How is the -fprofile-prefix-path option supposed to work? - linux

When compiling code for coverage instrumentation (to use with lcov later on), we're compiling from a base directory tree (let's call it A), and we want the .gcda files to be produced at a different place (because the target directory tree is different - let's call it B).
So, the compilation command looked like this:
gcc -O0 -g --coverage -fprofile-dir=B -c -fPIC -Wall -o A/otherpath/to/mySourceFile.o A/path/to/mySourceFile.c
When checking the contents of mySourceFile.o (with the strings command), I saw that the mySourceFile.gcda file was set to be generated in B/A/otherpath/to/mySourceFile.gcda
Which is the mangling of the path given through the -fprofile-dir option with the exact absolute path of the object file created - just as written in the documentation. So far, no problem - except that what I want would be the mySourceFile.gcda file to be generated from the B directory, WITHOUT the A part.
So, the documentation also mentions the -fprofile-prefix-path option, which is supposed to allow you to remove part of the path, so that the mangling doesn't add the old path to the new.
I tried using it in the following way:
gcc -O0 -g --coverage -fprofile-dir=B -fprofile-prefix-path=A -c -fPIC -Wall -o A/otherpath/to/mySourceFile.o A/path/to/mySourceFile.c
However, after checking through strings, once again, in the generated object file, the path was still B/A/otherpath/to/mySourceFile.gcda, whereas I expected it to be B/otherpath/to/mySourceFile.gcda (that is, I expected the A part to have been stripped by the -fprofile-prefix-path option.)
Obviously, it didn't work. Any insight why ?
( Compiler used is GCC 11.2.1, which is a version recent enough to know about the option. )

Ok, after some tinkering, I got results. Maybe not exactly what I was expecting, but close enough.
Let me start by saying that the A and B "directories" I mentioned in my question were absolute paths. And it didn't work well.
However, while keeping the absolute B (target) path, I tried not using the full A (source) path while compiling. More precisely, I didn't use it to specify the OUTPUT file name, for the object. Instead, I went to the base directory (the A path), and then, ran the command by specifying the output file path relative to the current (A) directory
Which would give the following command:
(From directory A)
gcc -O0 -g --coverage -fprofile-dir=B -fprofile-prefix-path=A -c -fPIC -Wall -o otherpath/to/mySourceFile.o path/to/mySourceFile.c
This time, the source command did show an interesting result, for the mySourceFile.gcda file:
B#otherpath#to#mySourceFile.gcda
As you can see, it's not exactly what I wanted (there are # instead of /), but mentions to A disappeared, and overall, I'm confident it should work as intended. Not utterly sure yet (I still have to test it on the target platform, which will need tinkering with the way the makefiles currently work), but confident nonetheless.
Also, if I didn't use the -fprofile-prefix-path in the command, then the string would mention the A path, like this (with the '/' inside the A path being replaced with '#' characters, obviously):
B#A#otherpath#to#mySourceFile.gcda
So, the option works, but only when using relative paths, not when using absolute ones, for the object file. Hope that helps people.
PS: I checked by changing the path to the source (.c) file. Whether specified using absolute, or relative, paths, it didn't change the outcome. What matters is specifying the path to the object file in a relative manner.

Related

Testing a modified version of readelf

I modified the readelf.c file in binutils-2.36.1/binutils/ such that it prints a few details differently with some flags such as "s","S","a" and doesn't affect the output of other flags.
I'm trying to test whether the changes I made to the file affected any other flags than the ones I intended(mentinoed above).
and therefore I generated a few tests of the following format :
./binutils/readelf -g ./readelfTests/Objects/ObjectFiles/object_1.o
./binutils/readelf -n ./readelfTests/Objects/ObjectFiles/object_1.o
./binutils/readelf -e ./readelfTests/Objects/ObjectFiles/object_1.o
./binutils/readelf -S ./readelfTests/Objects/ObjectFiles/object_1.o
and so on, you get the point.
the problem is the .o files I have are very basic with few sections and variables therefore running a test on them may not catch the errors in my code, I'd appreciate a way to get some .o files with a lot of sections and variables such that running tests on them may actually be effective.
or alternatively I'd appreciate a way to test my modified readelf file in an automatic way.

