Interlisp Monday meeting notes

[Herb Jellinek]

-------- Forwarded Message --------

Subject:	Recap for "Impromptu Google Meet Meeting"
Date:	Mon, 24 Aug 2026 18:26:06 +0000

Meeting Purpose Review PRs and sync on font rotation, T-Edit, and other projects. Key Takeaways - Font Rotation: Ron's PR fixes a critical bug where the bitmap height was used as character height, but a fundamental design flaw remains: rotating the entire font bitmap instead of individual glyphs. This creates complex issues with metrics and baselines that require a more consistent coordinate system model. - T-Edit Update: Ron is finalizing a T-Edit update that improves menu tracking and fixes a bug preventing multi-object copy-selection. The update also clarifies the logic for undocumented protected, sell-before, and sell-after image object properties. - New Font Strategy: Matt is building a BDF font pipeline, starting with Noto fonts. This requires a new medley-font-sources repo to separate large, generated files from the main source. Future work will explore adding kerning and ligature support. - Student Project: Johans' project was scoped down from a regex engine to a unit test framework. This provides a more achievable 3-month goal and offers a clear path to integrate automated tests into the Medley build process. Topics PR Review & Merges - PR #26991 (Frank): Merged. - PR #2726 (Ron): Approved. Adds font rotation display to the edit-font title bar. - PR #550 (Nick): Refactors C file includes for clarity using include-what-you-use. - Status: Awaiting Cygwin test (Frank) and feedback from LispCore users on unusual OSs (Herb). Font Rotation: Design Flaws & Fixes - Problem: The current font rotation method rotates the entire font bitmap, not individual glyphs. This breaks standard font metrics and creates complex issues. - Example: Using the bitmap height as character height was a critical bug, as a rotated bitmap's height can be thousands of pixels. - Fixes in Ron's PR: - Metrics: widths and y-widths vectors are swapped for 90°/270° rotations. - Offsets: Bitmap offsets are now correctly applied to the vertical dimension for vertical text. - Missing Vector: A zero-filled widths vector is created for vertical fonts, as the code doesn't default to zero if the vector is nil. - Remaining Challenges: - Negative Advance Widths: The 16-bit unsigned metric vectors cannot represent negative values needed for backwards text (e.g., 270° rotation). - Baseline Alignment: Ensuring baselines align when mixing fonts with different rotations (e.g., 90° vs. 270°) is untested. - Coordinate System: A consistent, coordinate-system-independent model for font metrics (ascent, descent) is needed to avoid ambiguity. T-Edit Update - Ron is finalizing a T-Edit update with two main improvements: - Menu Tracking: Corrects menu highlighting behavior. When the mouse moves over an option, it shows the next state (e.g., toggled on) and reverts to the current state when the mouse leaves, before the button is released. - Multi-Object Selection: Fixes a bug preventing copy-selection across multiple image objects. The when-operated-on function now correctly applies to all objects in a selection, not just the one clicked. - Undocumented Features: The update clarifies logic for protected, sell-before, and sell-after image object properties. - sell-before/sell-after: Direct the caret to a specific object. A new rule handles the edge case where two such objects are adjacent, which previously failed. New Font Pipeline & Strategy - Matt is building a pipeline to generate Medley display fonts from BDF files, starting with Noto fonts. - Repo Strategy: A new medley-font-sources repo is needed to house large, generated files (BDFs, TTFs) and utilities (e.g., a modified otf-to-bdf tool). - Rationale: Avoids bloating the main Medley repo with non-source assets. - Future Work: - Kerning: Extract kerning data from TTF fonts. - Ligatures: Implement support for multi-character ligatures (e.g., "ffi"), which poses complex selection and rendering challenges. - Postscript Font Coercion: The fonts-available function is broken for Postscript fonts. - Problem: It fails to report availability for logical font names that are mapped to physical files via the font coercion table. - Solution: The function must correctly resolve logical names through the coercion mechanism. Other Project Updates - Johans (Student): Project scoped from a regex engine to a unit test framework. A proposal is due this Thursday. - Herb: Updated the Zotero link checker to build a full in-memory collection model. Awaiting a response from ACM regarding bot detection issues. - Nick: Investigating the complex history of the mako-ether and 10-megabit Ethernet code to enable proper PUP packet formatting. - Paolo: Debugging a gravity-loops simulator where the first run fails silently, but subsequent runs succeed. Tracing shows state-updating methods are skipped on the first run. Next Steps - Frank: Test PR #550 on Cygwin. - Herb: Post a note to LispCore requesting testers for PR #550 on unusual OSs. - Ron: Finalize and release the T-Edit update. - Matt: Create a medley-font-sources repo for the new font pipeline. - Johans: Submit the unit test framework proposal by Thursday. - Paolo: Continue debugging the gravity-loops simulator. - All: Meet with the Blue Sky user (UNICEA) this Wednesday.

