Version 05.180a --------------- 1) Re-added the extension filtering to the Open dialog. It now uses the associated file extensions from the Settings dialog 2) Added an "Extract to..." option to the Explorer context menu. This allows you to right click a VMARC in Explorer and extract it, saving 1 entire click (yippee! ;-D) You'll have to go back into Settings and click Update to activate this change. 3) Corrected the extracted date being off by 1. I oopsed and didn't notice the month was 0-based. (The month in the VMARC header is 1-based.) You may still notice that the hour is off by one. This relates to daylight savings time and how your system is handling it. Right now, my system is set to Central Time with automatic adjustment for DST and extracted files are always one hour off. However, if I turn off the DST setting, the extracted files have the same date as the ones in the VMARC. 4) Changed include to allow compiling under OSX. Version 05.178a --------------- 1) Added user selectable character maps. For command line users, the "-u" option is used to specify the path names to the UCM files. For GUI users, change them in the Settings dialog. I chose to use UCM files for this as not all platforms have EBCDIC conversion tables. This a shame since iconv() is quite easy to use. But, my guess is that this will be a little used feature so having to acquire the correct UCM won't be too much of a hassle for folks. The UCMs can be obtained at: http://dev.icu-project.org/cgi-bin/viewcvs.cgi/charset/data/ucm/ 2) I recreated the internal conversion tables. They now represent an IBM-1047 to ISO-8895-1 mapping. This can be easily overidden using UCMs. 3) I added a new "-t" option to the command line version to allow translation without appending line feeds. The "-a" option still functions as before, but internally it turns on "-t" and causes linefeeds to be appended to text records. 4) A record length problem introduced in 05.174a was corrected. 5) The "Format" registry setting in VMAWin was not being handled correctly. 6) The CMS version now restores the date/time of the extracted files. It uses the same method as VMARC. This was the first instance where I had an issue of running a z/OS compiled module under CMS. The program needs to determine at runtime where it is executing and the uname() function is perfect for this. However, the utsname structure it uses is not the same size between the 2 systems, so when the module executed on z/VM, the stack would get overlayed. Took quite a while to figure out why, but a workaround was fairly simple. 7) Added the first time users message that suggests they setup the file associations. Version 05.174a --------------- Geez, what didn't change...let's see... VMARC file processing separated from UI processing Added subfile type (Text or Binary) guesstimate The first 1024 bytes are examined for characters < 0x20 and if any are found, it is considered binary. It doesn't always guess right, but it gets it most of the time. Fixed several problems with ASIS processing Added VMAWin - a little GUI interface Fixed several problems found by those brave souls that tried it early Disabled file time setting under CMS and z/OS until a workaround is found Better (more) error detection New distribution layout Version 05.167b --------------- Fixed filtering (missed a period) - Reported by Alex Brodsky Version 05.167a --------------- Added subfile filtering. Added -c option to convert subfiles names to lowercase Added -m option to allow filemode specification Added -s option to allow a slightly slow method of skipping subfiles Added setting of last access and modification times (this does not yet work for standard CMS files and z/OS datasets) Added a little summary of how many subfiles were located and/or bypassed Changed how to locate subfiles so as not to be locked into a static record length and to support self-extracting VMARCs