scons surrounds option with double quotes

I use scons (V1.1.0) for a project that contains a build step that involves the flex tool.
The definition for the flex command in the scons default rules is:
env["LEX"] = env.Detect("flex") or "lex"
env["LEXFLAGS"] = SCons.Util.CLVar("")
env["LEXCOM"] = "$LEX $LEXFLAGS -t $SOURCES > $TARGET"
which I don't want to change.
However, since -t causes #line directives to be created in the output file that refer to the file "<stdout>", this confuses the subsequent gcov processing.
As a solution, I found that -o can be used to override the file name flex produces into the #line directives (it still produces its output on stdout due to the -t option which apparently has precedence).
To achieve that, I added this in the project's SConscript file:
env.AppendUnique(LEXFLAGS = ['-o $TARGET','-c'],delete_existing=1)
I added the -c option (which does nothing) only to show the difference between how it is treated compared to -o.
An according debug print in the SConscript file results in the following (as expected):
repr(env["LEXFLAGS"]) = ['-o $TARGET', '-c']
This results in the following command line, according to the scons log:
flex "-o build/myfile.cpp" -c -t src/myfile.ll > build/myfile.cpp
So the -c option gets into the command line as desired, but the -o option and its filename parameter has double quotes around it, that must have been created by scons when expanding the LEXFLAGS variable.
When I use this definition for LEXFLAGS instead:
env.AppendUnique(LEXFLAGS = ['--outfile=$TARGET','-c'],delete_existing=1)
the resulting command line works as desired:
flex --outfile=build/myfile.cpp -c -t src/myfile.ll > build/myfile.cpp
So one could speculate that the blank in the -o case caused the double quotes to be used, maybe in an attempt to bind the content together into one logical parameter for the command.
So while my immediate problem is solved by using --outfile, my question is still is it possible to rid of the double quotes in the -o case?
Thanks,
Andy
SCons 1.1.0 is extremely old at this point. I'd recommend trying 2.3.0. But your analysis is correct; if an option (a single option, that is) has a space in it, SCons will quote it so it stays a single option. But you don't have a single option; you really have two, '-o' and '$TARGET'. Just break it up like that and it'll work.

Windres syntax error

I am working in MinGW environment (downloaded with their installer on 12/12/2011). I am attempting to compile a resource (.rc) file using Windres. The specific command I use is
Windres -O coff About1.rc -o About1.res
Windres generates at least 100 lines of warning messages reading: "warning: null characters ignored". Following this Windres emits: "Abouty1.rc:1:syntax error".
As a matter of fact, there are no null characters in the About1.rc file. In addtition, the first line of the file is an include statement: #include "dlgresource.h". I played around and eliminated this statement and it turns out that it doesn't matter what I put there, I get the same flurry of messages and the syntax error notification.
To make things more confusing, this same .rc file compiles without any problem using MSFT's rc.exe. The resulting .res file links smoothly with the program .obj file and runs perfectly.
I have no idea what is going on. Any ideas?
Thanks,
Mark Allyn
Your .rc file is probably encoded as UTF-16.
That's what's required in general by Microsoft's [rc.exe], in order to be able to deal with international characters, but GNU [windres.exe] can only deal with ANSI encoding.
One workaround is to convert the file to ANSI on the spot (possibly losing e.g. Russian or Greek characters):
> chcp 1252
Active code page: 1252
> type my.rc | windres --output-format=COFF -o my.res
> _
You probably used VS or a similar tool to generate the file. There are some parts of the character encodings that you cannot see resulting in null characters and etc.
Generate a new .res file with the same content, don't copy/paste the content, type it in yourself.
Try:
windres About1.rc -o About1.o
and then just use the resulting .o file instead of the originally intended .res file.
I've had the same troubles than you today. I know it has passed a lot of time from your question, but I'm writting this on the hope that it can be useful for someone.
First, I obtained an object file .o compiled using Cygwin, writting:
windres -o resource.o resource.rc
By doing that, you dont need to use the .res file, but the .o one, and you can then link this object with all the others, when you compile yout program, using GNU resources:
g++ Header_files CPP_files flags ... -o program.exe recource.o -lm
For instance.

