From:
cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Andrew Ferguson
Sent: Friday, September 09, 2005
7:51 AM
To: cdt-dev@xxxxxxxxxxx
Subject: [cdt-dev] managed build:
VCErrorParser patch
I feel a little bad about sending this one but thought
I'd see what you think.. :)
An old (released) version of our compiler is producing
output intended to conform to the Visual Studio output spec. but failing
because its outputting "Warning" rather than "warning".
I've worked around this by adding a new error parser which relaxes this
condition, but is potentially confusing for the user as an additional error
parser appears in the build settings. I'm thinking it would be preferable to
just have the VCErrorParser tweaked as below.
I realize adapting VCErrorParser for every non-conformant
tool output would not be a good road to go down, but was wondering if this case
is so thin that it might not count as the thin end of the wedge?
Index: VCErrorParser.java
===================================================================
RCS file:
/home/tools/org.eclipse.cdt-core/org.eclipse.cdt.core/src/org/eclipse/cdt/internal/errorparsers/VCErrorParser.java,v
retrieving revision 1.9
diff -u -r1.9 VCErrorParser.java
--- VCErrorParser.java 22 Jul 2005 02:36:06 -0000 1.9
+++ VCErrorParser.java 9 Sep 2005 10:53:13 -0000
@@ -60,7 +60,7 @@
}
IFile file =
eoParser.findFileName(fileName);
if (file != null ||
eoParser.isConflictingName(fileName)) {
- String desc =
line.substring(firstColon + 1).trim();
+ String desc =
line.substring(firstColon + 1).trim().toLowerCase();
if (file == null) {
desc = "*" +
desc; //$NON-NLS-1$
}