EMULAB Forum
clrmamepro [English] => clrmame Discussion => Topic started by: Roman on 09 August 2016, 20:35
-
4.030
added: supporting crc32 collisions within sets (equal crc32, different sha1/md5)
added: missing/total chds column for profiler and htmlReport
fixed: having set folders and archives can create wrong missing roms messages (mainly on Linux/ext4)
fixed: unneeded chds are not showing up if nothing of the set exists
fixed: MAME -listxml output was not taking changed temporary folder into account
misc: slight rebuilder speed update for rebuilding separated bios files
misc: updated unrar dll to 5.40.4
misc: compiled with VS Studio 2015 Update 3
-
Hi Roman, cmpro4.030 freeze for 2-3min. when scaning naomigd set ( with crc32 collision) .
After 3min. scaning continue.
Tested with clean cmpro - same result
Tnx:)
-
Hmmm...it doesn't do that here....
Well actually what it does in the scanner is that it internally turns on SHA1 checks for the files in question.
So in that mentioned BIOS set are something around 6 affected files which will be unpacked to memory, SHA1 is calculated and checked against the database. Like turning on the deep checksum sha1 check..but only for these few files.
If it's slow for you you might have an underpowered PC or solid 7z files....but still...2-3 minutes would be way too long.
will do some more testing later at home on a different machine.
By the way...when does it do it? During the "looking for unneeded sets" phase at the start of a scan...or when the scan procedure reaches naomigd when running through the sets?
-
..or when the scan procedure reaches naomigd when running through the sets?
<=== yes
Btw - my sets it's t7z full merge mode.
-
and your naomigd set does not miss any files I guess.....
-
yep, set it's full . Now test with rebuilding naomigd set - not luck.
attach naomigd t7z
-
what do you mean with "not luck"? Rebuilding fails? Or the rebuilt files still cause the slowdown in the scanner?
-
ok...I have a scenario where the slowdown happens...I will look at it later tonight..Thanks
-
ah yes ok....I found the problem....fixed version later tonight