Re: possible Global bug - not picking up symols?

[Shigio YAMAGUCHI]

Hello,
> search->    global dx_
> finds  ->    src/dataframe.cpp
>
> but
>
> search->    global index_
> finds  ->    (finds nothing)

Both of the above seem to be error.
Currently, gtags's C++ parser does not know 'member initialization list' of C++.
So, gtags accidentally treats the last initialization 'dx_(...)' as a function
definition. I have added the following item to the Known Bug list.

Known Bug list:
Gtags(Cpp parser) does not support member initialization list. [report]

Thanks,
Shigio

2020年4月30日(木) 23:42 jan <address@hidden>:
>
> Hi,
> this may be a bit difficult as I don't know how Global parses stuff.
> I assume it does less than a compiler, but does more than a simple
> parser as you are trying to keep relationships between symbols ("x is
> declared here and used there").
> So I'm not sure what to expect.
>
> Anyway, the easiest way is just to git clone gromacs from
> <https://github.com/gromacs/gromacs> and run gtags on it. You know
> that it is complete and compiles (use the v2020.1 tag, which does
> compile. HEAD may not). Then just run gtags and htags on it.
>
> Then in the HTML that htags has made go  to
> /src/gromacs/analysisdata/dataframe.cpp and you'll see stuff that is
> not underlined.
>
> ---
>
> However, that is not a minimal example (and gromacs is big), so
> instead get hold of these 2 files
>     dataframe.cpp
>     dataframe.h
>
> from <https://github.com/gromacs/gromacs/tree/master/src/gromacs/analysisdata>
>
> and put them in a directory "src" and run gtags and htags on them.
> Then search. This finds something, correctly:
>
> search->    global dx_
> finds  ->    src/dataframe.cpp
>
> but
>
> search->    global index_
> finds  ->    (finds nothing)
>
> Both index_ and dx_ are defined in 'dataframe.h':
>
>     class AnalysisDataFrameHeader
>     {
>     ...
>     private:
>         int  index_;
>         real x_;
>         real dx_;
>     };
>
> So they should be found. As I understand it, anyway.
>
> I've attached these 2 files but they may not get through as mailing
> lists tend to delete attachments.
>
> ---
>
> Is this OK? If you want it even simpler I can set up a simple github
> repo with just these files which you can pull from.
>
> Happy to do that if it makes things easier - let me know.
>
> thanks
>
> jan
>
>
> On 30/04/2020, Shigio YAMAGUCHI <address@hidden> wrote:
> > Hello,
> > Could you please give me a reproduce method to understand the problem?
> > It would be helpful if you could describe it as concisely as possible.
> >
> > Thanks,
> > Shigio
> >
> > 2020年4月30日(木) 21:26 jan <address@hidden>:
> >>
> >> (if anything below is unclear, please shout!  I will try to explain
> >> better)
> >> (also please acknowledge that it's been received, just so I know it's
> >> not got lost)
> >>
> >> Hi,
> >> this bug (if it is a bug) seems to have been around a long time. I can
> >> see a reference to something like this here, from 2007
> >> <https://lists.gnu.org/archive/html/bug-global/2007-10/msg00004.html>
> >>
> >> I installed Global yesterday, as part of setting up emacs as a proper
> >> IDE,  and started to play. I built the tag table and the very first
> >> thing I did was pick a random symbol and look for it.
> >> Global couldn't find it, which caused me a couple of hours more
> >> playing and reading the manual. I still don't get it. Maybe it's me.
> >>
> >> I am also very, very rusty with C++ so I may be getting things wrong
> >> there.
> >>
> >> The project is gromacs <http://www.gromacs.org/>.
> >>
> >> The size of the project source is 43.6MB. The total size of
> >> gpath/grtags/gtags is 17.4MB.  The documentation implies these would
> >> be about the size of the original source file, but this is ~1/3rd the
> >> size so I'm surprised.
> >>
> >> I looked for a constructor
> >> "AnalysisDataPointSetRef::AnalysisDataPointSetRef". It can't find this
> >> or even "AnalysisDataPointSetRef" by itself. It can't seem to find a
> >> number of symbols.
> >>
> >> If I generate the HTML it
> >> AnalysisDataPointSetRef::AnalysisDataPointSetRef shows up, but without
> >> underlining. In fact quite a number of symbols show up without
> >> underlining in that source file. The following is from the HTML file
> >> with recognised (underlined) items made upper case.
> >>
> >> AnalysisDataFrameHeader::AnalysisDataFrameHeader
> >>  (int index, REAL X_, REAL dx) :
> >>     index_(index),
> >>     X_(x),
> >>     DX_(dx)
> >> {
> >>     GMX_ASSERT(index >= 0, "Invalid frame index");
> >> }
> >>
> >> Non underlined items are
> >> AnalysisDataFrameHeader::AnalysisDataFrameHeader, index_, index, x and
> >> dx. Therefore it seems not to be picking these up?
> >>
> >> Looking for these it can find the underlined ones:
> >>
> >> >>>global dx_
> >> gromacs_working/src/gromacs/analysisdata/dataframe.cpp
> >>
> >> but
> >>
> >> >>>global index
> >> (nothing)
> >>
> >>
> >> From the same file, another example; similar constructor but
> >> conventional parameters are being inconsistenly picked up. As before
> >> underlined items are made upper case here:
> >>
> >> AnalysisDataPointSetRef::AnalysisDataPointSetRef
> >>                         (const AnalysisDataFrameHeader&  HEADER,
> >>                   const AnalysisDataPointSetInfo& pointSetInfo,
> >>                   const ANALYSISDATAVALUESREF&    values)
> >>
> >> So it can see HEADER and ANALYSISDATAVALUESREF but apparently nothing
> >> else?
> >>
> >> OS is windows 2008R2.
> >> Global version is 6.6.3
> >> The project compiles cleanly so it's not been confused by syntax errors.
> >>
> >> Happy to help further.
> >>
> >> thanks
> >>
> >> jan
> >>
> >
> >
> > --
> > Shigio YAMAGUCHI <address@hidden>
> > PGP fingerprint:
> > 26F6 31B4 3D62 4A92 7E6F  1C33 969C 3BE3 89DD A6EB
> >



-- 
Shigio YAMAGUCHI <address@hidden>
PGP fingerprint:
26F6 31B4 3D62 4A92 7E6F  1C33 969C 3BE3 89DD A6EB