From: Max T. W. <max...@ve...> - 2005-04-23 23:22:14
|
Earnie Boyd wrote: > > On 2:20:22 pm 2005-04-21 "Max T. Woodbury" <max...@ve...> wrote: > > I've been trying to build a useable MSYS from CVS HEAD with changes > > and failing. > > > > I figured that I might have better luck if I tried to build a base > > 1.0.10 distribution so I could be sure I hadn't messed up something > > without knowing it was important. I pulled and unpacked the src > > tarball from sourceforge, made a working copy of the resulting tree > > with cp -pR and tried configure and make. I seem to be missing some > > crucial step... > > > > Could you point me at or provide a description of what I should be > > doing? > > > > Did you follow these instructions? http://www.mingw.org/MinGWiki/index.php/Build%20MSYS Thank you that was the pointer I needed. Someone is in the process of updating this information. This should be very helpful indeed. However, there are some 'interesting' statements in the new description that raise questions when you think about them. 1) There is a recommendation to get TortoiseCVS. msysDTK already includes CVS. Why get a second version? 2) There is a recommendation to get GnuWin32 bison. There is a MinGW bison package. I suspect that the MinGW version is specifically adapted to MSys/MinGW, so why get the non-MinGW version? (Also, there is an implication that FLex is needed. Is this true?) 3) I did not see a sequence for building the test suites. I happen to believe strongly in regression testing, so I would like to be able to run any test programs available. What is needed for this? Note on the download page -- Originally I stopped scrolling when I came the the 'Previous' table simply because I didn't want to store the old stuff and missed the 'Snapshot' section as a result. Could you please move the 'Snapshot' section so that is follows 'Current' but precedes 'Previous'? ma...@mt... |