... |
... |
@@ -65,11 +65,10 @@ |
65
|
65
|
you <a href="">"https://lists.nongnu.org/mailman/listinfo/freetype-devel">subscribe</a>
|
66
|
66
|
to it.</p>
|
67
|
67
|
|
68
|
|
- <p>Note that starting with 2021 the time available for
|
69
|
|
- students has been halved. Most of our projects are
|
70
|
|
- modular, more or less, allowing students to propose a
|
71
|
|
- subset of the mentioned targets if this appears
|
72
|
|
- appropriate.</p>
|
|
68
|
+ <p>Note that starting with 2022 there are half-time and
|
|
69
|
+ full-time projects. Most of our projects are modular,
|
|
70
|
+ more or less, allowing students to propose a subset of the
|
|
71
|
+ mentioned targets if this appears appropriate.</p>
|
73
|
72
|
|
74
|
73
|
<p class="warning">Before contacting us, however, you should
|
75
|
74
|
get <em>really</em> acquainted with the topic you would
|
... |
... |
@@ -80,52 +79,6 @@ |
80
|
79
|
project, I want to contribute, please tell me what to
|
81
|
80
|
do!” again and again…</p>
|
82
|
81
|
|
83
|
|
- <dl>
|
84
|
|
- <dt>Develop a test framework for checking FreeType's
|
85
|
|
- rendering output</dt>
|
86
|
|
-
|
87
|
|
- <dd>
|
88
|
|
- <p>Right now, FreeType's rendering results of the
|
89
|
|
- current development version are not systematically
|
90
|
|
- compared to a baseline version, using continuous
|
91
|
|
- integration (CI) or something similar. This is
|
92
|
|
- problematic, since rendering regressions can be very
|
93
|
|
- easily missed due to subtle differences.</p>
|
94
|
|
-
|
95
|
|
- <p>However,
|
96
|
|
- a <a href="">"https://github.com/fundies/freetype2/tree/greg-gsoc">GSoC
|
97
|
|
- 2020 project</a> now provides a good framework that
|
98
|
|
- can act as a starting point for better integration.
|
99
|
|
- In addition, we are moving to the gitlab instance
|
100
|
|
- of <a href="">"https://gitlab.freedesktop.org/freetype">freedesktop.org</a>;
|
101
|
|
- this implies that access to
|
102
|
|
- its <a href="">"https://docs.gitlab.com/ee/ci/">CI
|
103
|
|
- abilities</a> should be straightforward.</p>
|
104
|
|
-
|
105
|
|
- <p>The idea is to select a representative set of
|
106
|
|
- reference fonts from font corpora (which already exist
|
107
|
|
- mainly
|
108
|
|
- for <a href="">"https://github.com/freetype/freetype2-testing/tree/master/fuzzing">fuzzing</a>).
|
109
|
|
- The fonts are used to produce glyph images for various
|
110
|
|
- sizes and rendering modes (anti-aliased, B/W, native
|
111
|
|
- hinting, auto-hinting, etc.). FreeType can already
|
112
|
|
- produce MD5 checksums of glyph images as part of its
|
113
|
|
- debugging output; these values should be compared
|
114
|
|
- against a baseline version of rendering results. If
|
115
|
|
- there are differences, HTML pages should be generated
|
116
|
|
- that contain comparison images of the baseline's and
|
117
|
|
- the current development version's rendering result,
|
118
|
|
- ideally indicating how large the differences between
|
119
|
|
- the images are by using some yet to be defined
|
120
|
|
- measure.</p>
|
121
|
|
-
|
122
|
|
- <p><em>Difficulty:</em> medium. <em>Requirements:</em>
|
123
|
|
- C, HTML, Unix build tools. <em>Potential
|
124
|
|
- mentors:</em> Werner Lemberg, Alexei Podtelezhnikov,
|
125
|
|
- Toshiya Suzuki (FreeType).</p>
|
126
|
|
- </dd>
|
127
|
|
- </dl>
|
128
|
|
-
|
129
|
82
|
<dl>
|
130
|
83
|
<dt>Update FreeType's build systems</dt>
|
131
|
84
|
|
... |
... |
@@ -399,7 +352,7 @@ |
399
|
352
|
<!-- ************************************************** -->
|
400
|
353
|
|
401
|
354
|
<div class="updated">
|
402
|
|
- <p>Last update: 28-Oct-2021</p>
|
|
355
|
+ <p>Last update: 24-Nov-2021</p>
|
403
|
356
|
</div>
|
404
|
357
|
</div>
|
405
|
358
|
</div>
|