Xcolor problem #1
Comments
|
Thanks for the clear test file, I can reproduce the problem, we will see
what we can do....
(mis-matched group errors were the "expected" failures which caused us to
set up this repository but tabu is a big package and finding all the places
it needs updating is tricky so test cases like this are very helpful)
|
|
@davidcarlisle Thanks so much for your reply. In 2019 (I gues since then I have the problem), I have downloaded upgrades for the following packages: |
|
I have a similar problem, here is my MWE: As soon as one removes the call to color the compilation works OK. Interestingly downgrading tabu to v2.8 does not solve my problem. Nevertheless I tend to say that the issue is in the tabu package. Hilmar |
|
Any progress on this issue? Multiple project run into problems (e.g. https://bugs.debian.org/920621). |
|
Version 2.9 seems to fix the xcolor conflict for fixed width (positive coefficient) columns, but not for flexible width (negative coefficient) columns. |
|
I have a similar issue where the row colors are not showing up at all. I just updated all my packages from ctan and (due to other errors i wanted to fix). Before the alternating row colors worked (but tabu nesting did not). Now the nesting errors are gone but so are the colors. I also tried it with the 2.10 release straight from github. My MWE: |
|
Any reason why on CTAN version 2.10 is not present? (I still see: Version 2.9 2019-01-11) |
|
@albert-github I had just submitted 2.10 to ctan but before they had unpacked it several more issues got posted here which suggested that 2.10 may have made some things worse as well as fixing some things, so I asked them to cancel the update. I haven't had time to look at tabu in the last couple of weeks so things are as they are, sorry. I set up this github so issues can be tracked and updates suggested, I can also do some fixes myself as time allows but I can't promise any fixed schedule. the original author posted as far back as 2011 that the package was unsupported. |
|
@davidcarlisle Thanks for the information, I know time is limited and I do appreciate the effort you and others put into the tabu package. |
|
I am experiencing a similar issue with the following MWE: Since I am operating a CTAN mirror with daily snapshots of the entire repository contents since some years, I could trace the problem down to an incompatible change of the I could figure out that the problems seem to be related to changes in That suggests the assumption that the Intermediate WorkaroundThe following is just an intermediate workaround, which solved the issue at least for my documents. I don't claim that this is a permanent fix, but it might allow to continue work on documents until proper fixes for The versions of the For TeX Live:
For MiKTeX:
|
|
I'd just like to put a smaller MWE. Comment the |
|
Another testcase (from https://tex.stackexchange.com/a/468958/43807): It also breaks if you replace Another testcase (from Debian #920459): This one also breaks with … because, here, the triggering point is negative X coefficient. I believe I have a fix for this, because all those examples work with it. I’m a bit scared of breaking anything currently working, though, so immediate review would be massively welcome. |
|
Not sure if that helps, but this issue is reproducible when using \documentclass{article}
\usepackage{tabu}
\usepackage{colortbl}
\begin{document}
\begin{tabu}{X[-1]}
hello
\end{tabu}
\end{document} |
|
@vladimiroff colortbl loads color thus it shows the problems caused by color (but those should now be mended with the code already in the repository). However, there are additional issues with colortbl (because the fix so far doesn't take into account changes made by that package to array package internals) and those still need to be addressed, e.g. #4, #11, #13. |
|
And here is another group of test cases, which
\documentclass{article}
%% Comment "xcolor" line then error "Missing } inserted" disappears.
%% Loading "color" instead of "xcolor" raises the same error.
\usepackage{xcolor}
\usepackage{tabu}
\begin{document}
% alright
\begin{tabu}{X}
bar
\end{tabu}
% alright
\begin{tabu} spread 100pt {l}
zz
\end{tabu}
% error
\begin{tabu} spread 100pt {X}
bar
\end{tabu}
\end{document} |
|
tabu / xcolor still broken for me MiKTeX 2.9 with tabu v2.9 (Sun Jan 20 07:11:06 2019) also tried DigNative's fix to no avail. New PC upgrade and fresh install. Been using LaTeX with tabu for years and have a large amount of documents as well as automated data report generation scripts that are reliant on it. Unless there's a fix or simple workaround I will probably be stuck on the older version of latex for some time. Here's an additional test case: |
|
@khronimo maybe have a look at how doxygen "fixed" / worked around it (doxygen/doxygen#7137) and the discussion there (doxygen/doxygen#6769). |
|
@khronimo I tried your example with a freshly installed MiKTeX with today's version of all packages and with the workaround suggested by me applied. Your example compiles fine for me, so it appears to me that the described steps have not exactly been carried out as described. To be very sure, I also tried with TeX Live, too, and can confirm that the workaround is still working for both distributions with most recent package versions. |
|
Thanks so much @DigNative I've been playing around and have finally managed to get it to work. I had to reinstall several times, even after a fresh install it wouldn't work following your instructions. Noticed it seemed to have loaded in packages automatically even though it was set to ask this time. Uninstalled and manually deleted everything from Program Data and my user AppData folders and it worked like a charm after a reinstall following your steps. |
|
Ah, I see @khronimo. I would expect that the issue on your system was caused by user-level package installations (e.g., due to packages loaded on the fly). Since these packages take precedence over the system-wide installed ones, the files for the workaround placed within the system-wide installation folder may have no effect. This concept of user- and system-level packages of MiKTeX already caused many not so trivial issues in the past for me and my colleagues, for example, when still old package versions have been used after a system-wide update of the MiKTeX installation (because the older and not updated user-level packages also in this case took precedence and where causing conflicts). |
|
I want to color a single row not alternating colors...any advice? |
|
@neuronflow |
|
@davidcarlisle oh thanks, for the quick reply and pointers! |
|
@davidcarlisle Thanks for the tips |
|
@assignUser word of caution: I question the “ |
|
@mirabilos thanks for the info, it looks like it is part of texlive though (haven't tested it yet) and the syntax for it looks a bit friendlier then tabu. But I do agree, with some patches (siunitx and threeparttable) I was able to use it for my work so far without any issues. On the other hand I would like to transition to something stable and maintained before it fails completely :D |
and how old is the Debian version of LaTeX core? |
|
Frank Mittelbach dixit:
and how old is the Debian version of LaTeX core?
texlive-base 2020.20210202-3
texlive-binaries 2020.20200327.54578-7
So, almost a year, but this is when the current stable stabilised.
This is what you’ll want to be compatible with for another ~3 years
in documents, until Debian and its derivatives have brought out new
releases and most users upgraded.
bye,
//mirabilos
--
I believe no one can invent an algorithm. One just happens to hit upon it
when God enlightens him. Or only God invents algorithms, we merely copy them.
If you don't believe in God, just consider God as Nature if you won't deny
existence. -- Coywolf Qi Hunt
|
probably more than a year then, as far as the LaTeX inside is concerned and yes the further you go back it means that you will see less issues with tabu (even though some are in there for ages). But a stable Debian is not what most users will have and our advice is not based on old releases, but on the fact that the package is not maintained for more than a decade and cuts across a lot of internal code that shouldn't have been touched (or at least not without active maintainance) and so I can only repeat David's statement: it is, with a fairly current LaTeX, totally broken and will get worse further --- it can't be recommended therefore. And even it it appears to be still working within current Debian stable it will not work once Debian updates. |
I still seem to have troubles with xcolor in my case.
compiling twice (I guess when the glossar should pop up for the first time) I get a lot of
Removing
xcolorseems to fix it.Here my code:
EDIT:
The report from above is with the latest ctan version of all my packages.
Most importantly, here a piece from the log
The text was updated successfully, but these errors were encountered: