Skip to contentSkip to conversation

Trouble building perfbook

DISCUSSION Andrew Donnellan 2 participants 4 messages Oct 13, 2015 – Oct 14, 2015

lore
Share & tools
lore
#0 lore raw

Hi perfbook list,

I'm having trouble building perfbook (latest git) - am encountering a weird fatal TeX error. I'm using the current TeXLive 2015 from Debian stable.

The TeX logfile is available at [0].

I've taken a look at the section of perfbook_flat.tex that it refers to, and I can't see anything particularly out of the ordinary. My LaTeX-fu isn't quite good enough to figure out what's going on here - any assistance would be appreciated!

Andrew

[0] http://pastebin.com/r4rMzF4d

Signature · 3 lines
Andrew Donnellan              Software Engineer, OzLabs
[email protected]  Australia Development Lab, Canberra
+61 2 6201 8874 (work)        IBM Australia Limited
#1 lore raw
Quoted 14 lines · Andrew Donnellan

Hi perfbook list,

I'm having trouble building perfbook (latest git) - am encountering a weird fatal TeX error. I'm using the current TeXLive 2015 from Debian stable.

The TeX logfile is available at [0].

I've taken a look at the section of perfbook_flat.tex that it refers to, and I can't see anything particularly out of the ordinary. My LaTeX-fu isn't quite good enough to figure out what's going on here
- any assistance would be appreciated!

Quoted 4 lines

Andrew

[0] http://pastebin.com/r4rMzF4d

This looks like the same LaTeX bug that Borislav Petkov ran into some time back. He came up with the workaround below, does this work for you?

Just out of curiosity, what distro are you using? I haven't run into this myself on either Ubuntu 12.04 or 14.04.

Thanx, Paul

------------------------------------------------------------------------

Apparently, it is still an open bug which seems unresolved even in tex upstream, AFAICT:

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=752773

I tried the suggested workaround first by replacing \lstinline in

/usr/share/texlive/texmf-dist/tex/latex/listings/listings.sty

with it but it complained about an undefined command. Then I went and did this:

\newcommand\lstinline[1][]{%
\leavevmode\bgroup % \hbox\bgroup --> \bgroup

     \def\lst@boxpos{b}%
     \lsthk@PreSet\lstset{flexiblecolumns,#1}%
     \lsthk@TextStyle
     \ifnum\iffalse{\fi`}=\z@\fi
     \@ifnextchar\bgroup{%
        \ifnum`{=\z@}\fi%
        \afterassignment\lst@InlineG \let\@let@token}{%
        \ifnum`{=\z@}\fi\lstinline@}}

and the pdf was built successfully.

This this goes into

/usr/share/texlive/texmf-dist/tex/latex/listings/listings.sty

which is part of texlive-latex-recommended. It needs to be fixed in the distro.

#2 lore raw
Quoted 3 lines · Paul E. McKenney

This looks like the same LaTeX bug that Borislav Petkov ran into some time back. He came up with the workaround below, does this work for you?

Appears to work - thanks!

Quoted 2 lines

Just out of curiosity, what distro are you using? I haven't run into this myself on either Ubuntu 12.04 or 14.04.

Debian 8.2 without backports. As it turns out that shipped with TeXLive 2014.20141024. It appears the Debian bug you linked was fixed in 2015.20150625, which my home machine running Debian testing would have picked up by now, but I generally prefer more stable things for my IBM machine...

Andrew

Signature · 3 lines
Andrew Donnellan              Software Engineer, OzLabs
[email protected]  Australia Development Lab, Canberra
+61 2 6201 8874 (work)        IBM Australia Limited
#3 lore raw
Quoted 6 lines · Andrew Donnellan
Quoted 3 lines · Paul E. McKenney

This looks like the same LaTeX bug that Borislav Petkov ran into some time back. He came up with the workaround below, does this work for you?

Appears to work - thanks!

Very good!

Quoted 8 lines
Quoted 2 lines

Just out of curiosity, what distro are you using? I haven't run into this myself on either Ubuntu 12.04 or 14.04.

Debian 8.2 without backports. As it turns out that shipped with TeXLive 2014.20141024. It appears the Debian bug you linked was fixed in 2015.20150625, which my home machine running Debian testing would have picked up by now, but I generally prefer more stable things for my IBM machine...

Well, at least a fix is in the works. ;-)

Thanx, Paul