A couple of fixes & tweaks #8250
Conversation
| @@ -381,6 +390,10 @@ function KoptInterface:renderOptimizedPage(doc, pageno, rect, zoom, rotation, re | |||
|
|
|||
| local cached = DocCache:check(hash, TileCacheItem) | |||
| if not cached then | |||
| if hinting then | |||
| Device:enableCPUCores(2) | |||
There was a problem hiding this comment.
Is that all that's actually useful in theory or just all that's actually useful in current practice?
There was a problem hiding this comment.
It's already barely useful here, as we're heavily single-threaded ;p.
It's mainly for the sake of the eInk driver itself, which is fairly terrible (read: resource intensive ;p).
* Make sure we have a BB to measure in getSize, in case the instance is recycled. (fix koreader#8241) * nil `line_num_to_image` early in `:free` * Hide the _renderText calls that are used across the whole module to simply update the text layout & instantiate the inner bb behind a wrapper function with a slightly less obscure name.
cores * Only keep a single core online most of the time. * Device: Add an enableCPUCores method to allow controlling the amount of online CPU cores. * Move the initial core onlining setup to Kobo:init, instead of the startup script. * Enable two CPU cores while hinting new (e.g., cache miss) pages in PDF land. * Enable two CPU cores while processing book metadata. * Drive-by fix to isolate the DocCache pressure check to KoptInterface and actually apply it when it matters most (e.g., k2pdfopt stuff).
NiLuJe
force-pushed
the
master
branch
2 times, most recently
from
a9c1041
to
1950ad1
Compare
|
(Now includes UB cleanup in |
"UB" = undefined behaviour ? |
* Don't run multiple collectSubprocesses tasks in parallel. UIManager:scheduleIn doesn't return anything, using that as a boolean trap made no sense. * Make the in-place removal of collected pids use a slightly more common idiom (table.remove in a reverse iteration of said array).
| @@ -1122,6 +1134,7 @@ function TextBoxWidget:free(full) | |||
| self.cursor_restore_bb:free() | |||
| self.cursor_restore_bb = nil | |||
| end | |||
| self.line_num_to_image = nil | |||
There was a problem hiding this comment.
Bad idea: thumbnails in Wikipedia results and full articles are not shown after this (the room for the image is still there, but not image displayed, and long press in the hold select texts around).
Remove this line and all works again.
(Except for the numerous segfault I get when existing KOreader :/ I get segfaults with or without this line. Somehow, I can't get any when running under gdb...)
Turns out we can't release line_num_to_image early, so, delay it until the final, full widget teardown. (Regression since koreader#8250, c.f., koreader#8250 (comment))
Turns out we can't release line_num_to_image early, so, delay it until the final, full widget teardown. (Regression since #8250, c.f., #8250 (comment))
The second argument is a ddjvu_render_mode_t Try to actually honor the user settings instead of enforcing COLOR while we're there. Fix koreader#8376 Regression since koreader#8250
(The 2 core switch is actually probably mostly overkill still, but, oh, well ;)).
collectSubprocesses.This change is