Fixed a lot of bugs and improved overall performance.Report: Updated "Copy & Move" functions, uses Yashied's excellent.Report: Updated "Secure Delete" function, uses wraithdu's excellent.Report: Added several new style and option settings.Redesigned SMF_TNP.exe to create thumbnails using AutoIt native functions / does not use AutoItObject dll anymore.Reworked the overall search loop and underlying SQL statements to improve speed.This second idea includes that dupes must be on the same drive, which must be NTFS formatted. Why not propose, after auto-selecting the 1st duplicate of a file, to hard link the other dupes to the 1st one? (either with a DLL call, either with integrated AutoIt FileCreateNTFSLink() ) Why not list all hard links to a file in a tooltip when hovering the hard link number of a file ( FindFirstFileNameW / FindNextFileNameW), I see 2 things which could be nice, the first could help me with my current duplicates project (as I try a lot, without success, to convert the functions to AutoIt), the second could make it useless, but I don't care I take advantage of this post to propose some ideas about SMF. I'll buy protection against whips, and clean the sources when I will have finished my current project. Well, it makes me realize that I push back this moment since the beginning (cleaning the source code), I'll publish them on my site as they are, and then I'll make a topic here. well, they are messy, and I see you, AutoIt MVP, coming with your whips ^^. I guess it is a shy attitude, this is really an insignificant tool compared to SMF.Īlso the source are a bit messy. Thank you about explaining what was going wrong, it is always interesting to understand things. "UAC Pass" is a really nice piece of software, I've started coding something similar I called "UAC Elevator" (using the "Schedule.Service" object), but development on that one stalled. I've fixed this in the upcoming version 1.9, which I'll release as soon as AutoIt 3.3.8.0 is released.
That's why the hashes were still consistent, but not comparable with the ones created with other calculation methods. The rest of the last chunk was filled with zeros. Now the issue was that all chunks were exactly 1MB large, even the last chunk which is only partially filled with the rest, e.g. I do this to make the calculation interruptable. The hashes are calculated in 1 Megabyte chunks.
I've checked the calculation and quickly found the issue. Thanks for reporting the issue, I didn't saw that!Īnd as you said they're consistently wrong.
#Mediainfo 0.7.61 full
You are absolutely right, full MD5 hashes (full CRC32 and full SHA1 hashes too) are wrong in SMF 1.8.
0 kommentar(er)
