On 02 Nov 2006, at 19:03, Ken G. Brown wrote:
With fpc, in trying to compile an example brought in from CW, and
using XCode 2.4.1 on a Dual 2.0 G5 10.4.8, I have started with a
Carbon template. In the target rules for processing pascal source
files, I see a custom script indicated to use output files $
(PROJECT_DERIVED_FILE_DIR)/$(INPUT_FILE_BASE).s
How do you look at the value that PROJECT_DERIVED_FILE_DIR is set to?
By looking at the build transcript. To see it, go to the "Build
Results" (cmd-shift-B), and then click on the small icon looking like
a couple of lines of text (next to a yellow warning triangle -- yes,
the user interface of that pane is bad).
And where would you change the setting? I cannot find it in any
documentation.
You can't change it, it's automatically set by Xcode. You can find
out about it (and tons of others) by having a look at the build
transcript mentioned above.
Is the custom script necessary?
Yes.
I am getting a compile error saying:
powerpc-apple-darwin8-gcc-4.0.1: no input files while looking
for .s file in the Build/projectname.build/DerivedSources directory.
Is a .s always created from Pascal Sources during compiles?
Yes.
I see a program.s file in Build/projectname.build/DerivedSources/
PPC/ directory but am not sure why it cannot be found or what to do
about it.
The compilation script should copy it to Build/projectname.build/
DerivedSources after compilation is finished, merging ppc and intel
versions of the assembler code if necessary (because in the past
Xcode did not allow specifying different output files for different
architectures). This should be done by the make_conditional()
function in the compilation script.
You may also want to try Erling Johanssen's alternate Xcode project
files, available from
http://www.fabfilter.com/frederik/fpc-app-templates-20060921-ejo.zip
They are written for Xcode 2.3 or later and therefore need far less
workarounds for Xcode bugs and limitations (and hence are less prone
to hard to find bugs). They also solve a few issues with the project
files we distribute (the main one is that it puts the compilation
script in a different target to avoid a race condition).
Jonas
_______________________________________________
fpc-pascal maillist - fpc-pascal@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
