![]() |
SUBSIM: The Web's #1 resource for all submarine & naval simulations since 1997 |
![]() |
#61 | |
CTD - it's not just a job
|
![]() Quote:
![]()
__________________
"...and bollocks to the naysayers" - Jimbuna |
|
![]() |
![]() |
![]() |
#62 |
GWX Project Director
|
![]()
Thank you, Hebe. Good stuff.
![]() |
![]() |
![]() |
![]() |
#63 |
Commander
![]() Join Date: Aug 2007
Location: Italia Venezia
Posts: 465
Downloads: 228
Uploads: 0
|
![]()
Thank you
![]() |
![]() |
![]() |
![]() |
#64 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
Hi All,
a very profound user has directed me to a Dat-file which crashes my Tool on makeDBs. I can reproduce this, and found the exact place where this ugly Bug is sitting and the reason why it is running wild. The underlying reason is this: i create a buffer in memory and sometimes i read beyond the end of the buffer, which is not allowed. So the Operating-System gives a "NoGo!" with a 0xC0000005 "STATUS_ACCESS_VIOLATION". A simple Programmer mistake. ![]() What bugs me is as big fat this little thing is/was, why it it does not crash always and only on specific files. Hmm, lets see - comes time, comes knowlege. What i like to say is - a bugfix is in progress. |
![]() |
![]() |
![]() |
#65 |
CTD - it's not just a job
|
![]() ![]() ![]()
__________________
"...and bollocks to the naysayers" - Jimbuna |
![]() |
![]() |
![]() |
#66 |
Ocean Warrior
![]() Join Date: May 2012
Location: In the sea, on land and above
Posts: 3,486
Downloads: 878
Uploads: 0
|
![]()
@Hebe Vollmaus
We are waiting for a fix for the latest version, which actually crashes sometimes.
__________________
![]() ![]() |
![]() |
![]() |
![]() |
#67 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
Bugfix need some time as i try to do also other things in it.
40+ GB of 10.300+ different SH-Dat-files on my HD needs around two and a halfe houre to scan. Only 2 crash on a Release-buld. None in Debug-build. ![]() |
![]() |
![]() |
![]() |
#68 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
*Bump*
Update to v2.1e I dont want you stand still in the rain for that long. So here it is. Bugfix #1 - makeDBs - Crash while scan Dat-files. Bugfix #2 - delAllIdx2NewMod - Abort scan after found first Dat-file without index in a drawer. (Filenumber in Library was much too less.) Added some Hints redirected into a file named "DB_idx.csv" when found mystic things as the lots of found ugly files are too much for a Screen-only Message when scan thousends of files. "Abort" here means the file is not copied, as it may be broken afterwards when deleting a index with a defect index-lenght. You may be amused how many files can not read by S3D, or how many files have more then one Index, or the index is not on End of file. And also here, there is no such mystic things in clean vanilla SH3v14. Whats left are these non-ansi Filenames/Drawernames by Cyrillic alphabets writing people like in "War Elite Mod" (WEM). Blucher.dat, NCA_BlБcher, what means Blücher.dat. Heavy cruiser named after "von Blücher". 'Ü' umlaut written by 'Б' or other things like that. If you have some dirs/files named like this you get a Warning "File not found". These files are ignored. This is by Design for now. Greetings. Have fun. -- 550 Mods found (9 activated) |
![]() |
![]() |
![]() |
#69 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
Some words about "==Sort order -- something after idx==".
Most of the files i have seen have this note added in "DB_idx.csv". This means what it says, there is something after the index. Sometimes the index is in the middle of the file, sometimes near the end. But it can also mean that the index-header-lenght is broken, somewhat too short. You can see this in S3D as an Unknown Chunk on the end with unusual huge numbers on Type and Subtype. For example Type: -732102642, Subtype: 1799828. In general all thouse files with these false index-header-lenght can be easily 'repaired' with a Hex-Editor. Naturally there are also other files which S3D will not read, not only those with index-problems. Greetings. |
![]() |
![]() |
![]() |
#70 |
CTD - it's not just a job
|
![]()
Oiy yeah!!! I know 2 modders that are VERY Happy!
![]() Thank you very much Hebe Vollmaus! ![]() ![]()
__________________
"...and bollocks to the naysayers" - Jimbuna |
![]() |
![]() |
![]() |
#71 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
*Bump*
Update to v2.2 The heavy changes are made under the hood. o ProgressBar -- for long run functions This Bar is File-based, which means it moves faster foreward on small files then as for big files. o Threads -- for long run functions To make newer Operating-Systems happy; "...not reponding" Window-title should be gone now. As a consequence Program can be EXIT inbetween long operations now for whatever reason. o New ButtonMenu [No MatFile Check] under [ini menu1024* Renum] Material-check can be de-activated. o New ButtonMenu [Split ini for SH5edit] under [ini menu1024* Renum] Creates a Directory called "_SH3-Menu_forSH5-ED" in the MODS-directory of JSGME in SH3-directory. It will cut the menu_1024_768.ini into pieces. May helpful when try to load parts of SH3-menu.ini into SH5-MenuEditor. With parts i really mean parts. Loading the whole created Dir will not work. Also all related TGAs may better convert to DDS, and also renamed in created files. The "Zone="-line is not injected. For converting TGAs i use "texconv.exe" out of DXsdk_oct2004. You may compare your menu_1024_768.ini with original SH5; maybe some info needed was deleted by the Modder which created your SH3-menu.ini. Also you have to read Documentations/Posts here on Subsim for handling the SH5-MenuEditor. Without it may hard to get any results. o change separate Changelog The history has been growing to much over time, to put everything in one ReadMe. Greetings. Have fun. -- 555 Mods found (12 activated) |
![]() |
![]() |
![]() |
#72 |
Krusty Krab
Join Date: Mar 2002
Location: Decks awash in the North Atlantic
Posts: 1,450
Downloads: 493
Uploads: 0
|
![]()
Thanks Hebe!
