diff --git a/README.md b/README.md index ced16ca..983b362 100644 --- a/README.md +++ b/README.md @@ -1,86 +1,96 @@ # MultiPar -v1.3.1.4 is public +v1.3.1.5 is public - This version has many changes. + This version has some trial functions. +I fixed some problems of previous versions. Though I tested the behavior on my PC, there may be a bug in other environments or in a rare case. Be careful to use this version for daily usage. When you want a stable version, you should not use this new one until other users try. +Unless you have a problem, v1.3.0.7 or v1.3.1.3 might be safe. If you see a strange behavior, odd problem, or failure, please report the incidinet to me. I will fix as possible as I can. - I improved MD5 hash calculation. -It becomes faster on my PC. -But, it may be slower on some CPUs. -If you feel that hash calculation becomes slow, -please report your case. + I increased max block size from 1 GB (old versions) to 2 GB (new version). +This change is for rare case. +Normally users should not set so large block size in most case. +For compatibility, setting less than 100,000,000 bytes = 95 MB is good. +Be careful, MultiPar cannot treat too large block size properly. +When you have a set of PAR2 files with more than 2 GB block size (such like 4 GB), +par2cmdline may support them. +I adjusted some GUI components for big numbers. +If you see something bad in your language UI, please let me know. - I implemented a function to calculate MD5 hash of multiple files at creating PAR2 files. + I implemented a function to calculate MD5 hash of multiple files at verifying source files. It seems to be faster on SSD. (I cannot test the speed by myself.) But, it is slow on HDD. It detects your drive type and switches function automatically. If it fails and happens to be slow on your PC, please report your case. Because I don't know SSD's property so much, I will need help of users. +Thanks John L. Galt for tests and bug report on SSD. - I tried to fix a location bug of MultiPar window on multiple monitors. -But, I cannot test the behavior on my PC. -There is a blur window problem on different DPI setting between multiple monitors. -At this time, I cannot solve this yet. -When someone has this trouble and want to help me, I will try to solve. + There was a bug in v1.3.1.4, and created PAR2 files happened to contain broken packets. +MD5 hash of some packets were wrong. +When a file size was multiple of block size, the problem occured. +Such PAR2 files are shown as damaged on MultiPar verification, +and v1.3.1.4 could not verify source files without checksum packets. +I fixed the bug in v1.3.1.5, and it can verify source files now. +Though it's possible to use such broken PAR2 files (created by v1.3.1.4), +you would better recreate new PAR2 files with this new version. +I'm sorry for the inconvenience. +Thanks nutpantz for this bug report. + + When you created many PAR2 files (more than 512), verification had failed in previous versions. +I (and many users) didn't see this bug for long time. +Normally people don't create so many PAR2 files. +Thanks Martin Klefas-Stennett for finding this rare problem. -[ Changes from 1.3.1.3 to 1.3.1.4 ] +[ Changes from 1.3.1.4 to 1.3.1.5 ] GUI update - Change - - An option "Don't search subfolders" is added for verification and reapir. + - Max block size is increased to 2,118,123,520 bytes. (1.97 GB) + - Max split size is increased to 2,147,287,040 bytes. (1.99 GB) - Bug fix - - It's possible to adjust opening window position on multiple monitors. - -PAR1 client update -- Improvement - - MD5 hash calculation becomes slightly faster. + - Memory allocation failure in verifying over than 512 PAR2 files was fixed. PAR2 client update -- New - - It's possible to set file access mode for debug usage. - - Change - - When source files are on SSD, hash calculation may become faster. + - Max slice size is increased to 2,147,483,644 bytes. (2 GB) + - Max split size is increased to 4,294,967,292 bytes. (4 GB) + - When source files are on SSD, verification may become faster. - Improvement - - MD5 hash calculation becomes faster on recent CPU. + - Setup of CRC-32 may become slightly faster on recent CPU. - Bug fix - - An access violation bug in restoring single source file was fixed. - -SFV/MD5 client update -- Improvement - - MD5 hash calculation becomes slightly faster. + - A bug in calculating hash of source files on SSD was fixed. + - A bug in verifying source files without Slice Checksum packet was fixed. [ Hash value ] -MultiPar1314.zip -MD5 : 44C31A7F81C6D0C24339D311FD37C10D -SHA-1 : 8871BE03522CF4C6ED96F460E65A4578A0C1D2E3 +MultiPar1315.zip +MD5: DB3661C2AD4D5B6404C7FC8C4CF5AE2B +SHA1: AECCB95F757163E439B9BB6372050D2AB1538D39 -MultiPar1314_setup.exe -MD5: E70D7A2222F99CB903B2A16365D9887E -SHA1: 47A9BDAA489F99067516B160984A9A02352E56BF +MultiPar1315_setup.exe +MD5: ECFC7F69FB5AF168478C24F5C06F025E +SHA1: 66EB941B3E293EB35A48FAD80DD94BC3980DA3E8 [ Hash value of other source code packages ] Old versions and source code packages are available at [OneDrive](https://1drv.ms/u/s!AtGhNMUyvbWOaSo1n_R8awJ_hg0?e=4V0gXu) now. -MultiPar_par2j_1314.7z -MD5: B505C54360170EE14BDC418A62A422C7 -SHA1: B2CAB6C0074E48A7E06973D841F3C88251410ACD +MultiPar_par2j_1315.7z +MD5: FB245B4AFDB45F6C89FFC6231ECA3868 +SHA1: FC32C6F0A88C2750C9A8408525359AED6DBE2B2C MultiPar_par2j_extra_1294.7z MD5: 6D165CDA2645924ACAFE902F02FAD309 @@ -98,9 +108,9 @@ MultiPar_ShlExt_1298.7z MD5: BE0F04DF1A6B936F23F6F01930562248 SHA1: 52818266B45ECE135EECFF12D8DA2640A6AD5075 -MultiPar_ResUI_1314.7z -MD5: 2C6BB0F25A9E95E5C38BA856FA9DC4A6 -SHA1: F4ED32963CD476F5DF722BE6536E116F0FE55A23 +MultiPar_ResUI_1315.7z +MD5: B8B6A9DA4BD9D418CFA90FD01CCC615A +SHA1: E5B4B16DBCAECACA2095A64006C117E04D3C9E74 MultiPar_Help_1314.7z MD5: 5D274F59A5B908B1E31D62CD4F4A0D54