#636 Viewing pdf causes a popup window to appear

None
closed
nobody
None
1
2014-08-24
2012-11-17
Alex Koudrin
No

In TS 2.5 when I compile and view PDF output I small window pops up to say that the pdf "does not look like a valid PDF document" and may be corrupt. The pdf still comes up, I just need to click Yes. It possibly has something to do with the fact that my document is not small (about 100 pages with 50 vector diagrams), but still it takes about 30-50 seconds to compile and view and it was fine in previous version.

The command I'm running is:

txs:///latex | txs:///makeindex | txs:///bibtex | txs:///latex | txs:///dvi-pdf-chain

I've attached the screenshot. I get this popup both on WinXP and Win7.

1 Attachments

Discussion

  • Tim Hoffmann
    Tim Hoffmann
    2012-11-18

    Yes, likely the viewer tries to load the pdf while it it not yet complete.

    Do you call view explicitly? If not, at which part of your command chain does the message appear?

     
    Last edit: Tim Hoffmann 2012-11-18
  • Alex Koudrin
    Alex Koudrin
    2012-11-18

    I agree, it is as if the internal viewer is trying to open the pdf too early, rather than first checking that the pdf has finished building.

    The Build & View command as above

    txs:///latex | txs:///makeindex | txs:///bibtex | txs:///latex | txs:///dvi-pdf-chain

    compiles everything and brings up the viewer. That is,

    txs:///dvi-pdf-chain

    expands to

    txs:///pdflatex | txs:///view-pdf

    with PDF Viewer configured as txs:///view-pdf-internal

     
  • Tim Hoffmann
    Tim Hoffmann
    2012-11-18

    so it happens in the call of txs:///view-pdf?

    Could you please verify this by temporarily removing this command from txs:///dvi-pdf-chain?

     
  • Alex Koudrin
    Alex Koudrin
    2012-11-18

    Yes, I've just tried reproducing it on this WinXp machine at no avail. Maybe I was wrong and it only happened on my Win7 machine at home. I'll do some experimentation in Win7 tonight (in 12 hours or so).

    Perhaps Win7 has something about it that make view-pdf to be called before pdflatex finishes. I'll let you know how I go.

     
  • Alex Koudrin
    Alex Koudrin
    2012-11-18

    I've managed to reproduce it here on WinXP.

    First, I'll correct one mistake:

    txs:///latex | txs:///makeindex | txs:///bibtex | txs:///latex | txs:///dvi-pdf-chain

    The

    txs:///dvi-pdf-chain

    expands to

    txs:///latex | txs:///dvipdf | txs:///view-pdf-internal

    After some edits to my latex file, the popup windows started to appear. I made sure I could reproduce it. Then I removed txs:///view-pdf-internal from the dvi-pdf-chain and it worked fine (but I had to start the viewer myself, of course). I then put the command back and watched the Messages window. Indeed I see:

    Process started: dvipdfm "master".dvi

    and then immidiately the internal pdf viewer appears (with nothing in it yet) and then the window pops up. Only after another 10 seconds the Messages window shows:

    Process exited normally

    And then the output appears in the dvi viewer.

    I've experimented keeping my original build config and found:

    The popup does NOT appear if I:
    - Close the viewer and then build.
    - Do a full clean of auxiliary files (Alt-F1) even if you leave the viewer open.

    The popup DOES appears if I:
    - Do a build other than a fresh build after a full clean of auxiliary files AND have the viewer open already. Sometimes I had to slightly modify and save one latex file to get the popup to appear. This one seems a little temperamental, but mostly reproducable.

    I hope this is helpful and gives you enough to go on to debug it.

     
  • in your configuration file application data/texstudio/texstudio.ini, please edit the line
    "Tools\Kind\Pdf=pdflatex, xelatex, lualatex, latexmk" to "Tools\Kind\Pdf=pdflatex, xelatex, lualatex, latexmk, dvipdf"

    Does the problem still occur ?

     
  • Tim Hoffmann
    Tim Hoffmann
    2012-11-21

    fixed by rev. 3473?

     
    • status: open --> closed
    • milestone: -->
     
  • no feedback, let's assume, that the fix works