On Tue, 29 Apr 2008, Duncan Murdoch wrote:
This is not a bug, it's a feature. It stops me from distributing versions with unintentional uncommitted changes.
It does have that effect, but the error messages
are pretty obscure:
c:\Rtools\MinGW\bin\windres.exe: rcico.rc:9: syntax error
and
Error on line 12 in
e:\R\R-svn\r-devel\src\gnuwin32\installer\R.iss: Value of [Setup]
section directive "VersionInfoVersion" is invalid.
Both problems are Windows-specific and the desire to not distribute
versions with unintentional uncommitted uncommitted changes is not
Windows-specific.
Note that the rcico.rc problem arises as the first step of compiling R
on Windows from clean source tree, so you cannot even test a proposed
change. I ran into it because I had to change src/gnuwin32/MkRules to
say that I put Inno Setup in its standard place, "C:/Program Files/Inno
Setup 5", and not in "C:/packages/inno".
I guess I don't need to have USE_SVNVERSION=yes, but doing that means
that version$"svn rev" shows me I have a non-standard version: that
will show the svn version as something like "123M" or "123:145S". It
is too bad it only is available on the Windows build. I think it would
be a good thing to have on the non-Windows builds also. They use 'svn
info' instead, which does not tell you that the local version is not
the standard one.
If we did that on all platforms then we could check the disribution by
checking that !is.na(as.integer(svn$"svn rev")).
It might also be nice to include the output of 'svn diff' in the
distribution so a user or tester could see how this version was
different than the official one.
On 29/04/2008 2:30 PM, bill at insightful.com wrote:
Full_Name: Bill Dunlap
Version: 2.8.0dev
OS: Windows XP
Submission from: (NULL) (70.98.76.47)
I tried for the first time to build R from source on Windows, where I
got the source code via svn. Per the Installation and Administration
manual, I altered src\gnuwin32\MkRules so it had the the locally
correct paths to HTML Help Workshop and Inno Setup 5. I also set
USE_SVNVERSION=yes, as suggested in MkRules itself. Then, while in
the directory src/gnuwin32 I ran 'make all recommended' and got an
error from windres very early in the build:
E:\R\R-svn\r-devel\src\gnuwin32>make all recommended
make --no-print-directory -C front-ends Rpwd
make -C ../../include -f Makefile.win version
make Rpwd.exe
gcc -std=gnu99 -I../../include -O3 -Wall -pedantic -c rpwd.c -o rpwd.o
windres --preprocessor="gcc -E -xc -DRC_INVOKED" -i rcico.rc -o rcico.o
c:\Rtools\MinGW\bin\windres.exe: rcico.rc:9: syntax error
make[3]: *** [rcico.o] Error 1
make[2]: *** [Rpwd] Error 2
make[1]: *** [front-ends/Rpwd.exe] Error 2
make: *** [all] Error 2
Line 9 of src\gnuwin32\front-ends\rcico.rc is
FILEVERSION R_FILEVERSION
The problem was that my change to MkRules caused 'svnversion' to put an
'M' (modified) on the end of the svn version it reports. This svn
version number is used in the R_FILEVERSION macro, which is used in in
the *.rc files. The resource file compiler, windres, appears to choke on
non-digits in R_FILEVERSION. (A comment in tools\GETVERSION indicates
it might choke on leading 0's as well.)
svnversion can also output a string of the form "123:125"
to mean the current source is from a variety of svn revisions,
between 123 through 125. An "M" or "S" may also be appended.
If you don't have Subversion installed then you might get
a svn version of "unknown". None of these non-digits is
acceptble to windres.
I patched tools/GETVERSION to fix up this problem. It removes
the trailing M or S and removes anything up to an including a
colon, so it reports the highest version in the range. If there
are still non-digits in there it makes the version string "0"
(not useful, but doesn't kill the build with a noninformative
error message).
I ran into another problem with the trailing 'M' from svnversion
when doing 'make rinstaller' in R_HOME/src/gnuwin32:
E:\R\R-svn\r-devel\src\gnuwin32\installer>type R-2.8.0dev.log
Inno Setup 5 Command-Line Compiler
...
Compiler engine version: Inno Setup 5.2.3 (ISPP 5.2.3.0)
...
[ISPP] Preprocessing.
...
[ISPP] Preprocessed.
Parsing [Setup] section, line 10
...
Parsing [Setup] section, line 20
Error on line 12 in e:\R\R-svn\r-devel\src\gnuwin32\installer\R.iss: Value of
[Setup] section directive "VersionInfoVersion" is invalid.
Compile aborted.
The offending line 12 in R.iss is
12 VersionInfoVersion=2.8.0.45381M
and it is put there by src/gnuwin32/installer/JRins.pl.
I patched JRins.pl to ensure the svn revision consisted
of only digits.
Then I could build R on Windows (make distribution)
with USE_SVNVERSION=yes set in src/gnuwin32/MkRules
and with a variety of svn revision settings.
Index: src/gnuwin32/installer/JRins.pl
===================================================================
--- src/gnuwin32/installer/JRins.pl (revision 45553)
+++ src/gnuwin32/installer/JRins.pl (working copy)
@@ -44,6 +44,11 @@
$SVN = <ver>;
close ver;
$SVN =~s/Revision: //;
+## inno setup requires that SVN be all numeric, not 123M, 120:123, "unknown",
or empty
+$SVN =~s/[MS]* *$//;
+$SVN =~s/^.*://;
+$SVN =~s/[^0-9]//g;
+$SVN =~s/^$/0/;
$RVER0 .= "." . $SVN;
open insfile, "> R.iss" || die "Cannot open R.iss\n";
Index: tools/GETVERSION
===================================================================
--- tools/GETVERSION (revision 45479)
+++ tools/GETVERSION (working copy)
@@ -24,6 +24,7 @@
y1=6
svn_rev=unknown
fi
+ all_numeric_svn_rev=`echo ${svn_rev}|sed -e 's/[MS]*$//' -e 's/^.*://' -e
's/[^0-9]//g' -e 's/^$/0/'`
echo "/* Rversion.h. Generated automatically. */"
echo "#ifndef R_VERSION_H"
echo "#define R_VERSION_H"
@@ -43,7 +44,7 @@
echo "#define R_SVN_REVISION \"${svn_rev}\""
## Using 1-digit year stops problems with leading zeros
# echo "#define R_FILEVERSION ${maj},${pl}${sl},${y1}${m}${d},0"
- echo "#define R_FILEVERSION ${maj},${pl}${sl},${svn_rev},0"
+ echo "#define R_FILEVERSION ${maj},${pl}${sl},${all_numeric_svn_rev},0"
echo
echo '#ifdef __cplusplus'
echo '}'
______________________________________________ R-devel at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
---------------------------------------------------------------------------- Bill Dunlap Insightful Corporation bill at insightful dot com 360-428-8146 "All statements in this message represent the opinions of the author and do not necessarily reflect Insightful Corporation policy or position."