For the docking...
I've seen these warnings, some of which might be significant. If anyone can give them a look, that would be helpful
libgdl/gdl-dock-object.c:99: warning: unused parameter 'data' libgdl/gdl-dock-object.c: In function 'gdl_dock_object_get_type': libgdl/gdl-dock-object.c:99: warning: missing initializer libgdl/gdl-dock-object.c:99: warning: (near initialization for 'object_info.value_table')
libgdl/gdl-dock-object.c: In function 'gdl_dock_param_get_type': libgdl/gdl-dock-object.c:833: warning: missing initializer libgdl/gdl-dock-object.c:833: warning: (near initialization for 'tinfo.base_init')
libgdl/gdl-dock-master.c: In function 'gdl_dock_master_get_type': libgdl/gdl-dock-master.c:129: warning: missing initializer libgdl/gdl-dock-master.c:129: warning: (near initialization for 'object_info.value_table')
(several more of those)
libgdl/gdl-dock-item.c:276: warning: assignment from incompatible pointer type
libgdl/gdl-dock.c: In function 'gdl_dock_dock_request': libgdl/gdl-dock.c:847: warning: comparison between signed and unsigned libgdl/gdl-dock.c:850: warning: comparison between signed and unsigned libgdl/gdl-dock.c:854: warning: comparison between signed and unsigned libgdl/gdl-dock.c:857: warning: comparison between signed and unsigned
libgdl/gdl-dock-item-grip.c: In function 'gdl_dock_item_grip_expose': libgdl/gdl-dock-item-grip.c:147: warning: assignment from incompatible pointer type libgdl/gdl-dock-item-grip.c:150: warning: passing argument 2 of 'gdk_draw_rectangle' from incompatible pointer type libgdl/gdl-dock-item-grip.c:138: warning: unused variable 'item_or_child_has_focus'
libgdl/gdl-dock-paned.c:568: warning: comparison between signed and unsigned libgdl/gdl-dock-paned.c:573: warning: comparison between signed and unsigned libgdl/gdl-dock-paned.c:577: warning: comparison between signed and unsigned
On Sat, Mar 29, 2008 at 11:08:44PM -0700, Jon A. Cruz wrote:
libgdl/gdl-dock-item.c:276: warning: assignment from incompatible pointer type
This looks like a legitimate bug: there shouldn't be a gpointer argument.
Btw, where does this source come from? I don't see this issue in gdl either v0.7.11 or svn head.
libgdl/gdl-dock.c: In function 'gdl_dock_dock_request': libgdl/gdl-dock.c:847: warning: comparison between signed and unsigned libgdl/gdl-dock.c:850: warning: comparison between signed and unsigned libgdl/gdl-dock.c:854: warning: comparison between signed and unsigned libgdl/gdl-dock.c:857: warning: comparison between signed and unsigned
These are inside a `if rel_x > 0 && rel_y > 0 && ...' block, so treating rel_x and rel_y as unsigned happens to be the right thing here.
To address this warning while at the same time reducing the chance of int overflow, I suggest changing `rel_x = x - a->x; if rel_x > 0 && rel_x < a->width' to `if x > a->x && x < a->x + a->width { guint const rel_x = x - a->x;'.
(This one OTOH is in upstream, at least in v0.7.11.)
libgdl/gdl-dock-item-grip.c: In function 'gdl_dock_item_grip_expose': libgdl/gdl-dock-item-grip.c:147: warning: assignment from incompatible pointer type libgdl/gdl-dock-item-grip.c:150: warning: passing argument 2 of 'gdk_draw_rectangle' from incompatible pointer type
The above two look fixable by changing bg_style to type GdkGC*.
libgdl/gdl-dock-paned.c:568: warning: comparison between signed and unsigned libgdl/gdl-dock-paned.c:573: warning: comparison between signed and unsigned libgdl/gdl-dock-paned.c:577: warning: comparison between signed and unsigned
Much the same problem/solution as above.
pjrm.
On Mar 30, 2008, at 5:29 AM, Peter Moulder wrote:
This looks like a legitimate bug: there shouldn't be a gpointer argument.
Btw, where does this source come from? I don't see this issue in gdl either v0.7.11 or svn head.
I think the main difference is that I'm setup to build with fewer warnings disabled.
When our codebase was 'dirtier' we needed to have only the worst showing, but now that things are cleaned up a lot it's probably a good idea to move on.
In general the best working solution I've seen is an iterative process where code is cleaned up to generate zero warnings, then bump the settings up to show a few more. Even in the Inkscape sources I've seen things where for a given 'annoying' warning, there are a few legitimate bugs hiding that need to be fixed.
On Sun, Mar 30, 2008 at 11:56:18AM -0700, Jon A. Cruz wrote:
On Mar 30, 2008, at 5:29 AM, Peter Moulder wrote:
This looks like a legitimate bug: there shouldn't be a gpointer argument.
Btw, where does this source come from? I don't see this issue in gdl either v0.7.11 or svn head.
I think the main difference is that I'm setup to build with fewer warnings disabled.
When our codebase was 'dirtier' we needed to have only the worst showing, but now that things are cleaned up a lot it's probably a good idea to move on.
Given the 0.47 cleanup goals, would it make good sense to max up the warning levels, at least up until release time?
Bryce
On Sun, Mar 30, 2008 at 01:43:32PM -0700, Bryce Harrington wrote:
When our codebase was 'dirtier' we needed to have only the worst showing, but now that things are cleaned up a lot it's probably a good idea to move on.
Given the 0.47 cleanup goals, would it make good sense to max up the warning levels, at least up until release time?
It sounds like a good idea. Whenever I do some tweaks or changes in the code I look out for any warnings in the area that can be fixed so I suppose having those visible will draw more attention
-- Marcin Floryan http://marcin.floryan.pl/ [GPG Key ID: 0D5581C5]
participants (4)
-
Bryce Harrington -
Jon A. Cruz -
Marcin Floryan -
Peter Moulder