It links to http://shevek.github.io/jcpp/docs/cobertura/, but that page returns a 404 Not Found error.

Will this work?

             }
             if (include(quoteincludepath, name))
                 return;
+        } else {
+            int idx = name.indexOf('/');
+            if (idx != -1) {
+                String frameworkName = name.substring(0, idx);
+                String headerName = name.substring(idx + 1);
+                String headerPath = frameworkName + ".framework/Headers/" + headerName;
+                if (include(frameworkspath, headerPath))
+                    return;
+            }
         }

         if (include(sysincludepath, name))

Lots of headers on Windows use #pragma once instead of ifndef guards (or at least, used to use it when I last checked), so this feature is quite important.

See MSDN documentation

JNAerator uses a hack that involves overriding Preprocessor.include (using a forked jcpp where this is possible), but it would be great if jcpp had native support for this (err... maybe it does already, thanks for letting me know in that case :-))

https://github.com/jeffplourde/jcpp/commit/1c078608bb6ecdda6253b3fe0fac56731d15a761#commitcomment-5105734

ochafik@5b44dee#commitcomment-5110790

Always call getMacro() not macros.get() in Preprocessor.

The Input Stream:

aaa,,

Parses as:

Returning fresh token [IDENTIFIER@1,0]:"aaa"
pp: Returning [IDENTIFIER@1,0]:"aaa"
Returning fresh token [,@1,3]:","
pp: Returning [,@1,3]:","
Returning fresh token [,@1,3]:","
pp: Returning [,@1,3]:","
pp: Returning [EOF]:<265>

The column information of the seconds comma should be 4.

The JCPP version is 1.2.6

Since this commit ca42036 the occurrence of minus sign followed by a number constant is returned by preprocessor as a single token. I wonder why the minus sign is not considered to be a separate token (as it was previously)? Is it required by some preprocessor specification?

mimuw-distributed-systems-group@155b634

Hello!

Is jcpp capable of lexing and removing C++ style line comments?

// This is a comment

I can't see anything in the API to do this, and there doesn't seem to be a Feature value that would enable it. I'm seeing line comments in the output of the preprocessor, so presumably it doesn't do it by default, either.

There were two problems using the ANT build task for JCPP under windows:

1- "etc/global.xml", line 15
there is an absolute path reference to /usr/share/ant/lib
It is better to use the ANT property ${ant.home} instead, so the script will also work in different installations:
<fileset dir="${ant.home}/lib">

2- "etc/targets/global-javadoc.xml", line 51:
the javadoc task makes a reference to directory "${global.dir.etc}/javadoc" ()
When I use a fresh copy from the sources, this directory does not exists, so the javadoc task fails (aborting the build process). A good workaround is to create this directory if it does not exists, before the javadoc task is called. So I've inserted the following two lines at line 6 of "etc/targets/global-javadoc.xml":
<!-- create javadoc etc directory if it does not exist-->
<mkdir dir="${global.dir.etc}/javadoc" />

NullPointerException is thrown for source.setActive(true) in Preprocessor._token() when #endif is missing and #ifdef is satisfied.
Here is a simple test case:

#define D
#ifdef D
    int a = 1;
#else
    int a = 2;
<EOF>

Similarly for:

#ifndef D
    int a = 1;
#else
    int a = 2;
<EOF>

Preprocessor hangs up when the /* */ comment is not closed.

The simplest possible test case:

/*<EOF>

Issue separated from #23 as that has a patch and this has not.

https://github.com/jeffplourde/jcpp/commit/c7c0cce48187776586339b2a3c1e30e415debb18

When we define a multi-line macro, such as:

6: #define THREE 1 \
7:    + \
8:    2

we could expect that calling token.getLine() for Token representing number "2" would return line 8, but surprisingly the entire define definition is regarded as one-line preprocessor directive, so the result is 6.

The tokens list representing macro THREE:

[HASH@6,0]:"#"
[IDENTIFIER@6,1]:"define"
[IDENTIFIER@6,8]:"A"
[(@6,9]:"("
[IDENTIFIER@6,10]:"a"
[,@6,11]:","
[IDENTIFIER@6,13]:"b"
[)@6,14]:")"
[IDENTIFIER@6,16]:"a"
[WHITESPACE@6,17]:"    "
[+@6,21]:"+"
[WHITESPACE@6,22]:"    "
[IDENTIFIER@6,26]:"b"
[NL@6,27]:"

I'm aware that tokens list in Macro is not public, but still line and column numbers should be correct.

