Add support for $ in fortran identifiers and clickable .i include files.

[navinp0304]

How do i make #include 'file.i' as clickable i.e it is showing as text not a link ? For me #include "file.h" works as a clickable link.
I tried adding it in suffix but it doesn't work. The format is #include 'file.i' not "file.i".

diff --git a/opengrok-indexer/src/main/java/org/opengrok/indexer/analysis/fortran/FortranAnalyzerFactory.java b/opengrok-indexer/src/main/java/org/opengrok/indexer/analysis/fortran/FortranAnalyzerFactory.java
index 3bc256d38d9..14a8fcf3ca1 100644
--- a/opengrok-indexer/src/main/java/org/opengrok/indexer/analysis/fortran/FortranAnalyzerFactory.java
+++ b/opengrok-indexer/src/main/java/org/opengrok/indexer/analysis/fortran/FortranAnalyzerFactory.java
@@ -49,7 +49,9 @@ public class FortranAnalyzerFactory extends FileAnalyzerFactory {
         "F95",
         "F03",
         "F08",
-        "F15"};
+        "F15",
+        "F77",
+        "I"};

Ctags has added support for fortran identifiers with $
universal-ctags/ctags#4033

Please add support for opengrok as well. This works well for me.

diff --git a/opengrok-indexer/src/main/jflex/analysis/fortran/Fortran.lexh b/opengrok-indexer/src/main/jflex/analysis/fortran/Fortran.lexh
index 86994f85da2..7e9577f4588 100644
--- a/opengrok-indexer/src/main/jflex/analysis/fortran/Fortran.lexh
+++ b/opengrok-indexer/src/main/jflex/analysis/fortran/Fortran.lexh
@@ -22,7 +22,7 @@
  * Portions Copyright (c) 2017, Chris Fraire <cfraire@me.com>.
  */

-Identifier = [a-zA-Z_] [a-zA-Z0-9_]*
+Identifier = [a-zA-Z_] [a-zA-Z0-9_$]*
 Label = [0-9]+

 Number = ([0-9]+\.[0-9]+|[0-9][0-9]* | [0][xX] [0-9a-fA-F]+ )([uUdDlL]+)?

[vladak]

Care to submit a PR ?

[navinp0304]

@vladak Yes but why aren't the include files clickable (without href) even after adding suffixes. I'll have that and then submit PR.

[navinp0304]

@vladak Yes but why aren't the include files clickable (without href) even after adding suffixes. I'll have that and then submit PR.

Hi @vladak , can you please help with this ?

[vladak]

For the include directive I think this needs to be processed by the Fortran analyzer. I believe https://github.com/oracle/opengrok/blob/master/opengrok-indexer/src/main/jflex/analysis/fortran/FortranSymbolTokenizer.lex needs to obtain similar rule as e.g. in

opengrok/opengrok-indexer/src/main/jflex/analysis/c/CSymbolTokenizer.lex

Line 54 in 0e0e613

"#" {WhspChar}* "include" {WhspChar}* ("<"[^>\n\r]+">" | \"[^\"\n\r]+\") {}

, just with the single quotes surrounding the file path element.

Is the file include a feature of (modern) Fortran actually ? Just asking, I have never worked with Fortran before.

[vladak]

The Fortran lexing rules already contain file/path matching:

opengrok/opengrok-indexer/src/main/jflex/analysis/fortran/FortranXref.lex

Lines 166 to 184 in 0e0e613

	<SCOMMENT, STRING, QSTRING> {
	{FPath} {
	chkLOC();
	onPathlikeMatched(yytext(), '/', false, yychar);
	}
	{File}
	{
	chkLOC();
	String path = yytext();
	onFilelikeMatched(path, yychar);
	}
	{FNameChar}+ "@" {FNameChar}+ "." {FNameChar}+
	{
	chkLOC();
	onEmailAddressMatched(yytext(), yychar);
	}
	}

and there is also

opengrok/opengrok-indexer/src/main/jflex/analysis/fortran/FortranXref.lex

Lines 89 to 96 in 0e0e613

	"<" ({File}|{FPath}) ">" {
	chkLOC();
	onNonSymbolMatched("<", yychar);
	String file = yytext();
	file = file.substring(1, file.length() - 1);
	onFilelikeMatched(file, yychar + 1);
	onNonSymbolMatched(">", yychar + 1 + file.length());
	}

so perhaps these need to be augmented and/or receive rule similar to the one used for C:

opengrok/opengrok-indexer/src/main/jflex/analysis/c/CXref.lex

Lines 98 to 113 in 0e0e613

	"#" {WhspChar}* "include" {WhspChar}* ("<"[^>\n\r]+">" | \"[^\"\n\r]+\") {
	chkLOC();
	String capture = yytext();
	Matcher match = MATCH_INCLUDE.matcher(capture);
	if (match.matches()) {
	onNonSymbolMatched(match.group(INCL_HASH_G), yychar);
	onFilteredSymbolMatched(match.group(INCLUDE_G), yychar, Consts.kwd);
	onNonSymbolMatched(match.group(INCL_POST_G), yychar);
	onNonSymbolMatched(match.group(INCL_PUNC0_G), yychar);
	String path = match.group(INCL_PATH_G);
	onPathlikeMatched(path, '/', false, yychar);
	onNonSymbolMatched(match.group(INCL_PUNCZ_G), yychar);
	} else {
	onNonSymbolMatched(capture, yychar);
	}
	}

(which should match the rule added to FortranSymbolTokenizer.lex)

Further, there is also

opengrok/opengrok-indexer/src/main/jflex/analysis/fortran/FortranXref.lex

Line 62 in 0e0e613

File = [a-zA-Z]{FNameChar}* ".inc"

which probably needs to be reconciled with the new state of include files in Fortran.

[navinp0304]

For the include directive I think this needs to be processed by the Fortran analyzer. I believe https://github.com/oracle/opengrok/blob/master/opengrok-indexer/src/main/jflex/analysis/fortran/FortranSymbolTokenizer.lex needs to obtain similar rule as e.g. in

opengrok/opengrok-indexer/src/main/jflex/analysis/c/CSymbolTokenizer.lex

Line 54 in 0e0e613

"#" {WhspChar}* "include" {WhspChar}* ("<"[^>\n\r]+">" | \"[^\"\n\r]+\") {}

, just with the single quotes surrounding the file path element.
Is the file include a feature of (modern) Fortran actually ? Just asking, I have never worked with Fortran before.

No it was there since Fortran 77 include 'file' or include "file" . Below is the link
https://docs.oracle.com/cd/E19957-01/805-4939/6j4m0vna1/index.html

[navinp0304]

Care to submit a PR ?

Submitted