Up to Scribe...
I'm interested to know if this recent build (16-9) is crashing at all? So far I havn't had any crash reports. Maybe this is the start of some stability (finally!). Changes:
- Fixed some of the mail fields not loading. This was causing the alternate HTML not to show up at all.
- Made the calendar events sub-hour accurate in the week view.
- Rounded new events in the weekly calendar view to the nearest hour.
- Fixed the printing of mail and contacts. The codepage was still the old code.
- Fixed the crash when copying to the clipboard.
- Implemented temporary fix for multipart alternative charset. I'm trying to fix this properly, but it's going to make a mess of the code so I'm trying to find a way to do it cleanly.
- Fixed selection of the right ReplyTo info and .Sig file for the incomming account when replying and forwarding.
- Added code to read in the old "codepage" integer stored in old email and convert it to a new style "charset" string so that old email should display in the right charset.
- Added hard coded file extension -> MIME type mappings to override the busted or missing registry settings in more Windows installs.
- .mbox or .mbx -> text/mbox
- .vcf -> text/x-vcard
- .vcs -> text/calendar
- Fixed replying to messages with accents or unicode type charsets.
- Fixed saving messages with the right codepage. They get converted to utf-8 to allow all combinations of charsets.
- Added Base64 Rfc2047 encoding for utf-8 charset strings to be more efficent.
- Made the default LGI fixed width font Courier 9pt instead of 10pt. This effects the internet headers, by making the text smaller its easier to read because it wraps less.
- Fixed the status LED's to go red when the connection fails.
- Fixed charset conversion to skip over bad utf-8, outputting '?'s instead and continues on instead of ending the conversion. This is handy when a badly formed email put unmarked charset stuff in the headers and I then try and use it as utf-8. Also messages that Scribe can't determin the charset of are now assumed to be utf-8. But don't fear, I'm not ignoring the old messages, their codepage is accounted for. So it's really an error case when the charset of a received mail defaults to utf-8.
- Changed the implementation of the "Date:" header not to use the C library functions and instead use my own GDateTime object because on various localized windows the C library functions seem to use local day/month names with accents and so on which breaks the mail rfc's stipulation that headers are to be 7bit. So for the time being the date header will be strictly english (and thus 7bit).
- Added the windows version number to the "X-Mailer:" header. So far the poll on the site seems to suggest that people don't mind that information being included.
- Changed the Internet Headers to be displayed in the default input codepage, which is windows-12xx on Windows, and us-ascii in Linux. Which might help with some clients sending 8bit chars without the proper encoding. Headers are supposed to be 7bit so it shouldn't matter.
- Fixed the crash in the GTextView3 when the control had been set to empty and the mouse over handler crashs. This was caused by the layout information which points into the text buffer not being cleared after the text buffer was free'd.
- Fixed the wrap & quote code to handle unicode (properly).
- [Possibly?] Fixed the "GHtml.cpp, Line:1140" crash.