Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(132)

Issue 496873003: Make RenderTextWin fallback match original font heights. (Closed)

Created:
6 years, 4 months ago by msw
Modified:
6 years, 3 months ago
CC:
chromium-reviews, yusukes+watch_chromium.org, yukishiino+watch_chromium.org, tfarina, penghuang+watch_chromium.org, nona+watch_chromium.org, James Su, ckocagil
Visibility:
Public.

Description

ABANDONED: THIS SHRINKS I18N FONTS UNREASONABLY; SEE BUG. Make RenderTextWin fallback match original font heights. Fixes browser UI Hindi text clipping in EN-US locales, etc. (if the system's default font doesn't support Hindi glyphs) (fallback previously yielded text too tall for UI controls) (shouldn't affect Chrome on Hindi Windows installations) See before/after pics at http://crbug.com/404999#c17 BUG=404999 TEST=Chrome --lang=hi on EN-US Win, etc. looks okay. R=sky@chromium.org,asvitkine@chromium.org

Patch Set 1 #

Patch Set 2 : Cleanup; still exploring options. #

Patch Set 3 : Make fallback font heights match the original font height (not size). #

Patch Set 4 : Cleanup. #

Unified diffs Side-by-side diffs Delta from patch set Stats (+9 lines, -27 lines) Patch
M ui/gfx/render_text_win.cc View 1 2 3 3 chunks +9 lines, -27 lines 0 comments Download

Messages

Total messages: 4 (0 generated)
msw
msw@chromium.org changed reviewers: + sky@chromium.org
6 years, 4 months ago (2014-08-25 18:58:54 UTC) #1
msw
msw@chromium.org changed reviewers: + asvitkine@chromium.org
6 years, 4 months ago (2014-08-25 19:08:28 UTC) #2
msw
Hey Scott and Alexei, please take a look; thanks!
6 years, 4 months ago (2014-08-25 19:10:11 UTC) #3
Alexei Svitkine (slow)
6 years, 3 months ago (2014-08-25 22:39:37 UTC) #4
The code itself looks good.

Two questions, though:

1. Back when I implemented the original logic, I distinctly remember non-XP
needing to derive based on size, otherwise the results wouldn't match the
original GDI code. Is this not the case? Have you checked it doesn't result in
incorrect size in other contexts? e.g. CJK languages

2. The non-XP path had a performance optimization where it wouldn't do anything
if the requested values already matched. Have you checked that this doesn't
cause a noticeable performance regression?

Powered by Google App Engine
This is Rietveld 408576698