Impromptu Google Meet Meeting August 24, 2026  •  79 mins  •  View Meeting or Ask Fathom Action Items ✨ Merge PR 26991 (Cygwin from SDL to X) Paolo Amoroso Approve PR 2726 (Edit Font) Paolo Amoroso Test PR 550 (include-what-you-use) on Cygwin Frank Halasz Email LispCore re: PR 550 (include-what-you-use) for unusual-OS testing Herb Jellinek Investigate T-Edit copy-select across image objects (DINFO) Ron Kaplan Email ACM re: Zotero link-checker bot-detection approval Herb Jellinek Create public font-sources repo; upload tools + Noto BDFs; link in Herb's XCCS/MCCS issue Matt Heffron Meeting Summary ✨ Enhanced Template Customize   •  Change Template Meeting Purpose Review PRs and sync on font rotation, T-Edit, and other projects. Key Takeaways Font Rotation: Ron's PR fixes a critical bug where the bitmap height was used as character height, but a fundamental design flaw remains: rotating the entire font bitmap instead of individual glyphs. This creates complex issues with metrics and baselines that require a more consistent coordinate system model. T-Edit Update: Ron is finalizing a T-Edit update that improves menu tracking and fixes a bug preventing multi-object copy-selection. The update also clarifies the logic for undocumented protected, sell-before, and sell-after image object properties. New Font Strategy: Matt is building a BDF font pipeline, starting with Noto fonts. This requires a new medley-font-sources repo to separate large, generated files from the main source. Future work will explore adding kerning and ligature support. Student Project: Johans' project was scoped down from a regex engine to a unit test framework. This provides a more achievable 3-month goal and offers a clear path to integrate automated tests into the Medley build process. Topics PR Review & Merges PR #26991 (Frank): Merged. PR #2726 (Ron): Approved. Adds font rotation display to the edit-font title bar. PR #550 (Nick): Refactors C file includes for clarity using include-what-you-use. Status: Awaiting Cygwin test (Frank) and feedback from LispCore users on unusual OSs (Herb). Font Rotation: Design Flaws & Fixes Problem: The current font rotation method rotates the entire font bitmap, not individual glyphs. This breaks standard font metrics and creates complex issues. Example: Using the bitmap height as character height was a critical bug, as a rotated bitmap's height can be thousands of pixels. Fixes in Ron's PR: Metrics: widths and y-widths vectors are swapped for 90°/270° rotations. Offsets: Bitmap offsets are now correctly applied to the vertical dimension for vertical text. Missing Vector: A zero-filled widths vector is created for vertical fonts, as the code doesn't default to zero if the vector is nil. Remaining Challenges: Negative Advance Widths: The 16-bit unsigned metric vectors cannot represent negative values needed for backwards text (e.g., 270° rotation). Baseline Alignment: Ensuring baselines align when mixing fonts with different rotations (e.g., 90° vs. 270°) is untested. Coordinate System: A consistent, coordinate-system-independent model for font metrics (ascent, descent) is needed to avoid ambiguity. T-Edit Update Ron is finalizing a T-Edit update with two main improvements: Menu Tracking: Corrects menu highlighting behavior. When the mouse moves over an option, it shows the next state (e.g., toggled on) and reverts to the current state when the mouse leaves, before the button is released. Multi-Object Selection: Fixes a bug preventing copy-selection across multiple image objects. The when-operated-on function now correctly applies to all objects in a selection, not just the one clicked. Undocumented Features: The update clarifies logic for protected, sell-before, and sell-after image object properties. sell-before/sell-after: Direct the caret to a specific object. A new rule handles the edge case where two such objects are adjacent, which previously failed. New Font Pipeline & Strategy Matt is building a pipeline to generate Medley display fonts from BDF files, starting with Noto fonts. Repo Strategy: A new medley-font-sources repo is needed to house large, generated files (BDFs, TTFs) and utilities (e.g., a modified otf-to-bdf tool). Rationale: Avoids bloating the main Medley repo with non-source assets. Future Work: Kerning: Extract kerning data from TTF fonts. Ligatures: Implement support for multi-character ligatures (e.g., "ffi"), which poses complex selection and rendering challenges. Postscript Font Coercion: The fonts-available function is broken for Postscript fonts. Problem: It fails to report availability for logical font names that are mapped to physical files via the font coercion table. Solution: The function must correctly resolve logical names through the coercion mechanism. Other Project Updates Johans (Student): Project scoped from a regex engine to a unit test framework. A proposal is due this Thursday. Herb: Updated the Zotero link checker to build a full in-memory collection model. Awaiting a response from ACM regarding bot detection issues. Nick: Investigating the complex history of the mako-ether and 10-megabit Ethernet code to enable proper PUP packet formatting. Paolo: Debugging a gravity-loops simulator where the first run fails silently, but subsequent runs succeed. Tracing shows state-updating methods are skipped on the first run. Next Steps Frank: Test PR #550 on Cygwin. Nick: Post a note to LispCore requesting testers for PR #550 on unusual OSs. Ron: Finalize and release the T-Edit update. Matt: Create a medley-font-sources repo for the new font pipeline. Johanns: Submit the unit test framework proposal by Thursday. Paolo: Continue debugging the gravity-loops simulator. All: Meet with the Blue Sky user this Wednesday. View Meeting →