![]()
__________________
Good judgment comes from experience. Unfortunately, the experience usually comes from bad judgment. |
![]() |
![]() |
![]() |
#73 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
[Bump]
Update to v2.2a Personaly i was cured long ago from Updateritis; my tool currently not. I must really not assume that all files in all modded SH3-installations around the globe will follow SH3-Filename conventions. In general Names like "Name.ext" (Name-Point-Extension) or for Museum "Name_en.log" are usually 'normal' Files. "Name" (without extension) or "Name.log" has crashed in v2.2. Bug fixed. Maybe i should add a new Function to scan for crap^w'unusual' Files before all. Also remarkable is that a Installation of the above typ may not have Cargo-Nodes on Type=102 ships. So my tool first checks if any Type 102 has these nodes, and if not, the Campaigne-check will nolonger blame these 'CargoExt=-1', since the in-game-Crash will not happen then. Some minor changes; please look into the History. Greetings. Have fun! -- 562 Mods found (100 activated) |
![]() |
![]() |
![]() |
#74 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
*Bump*
Update to v2.2b bugfix: another filename misinterpretation New Tab "HexFloat" under Calc-Button: The upper Editbox is about to find out what the 4-Hex-bytes mean. The lower Editbox is for a positive or negative Floating-Point number. As an example try to find other Hex-Values for a.g.mouse sensitivity patch. Thread: "mouse sensitivity?" post#12 and #16 https://www.subsim.com/radioroom/sho...8&postcount=12 He has been found 1x slower: 0e 74 da 3a (i.e. original value) -- 1/600 Use MS-Calculator: 12x slower -- 600*12=7200 -- 1/7200=1,388888888888889e-4 which is in Exponent-Notation. Put in lower Edit-Box (with Point instead of Comma), push Calc/[Enter] You get: 0x B4A21139 = 0.000139 So B4 A2 11 39 is 12x slower. or 0.5 halfe times slower -- 600*0.5=300 -- 1/300=0,0033333333333333 You get: 0x 0E745A3B = 0.003333 0E 74 5A 3B is 0.5 times slower. Greetings. Have fun! -- 564 mods found (100 activated) |
![]() |
![]() |
![]() |
#75 |
Lieutenant
![]() Join Date: Jun 2010
Location: Germany
Posts: 269
Downloads: 3
Uploads: 1
|
![]()
Ok, when the Eponent-Number comes in, the calculation is wrong.
So there must be another update. ![]() [Edit] Well, the HEX-values are ok. What is not fully correct is the Float-value after the "="-sign. In v2.2b the output is too short, and sometimes zero. The default precision of six is not enough to reinsert this number into the lower Float-editbox to get the same Hex-result. The numbers after the point must be at minimum of eleven to get an identical hex. Here is a resum of a.g. mouse values with there meaning: 05 slower: 0E 74 5A 3B -- 1/300 -- 0,0033333333333333 -- 0.003333333414 1x slower: 0e 74 da 3a (i.e. original value) -- 1/600 -- 0,0016666666666667 2x slower: 0e 74 5a 3a -- 1/1200 -- 8,333333333333333e-4 -- 0.000833333354 3x slower: b4 a2 11 3a -- 1/1800 -- 5,555555555555556e-4 -- 0.000555555569 4x slower: 0e 74 da 39 -- 1/2400 -- 4,166666666666667e-4 -- 0.000416666677 5x slower: 3e c3 ae 39 -- 1/3000 -- 3,333333333333333e-4 -- 0.000333333330 6x slower: b4 a2 91 39 -- 1/3600 -- 2,777777777777778e-4 -- 0.000277777785 7x slower: 34 a9 79 39 -- 1/4200 -- 2,380952380952381e-4 -- 0.000238095236 8x slower: 0e 74 5a 39 -- 1/4800 -- 2,083333333333333e-4 -- 0.000208333338 9x slower: 45 2e 42 39 -- 1/5400 -- 1,851851851851852e-4 -- 0.000185185185 10 slower: 3E C3 2E 39 -- 1/6000 -- 1,666666666666667e-4 -- 0.000166666665 11 slower: 0A E0 1E 39 -- 1/6600 -- 1,515151515151515e-4 -- 0.000151515153 12 slower: B4 A2 11 39 -- 1/7200 -- 1,388888888888889e-4 -- 0.000138888892 e-4 means 10 high -4 -- means move the Comma or Point four times backwards to get a 'normal' dez. If you use "online"-calculators such as https://evanw.github.io/float-toy/ keep in mind, SH3 use reversed Hex-values. [\Edit] Last edited by Hebe Vollmaus; 11-25-24 at 05:11 PM. Reason: Add |
![]() |
![]() |
![]() |
|
|