Including header files in cygwin

As you know the getch() and getche() functions don't work with the cygwin, a linux oriented one.
But can I include the conio.h header file of borland c and call the functions getch in my makefiles?
Will it work and can anyone tell me how to include the header files from different directories in cywgin.
I have a header file strcal.h in directory c:/makk/string/.
How do I include that header file in my makefile?
gcc -I/string small.c
It is not working and my current directory is makk.
In stdio.h, there is a getchar() function which is what you need. You can't just bring across the Borland header file since that just declares the function, it doesn't define it. Standard C has no need for getch().
To include header files in different areas, you use the -I directives of gcc to set up search paths.
So, if you have a /xyz/myheader.h file, you can do something like:
gcc -I /xyz myprogram.c
To get at c:/makk/string/strcal.h, you may have to use gcc -I /cygdrive/c/makk/string or, if you know you're actually in that makk directory, you can use -I string (note the lack of leading / since you want a relative path, not an absolute one).

g++ searches /lib/../lib/, then /lib/

According to g++ -print-search-dirs my C++ compiler is searching for libraries in many directories, including ...
/lib/../lib/:
/usr/lib/../lib/:
/lib/:
/usr/lib/
Naively, /lib/../lib/ would appear to be the same directory as /lib/ — lib's parent will have a child named lib, "that man's father's son is my father's son's son" and all that. The same holds for /usr/lib/../lib/ and /usr/lib/
Is there some reason, perhaps having to do with symbolic links, that g++ ought to be configured to search both /lib/../lib/ and /lib/?
If this is unnecessary redundancy, how would one go about fixing it?
If it matters, this was observed on an unmodified install of Ubuntu 9.04.
Edit: More information.
The results are from executing g++ -print-search-dirs with no other switches, from a bash shell.
Neither LIBRARY_PATH nor LPATH are output from printenv, and both echo $LPATH and echo LIBRARY_PATH return blank lines.
An attempt at an answer (which I gathered from a few minutes of looking at the gcc.c driver source and the Makefile environment).
These paths are constructed in runtime from:
GCC exec prefix (see GCC documentation on GCC_EXEC_PREFIX)
The $LIBRARY_PATH environment variable
The $LPATH environment variable (which is treated like $LIBRARY_PATH)
Any values passed to -B command-line switch
Standard executable prefixes (as specified during compilation time)
Tooldir prefix
The last one (tooldir prefix) is usually defined to be a relative path:
From gcc's Makefile.in
# Directory in which the compiler finds libraries etc.
libsubdir = $(libdir)/gcc/$(target_noncanonical)/$(version)
# Directory in which the compiler finds executables
libexecsubdir = $(libexecdir)/gcc/$(target_noncanonical)/$(version)
# Used to produce a relative $(gcc_tooldir) in gcc.o
unlibsubdir = ../../..
....
# These go as compilation flags, so they define the tooldir base prefix
# as ../../../../, and the one of the library search prefixes as ../../../
# These get PREFIX appended, and then machine for which gcc is built
# i.e i484-linux-gnu, to get something like:
# /usr/lib/gcc/i486-linux-gnu/4.2.3/../../../../i486-linux-gnu/lib/../lib/
DRIVER_DEFINES = \
-DSTANDARD_STARTFILE_PREFIX=\"$(unlibsubdir)/\" \
-DTOOLDIR_BASE_PREFIX=\"$(unlibsubdir)/../\" \
However, these are for compiler-version specific paths. Your examples are likely affected by the environment variables that I've listed above (LIBRARY_PATH, LPATH)
Well, theoretically, if /lib was a symlink to /drive2/foo, then /lib/../lib would point to /drive2/lib if I'm not mistaken. Theoretically...
Edit: I just tested and it's not the case - it comes back to /lib. Hrm :(

Resources