bugGNU Octave - Bugs: bug #52773, PDF user manual breaks up index...


bug #52773: PDF user manual breaks up index entries for graphics properties in strange way

Submitted by:  Dan Sebald <sebald>
Submitted on:  Mon 01 Jan 2018 03:53:34 AM UTC  
Category:  Documentation Severity:  3 - Normal
Priority:  5 - Normal Item Group:  Documentation
Status:  Works For Me Assigned to:  None
Originator Name:  Open/Closed:  Closed
Release:  dev Operating System:  Any

Add a New Comment(Rich Markup)

You are not logged in

Please log in, so followups can be emailed to you.


Thu 04 Jan 2018 12:07:04 PM UTC, comment #4:

On my Fedora fc27 system using okular I find both
figure deletefcn and axes deletefcn. Of course
"figure deletefcn" is not found as there is no
"figure deletefcn" string in the file.

There are a few other items that need attention.
I will submit a new bug report for them.

Essentially, in my system, the Manual looks exactly like
recent versions.

Michael Godfrey <godfrey>
Project Member
Thu 04 Jan 2018 09:34:59 AM UTC, comment #3:

It's very likely that your build of octave.pdf has a different layout for pages, i.e., the fonts, etc. all affect the amount of words on a page. In that case you may not be seeing the same problem because the "figure ..." doesn't span down a whole column.

I've installed okular here and that doesn't find "figure deletefcn" or "axes deletefcn" for me either. It's in the PDF file itself. (BTW, okular's search algorithm is rather slow compared to xpdf and xreader.)

The best I can describe this is that it's something similar to how an HTML viewer works (like Firefox or whatever). So often one tries to mouse-select a sentence on the page for copy/paste but upon selecting that sentence there is all of the text down/up/across the whole next column. That happens because internally the HTML code may be contiguous in one way but once decoded in a graphical form its layout might no longer be contiguous. It happens a lot when there are sidebars for adverts and such.

So, what I showed in the screen shot is a mouse-select where I try to select "figure deletefcn" but then it selects the bottom half of the first column and top half of the second column. (BTW, xreader seems to be the only viewer of the three that has this mouse-select capability...which I'm pretty sure is something that acroread has as well.)

Let's close this report. What I described is too arcane to find anything related in a web search. I've done a bit of searching and read dvipdf man page. Nothing seems to ring a bell.

I've actually done a little bit of coding with PostScript and my recollection is that there is a means to define some object (text or image) and repeatedly place it on the page, and this is something commonly done because the original issue with PostScript was bulkiness.

Dan Sebald <sebald>
Thu 04 Jan 2018 06:38:49 AM UTC, comment #2:

I'm not sure I see what the problem is. I use okular from KDE to view PDFs and I can search for "figure deletefcn" and it is found without trouble.

Rik <rik5>
Project Administrator
Tue 02 Jan 2018 02:46:38 AM UTC, comment #1:

I've looked at this one just a bit. I've confirmed that xpdf has the same behavior as xreader. It's clear that the Texinfo input doesn't matter. I.e., this

to use

i.e., curly parentheses around what should be treated as a lumped string of text, outright fails.

It's strange. I recompiled octave.pdf and the blocky behavior I illustrated in the original post's PNG figure has gone away for 'figure deletefcn'. It's because the index was reduced slightly (with what I was working on) and no longer is there a whole column of entries that being with "figure". That is, no longer is there this arrangement:

By "block" I mean like a letter group A is all one block, B is all one block, etc.

So, I would guess this is either a TeX behavior or a dvi2pdf behavior. I wonder why it is doing this. Is it because the "figure etc" originates from a table? Or is the responsible program to trying to compress the data as much as possible by locally defining "figure" in PostScript/PDF and replicating that object thirty times.

I will look for some command line switch for tex2dvi or something similar. Otherwise, I think this is something out of Octave's hands.

Dan Sebald <sebald>
Mon 01 Jan 2018 03:53:34 AM UTC, original submission:

There are automatically generated entries for graphics properties in the indeces. For example

figure colormap
figure createfcn
figure currentaxes

However, it turns out those are grouped like a table, so it is figure, figure, figure, ..., colormap, createfcn, currentaxes, ... as viewed by the xviewer search engine. I'm not sure if that is a LaTeX matter or whether LaTeX is being provided the input in that particular layout. It would be nice to group that as "figure colormap", "figure createfcn", etc. so that the search engine works to find those.

I happened to search for "deletefcn" when working on some examples for the manual and was surprised there was nothing there because I made reference to it. But the situation above prevented finding the entry; the attached screenshot illustrates the issue.

Dan Sebald <sebald>


(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:


Depends on the following items: None found

Items that depend on this one: None found


Carbon-Copy List
  • -email is unavailable- added by godfrey (Posted a comment)
  • -email is unavailable- added by rik5 (Posted a comment)
  • -email is unavailable- added by sebald (Submitted the item)

    Do you think this task is very important?
    If so, you can add your encouragement to it.
    This task has 0 encouragements so far.

    Only project members can vote.




    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2018-01-04 rik5 Open/ClosedOpen => Closed
    2018-01-04 rik5 StatusNone => Works For Me
    2018-01-01 sebald Attached File- => Added index_layout_figure_deletefcn_Screenshot_from_2017-12-31_21-45-18.png, #42779

    Back to the top

    Powered by Savane 3.3