc tags: html
02:22
htmlatk.c/text822v.c: wire up HTML mail link clicks (left=launch, right=copy to X cut buffer), text and table-cell links via new htmllinkview class, image-wrapped links via a Down-instead-of-Up workaround for imagev's own focus-grab eating the button-release; 3 real bugs found+fixed live along the way (wrong view class for cells, cell text defaulting editable, message-line word-wrap hiding the URL); wdc-confirmed working; docs updated (roadmap.md item 1 closed, porting-assessment.md item t., changelog, revival.md) check-in: b2d5334877 user: wdc tags: html
2026-09-14
16:12
roadmap.md/design doc: smart punctuation (item 1) was already fixed 2026-08-17, wrongly carried into yesterday's Open-items consolidation as still-open; correct it and renumber check-in: ac5ff70ce8 user: wdc tags: html
03:19
consolidate HTML-rendering status tracking: roadmap.md's Open items list is now the single canonical source; design doc keeps only design-level reasoning, no duplicate status check-in: 940fe4db91 user: wdc tags: html
03:01
html-mail-rendering-design.md: mark bug r. (forward-paging double-count) done, matching porting-assessment.md item r.; clarify the still-open pixel-accuracy issue is distinct from it check-in: 0a97e66162 user: wdc tags: html
02:54
textv.c: fix forward-paging double-count in MoveForward's last-line cap branch (bug r.), a self-inflicted regression from item h.'s 2026-08-18 fix; root-caused via live lldb against the writeds repro; docs updated, wdc-confirmed fixed check-in: 548bde2f38 user: wdc tags: html
2026-09-12
20:43
correct writeds finding: it works, was a grep-wrapper artifact; writeds is now the standing offline repro for the forward-paging bug (confirmed: text renders, paging asymmetry reproduces) check-in: df60495a5e user: wdc tags: html
04:15
revival.md/porting-assessment.md/porting-changelog.md: log a real ez-vs-srctext/html crash (bug s.) found re-confirming the forward-paging repro; correct an overclaimed writeds-based repro that also turns out broken check-in: 186aa1c303 user: wdc tags: html
03:33
revival.md/porting-assessment.md/roadmap.md: log the forward-paging under-report bug (Book Rack) properly in the durable docs before resuming work on it; roadmap.md date refreshed check-in: d08d4153c6 user: wdc tags: html
03:15
revival.md/porting-changelog.md: document the andrdir.h/Makefile dependency fix, moved out of Open Issues into Old bugs never found till now check-in: 71bfc7f567 user: wdc tags: html
03:14
src/overhead/util/hdrs/Imakefile: andrdir.h now depends on Makefile, fixing the staleness bug where a site.h-only change never triggered regeneration check-in: b792c1f500 user: wdc tags: html
2026-09-03
04:35
porting-changelog.md: note that the amsdemo message d7 (not just the isolated test copy) got the same piano desh fix check-in: 5c533845a4 user: wdc tags: html
04:34
src/ams/demo/d7: apply the same arbiter desh=150 fix to the real 'Mail for your Ears' amsdemo message (piano_isolated.ez was only the lifted-out test copy) check-in: a029698be0 user: wdc tags: html
04:26
revival.md/porting-changelog.md/roadmap.md: arb/piano sizing and settle-transition RESOLVED; corrects earlier child-vs-app theory, documents the real cause (lset-slot vs cel-own trailer field mixup) check-in: 31ca54a759 user: wdc tags: html
04:17
piano_isolated.ez: bound the arbiter's own desh=150 on its actual cel trailer field, fixing both oversized rendering and the settle-transition resize check-in: 22155567af user: wdc tags: html
2026-09-01
05:08
revival.md/porting-changelog.md: arb/piano empty-spacer fix documented, ANDREWDIR staleness bug found+worked around, overall sizing root-caused to lset child-vs-app forwarding bug, still open check-in: e33e18cab3 user: wdc tags: html
05:06
lpair.c: empty lset spacer leaves now clamp to STARTHEIGHT instead of echoing an unbounded offered height check-in: 83c3bc4173 user: wdc tags: html
