Typos & nitpicks
Fix some typos in the READMEs.
Nit #1: Use full path when cloning repo
If we use the basename, we're forced to run the script from the directory where
we want the whole thing to live. By using the full path, we can run the script
from anywhere.
Nit #2: Use more general name for variable (generator vs ninja)
Nit #3 (noticed by Renato): Set generator explicitly in llvm-build
When ninja is not found, we use make for the build. However, we don't tell
cmake about it and instead let it choose its own default generator. This commit
explicitly sets the cmake generator to "Unix Makefiles", in case cmake's default
generator for the current platform is not make.
In order to get this to work, it was necessary to make some changes to the
safe_run helper function, so we wouldn't lose the quotes around "Unix
Makefiles".
Change-Id: I9592645d517d006b0e680a2996d498e57f009400
diff --git a/README.txt b/README.txt
index b1cb7de..1e611cc 100644
--- a/README.txt
+++ b/README.txt
@@ -2,7 +2,7 @@
===============================
These scripts are meant to help you setup and develop around LLVM and its
-sub-projects. There are many perks when developing with LLVM, but the scripts
+sub-projects. There are many quirks when developing with LLVM, but the scripts
can only cover part of them, while others have to be explained via documents.
The directory is split in multiple parts:
@@ -17,7 +17,7 @@
Helpers: Scripts that setup the Git/Svn repositories, check-out, validate
branches, build and test LLVM in its multiple forms. Since the repositories have
an intricate relationship (symlinks inside the llvm tree), dealing with them is
-not part of the scripts, and they assume the setup is conciously correct.
+not part of the scripts, and they assume the setup is consciously correct.
Monitor: Scripts to help you monitor buildbots.