"ilinebr" on every link broke my entire project

[aotto1968]

Describe the bug
I get the following error messages (multiple)

1. End of list marker found without any preceding list items ( I have no clue where)
2. "ilinebr" attached on !! many !! links so that near everything is broken

warning: End of list marker found without any preceding list items
warning: unable to resolve reference to 'tcl_ObjectObject\ilinebr' for \ref command
warning: unable to resolve reference to 'tcl_ErrorObject\ilinebr' for \ref command

"linebr" is part of your SW

src/commentscan.l:DOCNL     "\n"|"\\ilinebr"
src/commentscan.l:<Comment>{B}*"\\ilinebr"{B}*            { // preserve spacing around \\ilinebr
src/commentscan.l:<Comment>(\n|\\ilinebr)({B}*(\n|\\ilinebr))+ { // at least one blank line (or blank line command)
src/commentscan.l:                                              else if (strcmp(yytext+i,"\\ilinebr")==0) addOutput(yyscanner,"\\ilinebr"),i+=8;
src/commentscan.l:                                              else if (strcmp(yytext+i,"\\ilinebr")==0) addOutput(yyscanner,"\\ilinebr"),i+=8;
src/commentscan.l:<SectionTitle>[^\n@\\]*/"\\ilinebr"     { // end of section title
src/commentscan.l:    else if (c=='r' && i>=7 && qstrncmp("\\ilinebr",s.data()+i-7,8)==0) // special line break marker
src/doctokenizer.l:  int nl2 = text.findRev("\\ilinebr");
src/doctokenizer.l:<St_Para>{BLANK}*(\n|"\\ilinebr"){LISTITEM}     { /* list item on next line */
src/doctokenizer.l:<St_Para>{BLANK}*(\n|"\\ilinebr"){MLISTITEM}     { /* list item on next line */
src/doctokenizer.l:<St_Para>{BLANK}*(\n|"\\ilinebr"){OLISTITEM}     { /* list item on next line */
src/doctokenizer.l:<St_Para>{BLANK}*(\n|"\\ilinebr"){ENDLIST}      { /* end list on next line */
src/doctokenizer.l:<St_Para>"\\ilinebr"   {
src/doctokenizer.l:<St_CodeOpt>"\\ilinebr" |
src/doctokenizer.l:<St_HtmlOnlyOption>"\\ilinebr" {
src/doctokenizer.l:<St_PlantUMLOpt>{BLANK}*{FILEMASK}{BLANK}*/"\\ilinebr" { // case 5: plain file name specified without title or attributes
src/doctokenizer.l:<St_PlantUMLOpt>"\\ilinebr" |
src/doctokenizer.l:<St_Title>"\\ilinebr"  {
src/doctokenizer.l:<St_TitleN>(\n|"\\ilinebr") { /* new line => end of title */
src/doctokenizer.l:<St_TitleQ>(\n|"\\ilinebr") { /* new line => end of title */
src/doctokenizer.l:<St_TitleV,St_TitleA>(\n|"\\ilinebr") {
src/doctokenizer.l:<St_Cite>(\n|"\\ilinebr")  { // new line
src/doctokenizer.l:<St_Ref>(\n|"\\ilinebr") { // new line
src/doctokenizer.l:<St_Ref2>"\""|\n|"\\ilinebr" { /* " or \n => end of title */
src/doctokenizer.l:<St_SkipTitle>(\n|"\\ilinebr") { return 0; }
src/doctokenizer.l:<St_SecSkip>(\n|"\\ilinebr")
src/doctokenizer.l:<St_Sections>(\n|"\\ilinebr")
src/doctokenizer.l:                         if (g_secTitle.right(8)=="\\ilinebr")
src/doctokenizer.l:<St_Snippet>(\n|"\\ilinebr")  {
src/doctokenizer.l:<*>"\\ilinebr"         {
src/doxygen.cpp:        newValue+="\\ilinebr ";
src/doxygen.cpp:      newValue+="\\ilinebr ";
src/util.cpp:    else if (c=='\\' && qstrncmp(p,"\\ilinebr",8)==0) i+=8,li=i,docLine++,p+=8;
src/util.cpp:    else if (c=='r' && b>=7 && qstrncmp(p-7,"\\ilinebr",8)==0) bi=b-7,b-=8,p-=8;
src/markdown.cpp:// test if the next characters in data represent a new line (which can be character \n or string \ilinebr).
src/markdown.cpp:  if (data[0] == '\\' && qstrncmp(data+1,"ilinebr",7)==0) return 8;
src/markdown.cpp:      m_out.addStr("> " + cellText + "\\ilinebr </" + cellTag + ">");
build/generated_src/commentscan.cpp:{ // preserve spacing around \\ilinebr
build/generated_src/commentscan.cpp:                                              else if (strcmp(yytext+i,"\\ilinebr")==0) addOutput(yyscanner,"\\ilinebr"),i+=8;
build/generated_src/commentscan.cpp:                                              else if (strcmp(yytext+i,"\\ilinebr")==0) addOutput(yyscanner,"\\ilinebr"),i+=8;
build/generated_src/commentscan.cpp:    else if (c=='r' && i>=7 && qstrncmp("\\ilinebr",s.data()+i-7,8)==0) // special line break marker
build/generated_src/doctokenizer.cpp:  int nl2 = text.findRev("\\ilinebr");
build/generated_src/doctokenizer.cpp:                         if (g_secTitle.right(8)=="\\ilinebr")

