On Sun, Dec 17, 2006 at 02:04:55PM +0100, Georg Baum wrote: > Am Sonntag, 17. Dezember 2006 13:28 schrieb Enrico Forestieri: > > > Thanks. However, now I am really confused. I can't reproduce anymore > > the assertion reported by Michael. After I got it the first time, I > > tried recompiling src/support/filename.C and src/support/filetools.C with > > debugging symbols and ascertained that encodeName() was indeed returning > > the right thing. The assertion apparently is not triggered anymore, even > > recompiling without debugging symbols! > > > > All I can think is that for some reason those files were not recompiled > > the first time? > > I don't know. The lyx2lyx changes I did should not make any difference. I > only moved the point where a FileName is created to some other location. > Michael, do you still get the assertion?
I was able to reproduce again the assertion, but it has nothing to do with non ascii characters. 1) Launch LyX giving a not existing file name as argument and confirm that you want create a new document with that name. 2) Make any change and try to save the file: you can't and LyX asks whether you want to rename the document and try again. 3) Choose to rename the file and try to save it. LyX asserts: $ ./src/lyx-qt4 ~/foo.lyx Assertion triggered in lyx::support::FileName::FileName(const std::string&) by failing check "empty() || absolutePath(name_)" in file ../../../src/support/filename.C:47 Abort (core dumped) BTW, the document is actually saved but I am not able to get a backtrace as LyX exits with code 0103000 even from within gdb. I notice that after point 1) no name is shown in the title bar and at 3) the file dialog shows the current directory instead of that specified on the command line and the proposed file name is simply ".lyx". -- Enrico
