lp:~pygame/pygame/trunk

Created by Thomas Kluyver on 2013-01-08 and last modified on 2015-05-03
Get this branch:
bzr branch lp:~pygame/pygame/trunk

Branch merges

Related bugs

Related blueprints

Branch information

Owner:
PyGame
Project:
Pygame
Status:
Development

Import details

Import Status: Reviewed

This branch is an import of the HEAD branch of the Git repository at git://github.com/illume/pygame.git.

The next import is scheduled to run in 2 hours.

Last successful import was 3 hours ago.

Import started 3 hours ago on pear and finished 3 hours ago taking 20 seconds — see the log
Import started 9 hours ago on pear and finished 9 hours ago taking 20 seconds — see the log
Import started 15 hours ago on russkaya and finished 15 hours ago taking 40 seconds — see the log
Import started 21 hours ago on pear and finished 21 hours ago taking 20 seconds — see the log
Import started on 2015-05-27 on pear and finished on 2015-05-27 taking 20 seconds — see the log
Import started on 2015-05-26 on pear and finished on 2015-05-26 taking 15 seconds — see the log
Import started on 2015-05-26 on pear and finished on 2015-05-26 taking 20 seconds — see the log
Import started on 2015-05-26 on russkaya and finished on 2015-05-26 taking 25 seconds — see the log
Import started on 2015-05-26 on pear and finished on 2015-05-26 taking 20 seconds — see the log
Import started on 2015-05-25 on pear and finished on 2015-05-25 taking 15 seconds — see the log

Recent revisions

2402. By Lenard Lindstrom <email address hidden> on 2015-05-03

Fix bug where pygame.color fails the unit test for x86-32 Linux

See changeset dc57da440ac3 for the same problem with pygame.math

For gcc 4.4 to at least gcc 4.8.2, -O2 optimization reduces the precision of
floating point arithmetic on doubles, compared to unoptimized code. This caused
two test cases in color_test.py to fail.

2401. By Lenard Lindstrom <email address hidden> on 2015-05-02

Fix a bug in sysfont.py exposed by commit ae2e36d

For Python 3.x, raw font information was incorrectly converted back to byte
strings. This caused incorrect parsing of the information, which showed up
as failed unit tests. This corrects the bug, Raw data is not converted to
unicode strings for both Python 2.x and 3.x.

2400. By Lenard Lindstrom <email address hidden> on 2015-04-30

Decapitate zombie bug 52 (Motherhampster Bugzilla)

It came back. A gcc 4.4 floating point optimization changes the precision
of floating point arithmetic for doubles. The altered precision slightly
changes the results of calculations, showing up as unit test fails. The bug
fix, however, was quite specific in its application, only enabled for gcc
version 4.4. Guess what? The problem still exists with gcc 4.8.2. This commit
broadens the scope of the original fix.

2399. By Lenard Lindstrom <email address hidden> on 2015-04-30

A possible solution to Issue #131

In surface.c the surf_convert function uses memcpy to copy the source format
to the target format. This includes the palette pointer. If the target format
has a non-NULL palette pointer, the palette is checked to ensure at least one
palette color is not black. It does not matter if the target pixel format
actually uses a palette or not. So unless the target has a palette, make sure
the palette pointer is NULL.

2398. By Jason M. Marshall on 2015-04-28

Removed hard-coded font lists. They're not necessary, and updating them every time that a new OS is released would be a burden. Now, sysfont will depend on the registry (on Windows) or X11 fc-list (on any other OS) to generate a system's font list.
Ad-hoc tested sysfont changes on Windows 7 with Python 2.4 and 3.4. Tested on Mac OS X 10.4 with Python 2.7.
Close #179.
Processed with autopep8.

2397. By Lenard Lindstrom <email address hidden> on 2015-04-23

Reindent alphablit.c

Make alphablit.c control structures easier to follow. This will help when
converting it to SDL 2.

2396. By Lenard Lindstrom <email address hidden> on 2015-04-14

Close Issue #243

Issue #243: For a string with trailing spaces, freetype ignores the last space
            in boundary calculations

After the layout loop, determine if the last character in the text string
is a space by checking if its bitmap is width 0. If so, update the text
boundaries — min_x, max_x, min_y, max_y — with the current pen position, which
already includes the space's advance.

Also update change the Font.render_to example in the freetype module doc page to
reflect the removal of this "feature".

2395. By Lenard Lindstrom <email address hidden> on 2015-04-10

Close Issue #273

Issue #273: Memory overrun when text without descenders is rendered with
            underlining

When the fractional top and bottom of an underlining fit within a single row,
the difference between the line bottom and the lower pixel boundary, a
negative number, was misinterpreted as meaning there was a second row to
fill. If the line was already at the bottom row of the surface, then the
next row overwrote invalid memory. This problem was fixed in ft_render_cb.c.

To avoid rounding problems with underlines having an odd thickness value,
changes were also made in ft_render.c.

2394. By Lenard Lindstrom <email address hidden> on 2015-04-08

Close Issue #242

Issue #242: freetype.Font.get_rect() ignores style defaults when
            the style argument is not given

Make the default value of the style argument be FT_STYLE_DEFAULT, as
documented for get_rect().

2393. By Lenard Lindstrom <email address hidden> on 2015-02-24

msg.txt

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar repository format 2a (needs bzr 1.16 or later)
This branch contains Public information 
Everyone can see this information.