147453 - Page content not rendered consistantly in non-standard DPI settings

Status: RESOLVED DUPLICATE of bug 134942

(Core :: Layout, defect)

Description

[Zac Jacobson]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.0+)
Gecko/20020525
BuildID:    2002052508

Windows 2000 lets the user set the Font Size to Normal (96dpi), Large(120dpi),
or "Other...". When set to 113dpi (I haven't tried other settings), many things
are rendered... oddly.

Table borders: the bottom or right border is often not rendered to the correct
width. See Bug 142048 which I submitted demonstrating this behaviour before
playing with my dpi settings.

Images: scrolling often results in a white line, 1 pixel high being placed
diagonally across an image where the image is rendered accross the edge of the
viewable page. If the image is a link, clicking on it restores it to it's
intended image.

Fonts: text is often rendered shrunk by one pixel in the middle, this is most
noticable in italic fonts. Selecting the text corrects it, I assume because it
gets re-rendered.

Reproducible: Always
Steps to Reproduce:
1. In Windows: Open display properties. chose the "Advanced" tab. Click on the
"Advanced" button.
2. Under the "General" tab in the new dialog, there is a "Display" section,
containing a "Font Size" selection box. Choose "Other".
3. Drag the ruler until it's set at 118% normal size (113 dpi).
4. You may need to install some fonts, and you'll definitely have to restart.
5. Check out http://www.slashdot.org/ there's lots of italics and pictures, as
you scroll, you should notice some of the behaviour I describe.
6. Check out the test case in Attachment 82206 [details] (Bug 142048)... the top table is
rendered without a bottom border.

Expected Results:  I would have expected Gecko would not assume a font dpi when
rendering. I would have hoped that it would behave the same at 113dpi as at 96dpi.

I am using a Samsung SyncMaster 950p monitor on a 32MB ASUS AGPV7700 (nVidia
GeForce2 GTS)

Comment 1

[Boris Zbarsky [:bzbarsky]]

This is a duplicate of the various other "rounding errors at some DPI values" bugs.

Comment 2

[Zac Jacobson]

Boris: You are right... my bug searches were apparently in the wrong direction.
Bug 134942 (single pixel rounding error tracking bug) seems to be an
all-encompasing place to put this... so I guess I'll do that.

*** This bug has been marked as a duplicate of 134942 ***