Re: [MiNT] ***SPAM*** Re: Issue with Mint 1.19 and Litchi

[Pierre Ton-That <ptonthat@club-internet.fr>]

- FireBee RAM is about 0.5GB, it was and should be easy to malloc 70MB, even with fragmentation.
- tests were done under MiNT+XaAES (only these have changed), Thing, a few accessories, GBE and Litchi (which called the gemdos(68...)).

Thanks for the CVS URL, I used the diffs yesterday, but found nothing yet. Lonny advised me to search also in MiNT memory parameter (mint.ini may have changed when upgrading). I will spend some precious time to inquire more this week-end.

Cheers,

Raj

Le 05/02/2015 09:54, Adam Klobukowski a écrit :

Are you sure it is not a problem of memory fragmentation?

AdamK

2015-02-04 21:14 GMT+01:00 Pierre Ton-That <ptonthat@club-internet.fr>:

Bonsoir,

- GFA code may be not speakable enough for most of you.
- gemdos(68 in TTRAM, with size of 70MB) was OK in MiNT 1.18. This 70MB file was a one of 'Elansar' game archive.
- the same gemdos(68...) returned negative value in MiNT 1.19, I had to add some code to reduce size until the gemdos(68...) succeeded and download parts instead in one row.

If nobody can anwser this question, where can I find the history (github or svn)?

Cheers

Raj

Le 04/02/2015 20:47, Alan Hourihane a écrit :

Hi,

Sample code speaks volumes. If you have a test case, then it's easier to
determine if it's on purpose or not.

Thanks,

Alan.

On 04/02/15 19:17, Pierre Ton-That wrote:

Hi all,

Speaking of MiNT changes, from 1.18 to 1.19, I noticed one can't
malloc very big blocks. There is a limitation (about 1 MB ?). Could
not malloc 70 MB (MegaBytes) for a very big file for download in one
row. FireBee RAM is large enough.
Is this a feature or a bug? Was it done on purpose?

Cheers,

Raj

Le 03/02/2015 23:07, Vincent Rivière a écrit :

By looking at the source changes between that October version and the
current one, I see:

--

Semper Fidelis

Adam Klobukowski
adamklobukowski@gmail.com