Chris Young
31c08f39be
Basic event loop, only handles keypresses and window closure so far
2017-01-01 14:53:37 +00:00
Chris Young
c40e8197c2
Allow for the prospect of in-window scrollbars
2017-01-01 14:34:04 +00:00
Chris Young
8196f5b602
Basic template for an Amiga core window
...
Drawing and event processing not written yet
2017-01-01 13:28:11 +00:00
Chris Young
86a0f14d0a
Make the documentation reflect reality.
...
ExtMem is enabled by default for now to give it a good workout. Disabled may be a more sensible default.
2016-12-31 14:47:33 +00:00
Chris Young
8af9a8407b
Don't unmap ExtMem immediately as:
...
(a) it is upsetting the gif decoder (although I suspect it is caused by the OS struggling with the excessive remaps)
(b) We are probably going to need to map it back in imminently for display
2016-12-31 14:43:06 +00:00
Chris Young
36b26b861f
Update to test ExtMem on OS4.1FEU1
...
Mostly works, but falls over when there are lots of tiny GIF anims on screen
2016-12-31 14:17:22 +00:00
Chris Young
43bbdc520d
Fix merge error
2016-12-31 14:04:14 +00:00
Chris Young
b7551fec72
NULL iextmem
2016-12-31 14:03:34 +00:00
Chris Young
570212add2
Don't map an extmem area is it is already mapped, just extend the lifetime
2016-12-31 14:03:34 +00:00
Chris Young
adac4d98bb
More logging
2016-12-31 14:03:34 +00:00
Chris Young
982a59744f
Revert "Disable ExtMem in palette-mapped modes"
...
This reverts commit 4db40c3f27d0ade8fc76f57cd383e5ff52ad7c93.
2016-12-31 14:03:33 +00:00
Chris Young
205abcf189
Debug logging
2016-12-31 14:03:33 +00:00
Chris Young
937dce4f2e
Schedule unmapping the extmem object to ensure it releases main memory
2016-12-31 14:03:33 +00:00
Chris Young
493036a920
Documentation
2016-12-31 14:02:40 +00:00
Chris Young
a83afda791
Disable ExtMem in palette-mapped modes
...
Something handling palette mapped images causes a freeze.
2016-12-31 14:01:29 +00:00
Chris Young
0804c43bf1
Allocate uncompressed bitmap data in extended memory.
...
This currently isn't working correctly - it simply freezes at some point after loading the page.
2016-12-31 14:01:29 +00:00
Vincent Sanders
48f868a955
move tree compatability layer to amiga frontend
2016-12-31 01:04:27 +00:00
Chris Young
76673bbf65
add a todo
2016-12-31 00:58:03 +00:00
Chris Young
5e523b608a
Simply the event loop for the shared message port
...
This now simply calls the event callback for each window, which returns TRUE is the window was closed during event processing.
The window loop now restarts if any windows were closed, which fixes a potential rare issue with delayed event processing.
2016-12-31 00:51:05 +00:00
Chris Young
9df30eb78f
Update local history window to use event callback
2016-12-31 00:43:06 +00:00
Chris Young
d8eec1b048
Update print window to use event callback
2016-12-31 00:37:59 +00:00
Chris Young
d9c8d1c70c
Update 401 login window to use event callbacks
2016-12-31 00:30:42 +00:00
Chris Young
d90e82d3f1
Simplify closing of windows on quit
...
If they have a close callback it will be called.
Some windows are attached to browser windows and closed elsewhere without using the callbacks.
2016-12-31 00:14:44 +00:00
Chris Young
4893a9f9af
Update download window to use event callbacks
2016-12-31 00:11:06 +00:00
Chris Young
6ba5edaede
Update GUI options window to use event callback
2016-12-31 00:02:21 +00:00
Chris Young
47379c0444
Update treeviews to use event callback
2016-12-30 13:06:42 +00:00
Chris Young
9379a64c6d
Start tidying up Amiga event-handling code
...
As we used a shared message port, we call different event-handling functions for different window types.
This changes that so we register the event-handler as a callback so we can handle event processing for all windows generically.
Currently the main browser window and find window use the new approach, so the handling still has a long list of window types.
This should also make window_list private to gui.c
2016-12-30 00:59:12 +00:00
Chris Young
9d77dc83bd
Never warn the user about multiple tabs being open when we're in the process of quitting.
2016-12-15 18:23:52 +00:00
Chris Young
f04103398b
Avoid leaking memory when typing
2016-12-11 23:04:13 +00:00
Chris Young
ba0895121b
Ensure we close all opened resources when we early exit
2016-12-11 16:16:11 +00:00
Chris Young
9d694f56bb
Tidy-up init a bit
2016-12-11 16:04:16 +00:00
Chris Young
433418a8e9
Finalise bitmap code last as purging the cache needs it
2016-12-11 15:24:56 +00:00
Chris Young
34a587849c
Move as much closedown as possible into gui_quit
2016-12-11 15:18:08 +00:00
Chris Young
68e8152f3e
Ensure OpenURL list is always freed
2016-12-07 00:00:54 +00:00
Chris Young
8f9a652774
Revert "Ensure we free the pen list memory"
...
This reverts commit 74d7c2d2c2
2016-12-06 23:58:47 +00:00
Chris Young
ecf2fdf612
Free the correct memory pointer
2016-12-06 23:39:46 +00:00
Chris Young
74d7c2d2c2
Ensure we free the pen list memory
2016-12-06 23:39:04 +00:00
Chris Young
a440907b78
Fix memory leak
2016-12-06 23:34:24 +00:00
Chris Young
c0a15de1e4
Ensure pool init returns non-NULL so it isn't treated as an error.
2016-12-06 23:23:01 +00:00
Chris Young
2b6a416970
Replace custom itempool functions with malloc/free as clib handles these more efficiently than our pools.
2016-12-06 23:17:40 +00:00
Chris Young
f38b72a583
Avoid potential memory leak
2016-12-06 17:54:58 +00:00
Chris Young
2dd97b0b8e
Fix brokenness caused by recent changes
2016-12-05 23:32:55 +00:00
Chris Young
4b91c43003
Add missing nsoption_finalise
2016-12-05 23:08:17 +00:00
Chris Young
2fe884be65
Fix memory leak searching for resources
2016-12-05 20:30:48 +00:00
Chris Young
71afaacc80
Fix last tab title not being freed
2016-12-05 20:22:21 +00:00
Chris Young
2a9e89de75
Fix potential memory leak
2016-12-05 19:31:04 +00:00
Chris Young
54e31b65d1
Guard against crashing when opening the splash screen
...
Add extra logging for when things go wrong at this stage
2016-11-28 19:35:15 +00:00
Chris Young
0bc32aa654
Add more clib2 stats
...
JSON formatted data can now be saved with ARexx "SLABSTATS stats.json"
2016-11-27 17:10:03 +00:00
Chris Young
4fb38f574a
AmigaOS low memory handler attempt 2
...
Neither freeing slabs nor purging the cache are safe to call from interrupts, so instead we schedule these to run on NetSurf's process.
If NetSurf is busy, there may be significant delay, so the priority is set low to ensure this is called only when other - faster - memory handlers have been exhausted.
I don't know if this works, or is sane, as I'm not entirely sure how to test it.
Something similar needs adding for OS4 but low memory is less of an issue there, so will see if this works on OS3 first.
2016-11-27 16:25:27 +00:00
Chris Young
a76a36915c
Ensure the memory for the mem-handler interrupt is allocated
2016-11-27 15:36:30 +00:00
