[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Emacs-diffs] /srv/bzr/emacs/trunk r111707: * doc-view.el: Use (and

From: Tassilo Horn
Subject: Re: [Emacs-diffs] /srv/bzr/emacs/trunk r111707: * doc-view.el: Use (and prefer) soffice as default ODF->PDF
Date: Tue, 12 Feb 2013 10:59:42 +0100
User-agent: Gnus/5.130006 (Ma Gnus v0.6) Emacs/24.3.50 (gnu/linux)

"Stephen J. Turnbull" <address@hidden> writes:

>  > >> I'm not sure I get the logic here: you need to have soffice
>  > >> installed if you want to use unoconv, but being able to use
>  > >> unoconv instead of soffice is good.
> That's true for the *user* at the command line, but for Emacs it
> depends on how unoconv is being used.  If the UI provides a Swiss Army
> knife able to convert any given format to any other by delegating the
> format grokking to unoconv, yes, it's sort of good, assuming that
> Emacs will never pass incorrect arguments to unoconv.[1] In that case
> unoconv (and any other "universal converters") should be tried only
> after specific converters that do a better job are tried.[2]
> If in fact what happens is that Emacs parses the formats and only
> handles formats that it recognizes, unoconv is useless to Emacs.  (I
> suspect this is rarely what most users want[3], but I prefer it.  If
> Emacs can't do a good job automatically, it should leave it up to me!)

Ok, to clear up the situation a bit: unoconv is probably more powerful
than soffice.  For example, it is possible to run it as a server that
accepts conversion requests from multiple clients.  doc-view doesn't
need that.

doc-view just uses either unoconv or soffice to convert OpenDocument or
Microsoft Office documents to PDF.  Both tools can do that just fine,
and they use the same LibreOffice/OpenOffice internal APIs for the job.

I've converted some example documents with both of them, and the output
is almost identical: identically displayed, identical with respect to
file size of the generated PDF, but diff says they differ nonetheless.

However, soffice is appoximately 2.5 to 7 times faster.  It seems that
unoconv has some significant startup overhead compared to soffice.  I
hadn't expected that, but in the end it justifies that soffice is picked
up by default when both soffice and unoconv are installed on a system.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]