Re: Testsuite changes this week

Hi & thanks for running these weekly tests,

Am 16.09.2014 um 08:49 schrieb Bryce Harrington:
>                    git     git      git       git      git      git
>                   05/03   06/09    07/21    08/04    08/29    09/12
> -------------------------------------------------------------------
> Tests run:        24838   24813    24839    24838    24838    24838
> Passed:           15774   16015    16024    15889    15989    15808
> Failed:            6476    6226     6228     6372     6272     6451
> Expected Failed:    332     315      332      332      332      332
> Error:                1       1        0        0        0        0
> Crashed:             32      33       32       22       22       24
> Untested:          2223    2223     2223     2223     2223     2223
> Total:            24838   24813    24839    24838    24838    24838
> 
> Here's the tests that changed results since the previous run:
>   
>    http://www.bryceharrington.org/Cairo/changes-20140829-to-20140912-trent.txt
> 
> 102 xcb and 70 xlib tests failed here, including a bunch of
> *twin-antialias-* tests.  Anyone have ideas on these?

There weren't many changes to cairo-xcb and git bisect confirms:

e691d242d592a8556e25659fb091a2031abee4c9 is the first bad commit
commit e691d242d592a8556e25659fb091a2031abee4c9
Author: Lukáš Lalinský <[email protected]>
Date:   Wed Sep 3 22:53:55 2014 +0200

    xcb: Initialize font options from Xft resources

    There is a similar code in the Xlib backend. The logic here is the same, but
    XCB doesn't support X resources directly, so there is some custom code
    to get and parse the resources from the root window.

    Signed-off-by: Lukáš Lalinský <[email protected]>
    Reviewed-by: Uli Schlachter <[email protected]>

The attached patch makes these tests pass again. Both hunks are required. The
first one fixes a shameful oversight (whoops), the second one effectively
reverts the above patch, which is not what we want.

Something is merging font options wrongly? Why does this code work for
cairo-xlib? I don't know...

For the xlib backend: Are you running with --enable-xlib-xcb? Aka "are you
testing xlib at all?".

Not much time right now (or in the near future), so can't look too closely into this

Cheers,
Uli
-- 
“I’m Olaf and I like warm hugs.”

-- 
cairo mailing list
[email protected]
http://lists.cairographics.org/mailman/listinfo/cairo

diff --git a/src/cairo-xcb-screen.c b/src/cairo-xcb-screen.c
index 30295af..537e447 100644
--- a/src/cairo-xcb-screen.c
+++ b/src/cairo-xcb-screen.c
@@ -242,6 +242,7 @@ _cairo_xcb_screen_get (xcb_connection_t *xcb_connection,
 
     screen->connection = connection;
     screen->xcb_screen = xcb_screen;
+    screen->has_font_options = FALSE;
 
     _cairo_freelist_init (&screen->pattern_cache_entry_freelist,
 			  sizeof (struct pattern_cache_entry));
@@ -460,7 +461,7 @@ _cairo_xcb_screen_get_font_options (cairo_xcb_screen_t *screen)
 	_cairo_font_options_init_default (&screen->font_options);
 	_cairo_font_options_set_round_glyph_positions (&screen->font_options, CAIRO_ROUND_GLYPH_POS_ON);
 
-	if (! _cairo_xcb_connection_acquire (screen->connection)) {
+	if (0 && ! _cairo_xcb_connection_acquire (screen->connection)) {
 	    _cairo_xcb_init_screen_font_options (screen);
 	    _cairo_xcb_connection_release (screen->connection);
 	}