Input (groovy):

def pp = new Preprocessor();
pp.addInput(new StringLexerSource("""
#define ONE(arg) one_##arg
#define TWO(arg) ONE(two_##arg)

TWO(good)
TWO(     /* evil newline */
    bad)

ONE(good)
ONE(     /* evil newline */
    bad)
""", true))
println new CppReader(pp).text.trim()

Output:

one_two_good
    one_two_ bad

    one_good
    one_
 bad

Expected output:

one_two_good
one_two_bad

one_good
one_bad

Float number "0.1" parsed as three tokens: octal numeric "0", dot ".", numeric "0" intead of float numeric "0.1".

jcpp should support __VA_ARGS__.

Here's a fix ready to be cherry-picked from JNAerator's fork:

ochafik@444a57c

(can send a PR, but it currently depends on #4)

Hi Shevek,

In Preprocessor.pragma() you use token() (2 times). I am almost sure
it meant to be source_token() or _token(). Am I right?

Thanks,
Igor

JCPP cannot handle absolute paths in include directives. They are treated as relative. For example,

include "/home/me/test.h"

will not work. On the other hand

include "/test.h"

will work looking for the file in the current directory.

There is no wiki either.

Just wanted to try out this project, but ran into some issues. After patching #40, I run into this:

$ ./gradlew run --args="--help"

> Task :compileJava
[...]
> Task :run FAILED
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
Exception in thread "main" java.lang.NullPointerException
        at org.anarres.cpp.BuildMetadata.<init>(BuildMetadata.java:40)
        at org.anarres.cpp.BuildMetadata.getInstance(BuildMetadata.java:29)
        at org.anarres.cpp.Main.run(Main.java:61)
        at org.anarres.cpp.Main.main(Main.java:51)

FAILURE: Build failed with an exception.
[...]

It looks like it's failing to load /META-INF/jcpp.properties from the built resources. I don't see any file with that name, though.

Hello there!

I thought about extending jcpp to support preprocessing of GLSL (OpenGL Shading Language). This would require to add new preprocessor directives such as "#version" and "#extension", which can be basically ignored in my case. I'm currently modifying the classes as I need them but you might be interested in adding the ability to just extend your preprocessor in a more reasonable fashion for such use cases. Such as adding hooks or just subclassing Preprocessor or something similar.

Regards
Holger

1. Empty (missing) variadic arguments should be accepted.
2. ``, ## VA_ARGS'' should remove the comma, if the variadic argument is empty.

Hello! The following file yields a ClassCastException:

#define EXAMPLE_X 'a'

#if EXAMPLE == EXAMPLE_X
#endif

I have a small repro case here:

https://github.com/io7m/jcpp-bug-20170423

The bug appears to be at:

https://github.com/shevek/jcpp/blob/master/src/main/java/org/anarres/cpp/Preprocessor.java#L1598

The value held by the token isn't a Character but a String.

When Main run fails, it does not return with nonzero exit code, which means that in a shell script, an invoker does not have a way to know that the process failed.

gcc pack more than six sequent empty line into one line with using of # directive
like this one:

before:
\r
\r
\r
\r
\r
\r
void foo(){}
after:
# 6 "filename.ext"
void foo(){}

https://gcc.gnu.org/onlinedocs/cpp/Preprocessor-Output.html

this is dramatically decrease size of output file

Hello!

Any chance you'll be publishing Automatic-Module-Name entries for the jcpp jar?

http://blog.joda.org/2017/05/java-se-9-jpms-automatic-modules.html

I'd suggest: Automatic-Module-Name: org.anarres.jcpp

JogAmp/jcpp@ff2b708

Although jcpp parses #include_next, it doesn't seem to skip the current file's directory.

See http://gcc.gnu.org/onlinedocs/cpp/Wrapper-Headers.html

Here's a fix from JNAerator's fork:

ochafik@823a020

Hi Shevek,

In pop_source() you have:

return line_token(t.getLine() + 1, t.getName(), "2");

It seems like there should be no "+ 1" there since it creates
line number mismatch.

Thanks,
Igor