Kayvan A. Sylvan wrote: >> > I can probably spend a bit of time isolating exactly which >> > change broke lyx on Cygwin, after I have finished some work for >> > my employer and finished an assignment for school. >> >> Good. Thanks. This all smacks of something going wrong with the reading >> of the PATH variable (and maybe the writing to is). getEnvPath and >> setEnvPath in support/filetools.C. At a guess, what happens if you >> remove the cygwin_path_fix? > > I don't know, I haven't tried it, but as of the 1/16/05 repository, the > cygwin_path_fix patch was in and I have no problems. > > It's the changes that were made afterwards that make LyX unusable (no > external programs seem to run, including the Tools->Reconfigure script).
At the risk of being called a PITA, there's not much I can do to help from here. I have a linux box and Tools->Reconfigure works just fine for me. I have only LyX 1.3.x installed on Windows (native build) and that is working fine. However, it's in a state similar to your 16 Jan 1.4.x code as I have not applied a number of small changes to the 1.3.x tree yet. I've just made a diff of all the changes to the 1.4.x. tree since 16 Jan. $ cvs -q diff -D'16 Jan 2005 00:00' -Dnow > changes_since_16Jan.diff Stripping out Asger's MSVC additions to the boost libraries leaves a patch that's 2300 lines long. Of this we can discount changes to: src/BranchList.C src/Sectioning.C src/buffer.C src/cursor.C src/dimension.h src/lyxlex_pimpl.C src/output_docbook.C src/rowpainter.C src/sgml.C src/frontends/* src/insets/* src/mathed/* src/support/gzstream.C src/support/os_win32.C That leaves changes to these files only: src/lyx_main.[Ch] src/lyxfunc.C src/lyxrc.[Ch] src/support/filetools.[Ch] src/support/package.C.in Of these, almost all of the changes to lyxrc.[Ch] are superficial. I added a path_prefix variable and act on it, but that shouldn't affect you. I also removed the call to os::cygwin_path_fix from LyXRC::read, invoking the function instead in lyx_main.C's init function() and in lyx_cb.C's actOnUpdatedPrefs() function. So, I'd be surprised if that change broke anything. Anyway, I hope this is enough info to help you track down the problem. Regards, -- Angus