2026-08-31
03:56
revival.md/porting-changelog.md: arb/piano resize bug re-diagnosed - timing theory disproved, two independent leaf bugs found (unbounded tune textview, echoing empty-spacer lset), both still open check-in: 9e457dab20 user: wdc tags: html
2026-08-21
07:39
xfontd.c: root-cause the X BadValue noise (courier pixel_size=0 has no core-protocol scalable resource; skip that guessing cascade once fontconfig already resolved a font) + fix vertical line spacing (GetFontSummary was a second, missed live consumer of core-only metrics) check-in: 1d3dfaf1d9 user: wdc tags: html
07:35
txtvinfo.h: widen vi_width/vi_height short->long (35yr overflow, arb/piano demo); add repro fixture + doc a second, still-open sizing bug check-in: cba1ea1994 user: wdc tags: html
05:57
xfontd/xgraphic: real fontconfig resolution for Xft text (was antialiasing whatever bitmap core-matched); fixes icon-font substitution + Arial rejection found live; WidthTable/DrawChars alignment now use resolved font's own metrics check-in: b1d7b26a4c user: wdc tags: html
02:46
html-mail-rendering-design: log forward-paging under-report bug (Book Rack, partially diagnosed) check-in: 198c0dd170 user: wdc tags: html
02:09
htmlatk: apply table/td style+align, font-family, text-align; fix align float-vs-justify conflict; document XftFontOpenName gap in porting-assessment.md check-in: 2a2a6b6aa7 user: wdc tags: html
2026-08-20
20:43
httpimg: drop fetch-count cap, make image fetch timeouts configurable (fixes Book Rack cover/photo drops) check-in: e491528cfb user: wdc tags: html
20:19
add lpair_AUTOHEIGHT: stop destroying rows 1..N in floated-table content leaves (fixes Book Rack synopsis/Buy-link loss) check-in: 06dd42f9c8 user: wdc tags: html
06:14
add lpair_VCENTER: table-cell valign=middle for htmlatk's lset rows (interim; nested-table/href padding case still open) check-in: 566ffcacc8 user: wdc tags: html
05:17
docs: record icon-clipping root cause+fix (item q., porting-assessment.md; html-mail-rendering-design.md); log background-color gap check-in: eb1b1ab510 user: wdc tags: html
05:13
htmlatk.c: pad fixed icon-column width by 6px (textview's embedded-border reservation shorts it); also cap it to icon-scale to avoid starving row siblings; round image_Zoom's percentage calc instead of truncating; htmlatktest.c: add writedsr mode for local-file-backed ez repro check-in: e0eb4f3946 user: wdc tags: html
03:31
docs: log divider-bar/fixed-icon-column/image_Zoom fixes and new lset-width-reflow open bug check-in: 7250f45dce user: wdc tags: html
03:31
lset/lpair: suppress divider bar (nobar bit on movable, not a struct field); htmlatk.c: fixed-pixel icon column for floated-pair splits; image.c: round instead of truncate in image_Zoom's buildZoomIndex check-in: e641a29205 user: wdc tags: html
01:23
roadmap.md: mark scrollbar endzone bug resolved (was already fixed 492e9b6e24, doc was stale) check-in: dd64d20bf4 user: wdc tags: html
01:22
docs: record the floated-adjacent-tables fix (National Grid issue 1 of 5) check-in: 67743d63d3 user: wdc tags: html
01:21
htmlatk.c/htmlpart.c: side-by-side layout for align=left/right floated table pairs (National Grid gas-meter icon+text) check-in: 440265b65d user: wdc tags: html
2026-08-19
22:40
design doc: record image bevel/scaling/Zoom-bug fixes; log 5 new National Grid header layout issues to investigate next check-in: ced00b09cd user: wdc tags: html
22:39
image.ch/imagev.c/htmlatk.c: suppress imagev's bevel frame + scale HTML images to declared width=/height=; image.c: fix image__Zoom uninitialized-pointer bug for truecolor images check-in: 563523f25f user: wdc tags: html
20:20
docs: recover the A/B/C/D image-fetch design rationale (blocking curl chosen over async/in-process/disk-cache), note the curl prerequisite, and document caching's per-render-pass scope check-in: da46e58bd7 user: wdc tags: html