Quote:
Originally Posted by ashkulz
That helps a lot, as it looks the ebook crashed just after trying to access the eBookwise photo. Assuming that it needs to decode the whole picture in memory, then it requires roughly 0.5M of memory (a back of the envelope estimate). Can you tell me how much internal memory and how much free memory on the SM card was present? Because then it begins to look like the low memory issue is the root cause which may be turning up when a site has a lot of images, not when browsing mobile-friendly sites which anyway do not use images that much. |
It doesn't use the 8 MB internal storage (or external SM) as internal memory or swap area. The internal memory on the EBW1150 is 4 MB for both code (builtin reader app) and data per Garth's email to me and info posted
here. It's running out of internal (4 MB) memory when it loads that .gif image at the end of surfing; but not when I surf there first since I passed that point before (see my EBW1150 album)! I believe, the 4 MB internal memory is the culprit!
Quote:
|
One quick test: can you create a new HTML in your content and include the image and see if it crashes if you navigate locally? If not, then does adding 2 or 3 different large images causes it to crash? If it does crash, then we need to benchmark the maximum size of the images allowed and if they are encountered, then do some handling in impserve for that.
|
I don't need to do that because the content directory I uploaded with Linreb-0.2 had a large picture (REBTD-fullpage.png @ 0.48 MB) followed by a smaller sized file of the same image (REBTD-fullpage256.png @ 0.13 MB) on the second page of the next.html which would crash the EBW1150 sometimes. The follower image (REBTD-fullpage256.png) would render without problems. So it now seems that I
have seen this behaviour before...
A solution (for large images crashing the reader) could be for a plugin to 'convert' all images (over a certain size) to 16 or 256 colors, if possible. :nudge: :nudge:
Quote:
That's a typo which has already been fixed yesterday but not put in the archive uploaded here.
Hmm, I don't think there's a way in which you can rollback to an older version of the firmware. If we could verify that, we can then report a regression to ETI...
That ties in well with the analysis for nrapallo above. Can you try to remove all books from the internal storage and move them to the SM card? I suspect that the 1150 may be using the internal storage as swap space if it runs out of memory (which is more likely to happen in a large IMP file).
I'd advise you to do the same, as we need to verify the swap space theory too...
|
The (limited) 4 MB internal memory seems to be the culprit with reader resets.
