Build of LaTeX documentation fails with ! Dimension too large. in bitbuffer_8h.tex.

[dreirund]

After workarounding #2825, build of LaTeX documentation fails:

[131]) (./baseband_8h_source.tex [132] [133]) (./bitbuffer_8h.tex
<bitbuffer_8h__incl.eps>
! Dimension too large.
<to be read again> 
                   \relax 
l.22 ...hics[width=350pt]{bitbuffer_8h__dep__incl}

Build procedure:

---

git clone https://github.com/merbanan/rtl_433.git
cd rtl_433

Now patch Doxyfile.in with

diff -rU1 rtl_433.orig/Doxyfile.in rtl_433/Doxyfile.in
--- rtl_433.orig/Doxyfile.in	2024-02-07 15:59:47.959870493 +0100
+++ rtl_433/Doxyfile.in	2024-02-07 16:00:06.606536622 +0100
@@ -793,3 +793,3 @@
 
-WARN_AS_ERROR          = YES
+WARN_AS_ERROR          = NO

Continue build:

cmake -DBUILD_DOCUMENTATION=ON .
make
cd doc
make

---

More output of the last make command:

[...]
Package longtable Warning: Column widths have changed
(longtable)                in table 6.3 on input line 319.


Underfull \hbox (badness 10000) detected at line 320
[][][]
[131]) (./baseband_8h_source.tex [132] [133]) (./bitbuffer_8h.tex
<bitbuffer_8h__incl.eps>
! Dimension too large.
<to be read again> 
                   \relax 
l.22 ...hics[width=350pt]{bitbuffer_8h__dep__incl}
                                                  
?  

[zuckschwerdt]

I don't see this exact error, but the mix of 8-bit ASCII (some unknown variants of ISO-8859) and UTF-8 seems to be a problem. Not sure how we could deal with this? Mandate e.g. ISO-8859-1 or UTF-8? And how to declare that in the files /to Doxygen?

[zuckschwerdt]

I have now cleaned out some of the more unexpected characters. The only remaining chars that upset LaTeX are now ≤, ₀, π in acurite.c and tpms_renault_0435r.c. They look nice, but I guess need replacing to.

[dreirund]

I have now cleaned out some of the more unexpected characters. The only remaining chars that upset LaTeX are now `≤`, `₀`, `π` in acurite.c and tpms_renault_0435r.c. They look nice, but I guess need replacing to.

I don't know by myself, but there are LaTeX variants (LuaLaTeX? XeTeX?) that can natively handle UTF-8 input. Maybe switching to that? Anyway, using ASCII-only in code (and comments) is always a good idea for portability reasons :-).