Expected behavior
no warning -> just do the job

To Reproduce
impossible - it is my entire project :-(

Version
1.8.19 (84084be)

[aotto1968]

one source of the bug is an ref in a "markdown table" - the "tcl_ObjectObject" in the markdown-table should expand to "MqObjectC" like the text under the table (MqObjectC do some stuff) -

Problem: markdown table

[albert-github]

I see, at the moment, the following problems:

1. warning: End of list marker found without any preceding list items
   This typically happens when there is a dot (.) at the beginning of the line.
2. warning: unable to resolve reference to 'tcl_ObjectObject\ilinebr' for \ref command
   This problem is caused by the not handling of a \ilinebr case in doctokinizer.l. The \ilinebr result from a ^^ in an ALIASES.

As you can understand it will be a bit hard to find the cause of the problem (other than some general remarks) without a small example. Is your code propriety code or can you share it (still a small example would be best)?

Regarding the first problem based on the remark above and the line number as mentioned with the with the warning it should be possible to pinpoint the problem (note a guess is that it might also come from an ALIASES where there is ^^.).

Regarding the second problem, this looks like a forgotten case. Also here probably you will be able to pinpoint the problem a bit bu means of the line / file as given with the warning, based on this you probably van create a s small example. At least giving the ALIASES might already help here.

In both cases adjusting the INPUT tag only to one relevant file for each problem might help considerably (of course some other warnings might pop up but those can be ignored during the investigations).

[aotto1968]

thanks for the help… the cause was on missing "}" in a very large project … this created an extremely flood of error-messages without any relationship to the inital missing "}" -

-> it is an problem if the error-messages are not related to the initial problem

The problem was at the "system_primitive_types" item:

 ...
  <dt>  SYSTEM    \anchor \DocNS{idx_system}
    <dd>
      \RNS{system}                      "Philosophy"      ,
      \RNS{system_primitive_types      "PrimitiveTypes"  ,
      \RNS{system_managed_object}       "ManagedObject"   ,
      \RNS{system_storage_management}   "StorageManagement"
  <dt>  CLASS     \anchor \DocNS{idx_class}
    <dd>
...

[albert-github]

So when I understand it correctly the problem is solved and was "usage"

Didn't you get a message like:

Reached end of file while still inside a (nested) comment.  Nesting level 1 (probable line reference: 4)

Still not 100% clear but giving an indication of where the problem occurred.

[aotto1968]

yes - I saw this message but together with ~1000++ other messages - and I don't understand that this message point to a missing "}" - I for example search for a problem with "\if -> \endif" etc … you don't see a missing "}" in a large block of code…

The problem of debugging was: sometimes it works (with an other setup) and sometimes not - sometimes doxygen complains about aliases which are defined but not found… etc. -> part of the output was generated - part not -> it was a total "random" error.

the core problem was - I think - that doxygen run out of internal memory without "}" and than start to overwrite existing memory…

valgrind/sanitze etc does NOT complain -> this seems to be a sign that already allocated memory was "overwritten" inside doxygen…

[albert-github]

Doxygen tries to match the open and close { / } and similar, and here it was missing a } and running to the end of the code wile doxygen thinks that also the comment close is part of, here, the argument of the \RNS.

[albert-github]

Even though the problem is a usage problem, the remark:

The problem of debugging was: sometimes it works (with an other setup) and sometimes not - sometimes doxygen complains about aliases which are defined but not found… etc. -> part of the output was generated - part not -> it was a total "random" error.

made me think a bit ...

I tried a similar example with /// and /** comments and

• with /// I didn't get a message but the comment was silently omitted
• with the /** comment we got the warnings and also some output.

aa.h

/// \file

/// the fie
///
///   \RNS{system_primitive_types      "PrimitiveTypes"  ,<br>
///
void fie(void);

bb.h

/** \file

    the fie

      \RNS{system_primitive_types      "PrimitiveTypes"  ,<br>
*/
void fie1(void);

Doxyfile:

ALIASES += "RNS{1}=RNS \1"
QUIET=YES

The warnings:

.../bb.h:9: warning: Reached end of file while still inside a (nested) comment. Nesting level 1 (probable line reference: 1)
.../bb.h:7: warning: Member fie1(void) (function) of file bb.h is not documented.
.../bb.h:5: warning: Found unknown command '\RNS'

I would have expected that also in the case of /// there would have been a message.

Example: example.tar.gz

[aotto1968]

I use the same code with multiple targets for ONE target I get the

warning: Reached end of file while still inside a (nested) comment

and for the other target not - I don't get the message:

warning: Found unknown command '\RNS' 

but a lot of other like:

warning: Found unknown command '\function'

The "function" alias has nothing todo with the "RNS" alias -> probably a scaling problem and a memory overrun.

I use the /** ... */ syntax in this file.