In summary, this means that the "official" building instructions are misleading with regard to Windows, to say the least.
I'm not used nor do I like to teach professionals how to do their jobs, but what about adding a clear warning on the README directed to people like me trying to build OpenJDK7 on Windows? On 2010-05-05 19:59, Kelly O'Hair wrote: > > Many months back we were very close to formally converting to Visual > Studio 2008 (VS2008), > in the process of getting ready, I updated the README-builds.html file > to reflect VS2008 use > (at http://hg.openjdk.java.net/jdk7/build/raw-file/tip/README-builds.html). > That was a mistake on my part, the effort to convert to VS2008 stalled > and most > developers (including our formal JDK7 Release Engineering team), at > least internally > are still using the older VS2003 compilers. > > Our current thinking is that VS2010 makes more sense now and we are > trying to move > forward on that effort, but it's slow going. > > So as people make changes, I suspect they will create source and issues > for compilers > newer than VS2003 (post-VS2003), we just need to fix them as we run into > them. > Once we officially convert to a newer compiler, enforcement will be > easier, but the > opposite may start happening, it might be that VS2003 might start having > problems > compiling the sources as changes are made in the post-VS2003 world. But > once we > leave VS2003, it will probably become a permanent parting very quickly. > It's extremely difficult to keep all the VS compilers able to compile > the same > source, enforcement is difficult to say the least. > > In this case, I would use the awt team alias and find out from them what > they > plan to do to fix this, the alias is awt-dev AT openjdk.java.net. > > -kto > > On May 5, 2010, at 4:16 AM, Raffaello Giulietti wrote: > >> I'm on Vista 32-bit and have VisualStudio 2008 Professional installed. >> >> After some days hunting for backslashes, spaces and colons in paths, I >> now encounter a compilation error that seems to be known >> (http://bugs.sun.com/bugdatabase/view_bug.do;jsessionid=cc2b5a48cddee8e8cdbb6795b07?bug_id=6936319). >> >> >> Does anybody know how to circumvent or fix it? >> >> With this error, and perhaps even other ones I've not yet seen, I'm >> wondering how Sun/Oracle can build the binary snapshots if they are >> supposed to use the same tool chain that I'm using :-( >> I'm beginning to think that they possess some special magic or some >> undocumented killer tool. >> >> Thanks in advance for any support >> Raffaello >
