Convert Crossword Compiler ([url removed, login to view]), currently a VCL program in Delphi 2007, to a Delphi XE3 VCL project. There are about 82 forms and 100K lines of code. This is a fairly large and reasonably complicated piece of software. However this project is not to fully unicode-enable everything, but to reproduce all existing functionality perfectly in unicode-enabled XE3 ready for future more general unicode support. (a previous project involving full unicode conversion proved too complex). File formats etc should remain the same as now, and internal algorithm code can remain as-is (so some types will need to be converted to be explicity AnsiChar with charset code), but all gui should use unicode controls, with consistent conversion from unicode-ansi(in the correct charset) as required. Some thought will be required to do this in an intelligent and future-upgradable way.
Some parts of the software are already unicode enabled, so they should of course simply be converted to use native XE3 unicode functions. Some of the code is very well written, some of it is a bit more complicated/old-fashioned (dating back 20 years). There are many custom string classes (lists of words, etc) and a some assembler routines. Some current dependence on JEDI/TNT for unicode support should be removed to use native support.
Once complete there should be no implicit ansi-unicode conversion compiler warnings, and you should thoroughly test for consistency with the original version (sample puzzle and word list files provided). I expect regular progress reports, with sharing of code via git/svn repository; payment once new code fully compiling and fully tested (by us both). Testing is key; be prepared to spend 50% of your time testing: this is a 99.9% debugged, robust and widely-used code, I would like it to remain that way. There may be a followup project to implement more general unicode support and new features subject to mutual agreement.
Please give examples of previous delphi work.
17 freelancers are bidding on average $2345 for this job
hello, I`m more concerned about 3rd party components, if there are many, the conversion could be difficult, otherwise,I believe an intelligent approach can be applied thank you, Ionel