conversion to cmake building system #1

Ok, hold your horses people, It's work in progress...anyway I'd like to have your thoughts about.

Lot has to be done in order to get rid of -UTQT_NO_COMPAT

The po files install isn't good, zh_TW.Big5.po is renamed zh_TW.po.

I've "undef" _HPUX_SOURCE since I don't think we have HP-UX users.

USE_SSL is defined if we have OpenSSL installed, ssl is pretty mandatory these days, tell me if you want It as an option, I'll look into the others packages how to do It.

Thanks for the patch for conversion and the following good patches for cleaning and minor fixes. It's a good job, although I have some comments.

Regarding _HPUX_SOURCE, I think we should not set this up. When someone will build on HPUX, I assume that this definition will be set by the compiler or by some system header. Definition removed from config.h.

Regarding USE_SSL, the original procedure was wrong for two reasons:

With automake it was common that the options had three possible states ON / OFF / AUTO == what is the current state of the system. Because of this, the result of the build was not clearly determined. With CMake we want to have a clearly determined result. So we want to have the option WITH_SSL, which will have ON or OFF values.
For SSL, KSSL from tdelibs is used, not direct SSL library calls. Therefore, there is no need to test the availability of an SSL library or link the SSL library.

There was no option WITH_MLED and test for HAVE_MLED. At the same time, it had a misleading description in config.h.

When I made edits, I also solved the last part of code blocking to remove -UTQT_NO_COMPAT.

Patches rebased to the current head.

It seems we are ready to merge.

Regarding the zh_TW.Big5.po file, I propose to convert the file charset from Big5 to UTF-8, update the charset information in the PO file header and rename the file.

I'm a bit confused here, for instance if I check the charset of the file:

file -i zh_TW.Big5.po
zh_TW.Big5.po: text/x-po; charset=iso-8859-1

If I run iconv to change the charset:

iconv -f BIG5 -t UTF-8//TRANSLIT zh_TW.Big5.po -o zh_TW.po
file -i zh_TW.po
zh_TW.po: text/x-po; charset=utf-8

But the translation is not the same, while if I type:

iconv -f iso-8859-1 -t UTF-8//TRANSLIT zh_TW.Big5.po -o zh_TW.bis.po
file -i zh_TW.bis.po
zh_TW.bis.po: text/x-po; charset=utf-8

The characters in the translation seem to be the same like in the original po file ( zh_TW.Big5.po)

I tried it on my machine and when I used:

iconv -f Big5 -t UTF-8 po/zh_TW.Big5.po > po/zh_TW.po

...I get a file where the characters look like expected characters. For example:

#: notify.cpp:34 notify.cpp:40
msgid "You have new mail!"
msgstr "收到新Mail"

While when I used iso8859-1 as the input charset, the result was a strange clutter of characters:

#: notify.cpp:34 notify.cpp:40
msgid "You have new mail!"
msgstr "¦¬¨ì·sMail"

It seems that I can push the converted renamed file.

Maybe It's my user's settings.

"iconv -f BIG5 -t UTF-8" should work, I see some Chinese characters, same you reported as example.

Converted renamed zh_TW file pushed.

Merged and pushed upstream.