reproduce-devel
[Top][All Lists]
Advanced

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

[bug #58719] Crash when packing a project with dist if no tikz figures w


From: Mohammad Akhlaghi
Subject: [bug #58719] Crash when packing a project with dist if no tikz figures were generated
Date: Tue, 7 Jul 2020 09:28:12 -0400 (EDT)
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Firefox/78.0

Update of bug #58719 (project reproduce):

                  Status:                    None => Fixed                  
             Assigned to:                    None => makhlaghi              
             Open/Closed:                    Open => Closed                 

    _______________________________________________________

Follow-up Comment #1:

Thanks a lot for pointing this out Raul, a fix has been pushed to the Maneage
branch as Commit e1f10ac4516f
<http://git.maneage.org/project.git/commit/?id=e1f10ac4516f>. From now on, if
the project doesn't have any PDF files in the 'tex/tikz' directory, they will
not be copied into the distribution tarball :-). 

Its a simple fix and I also tested it, so I am closing this bug. But it would
be great if you can try it out in your project and report any possible
problems it may have ;-).

But generally, like all the other low-level tools we are advocating with
Maneage (which aren't commonly used, like Git or Make), we should also be
actively encouraging PGFPlots <https://ctan.org/pkg/pgfplots> as an optional
feature (its not mandatory to use Maneage in a project like Git or Make, but
it really helps). Here are the reasons: 

* The best thing about PGFPlots from a reproducibility perspective is that *it
has no dependencies* beyond LaTeX, thus avoiding the MANY dependencies of
plotting tools like Matplotlib (as shown in Figure 1, page 5 of Alliez+2019
<https://hal.archives-ouvertes.fr/hal-02135891v2/document>). 

* With PGFPlots the figure's final data tables go into LaTeX (thus encouraging
people to preserve the final figure datasets, not just the final figures). As
opposed to Python programs using something like Matplotlib, because people
often do some internal analysis in the same Python program that draws the
plots, so they are less likely to preserve the actual plotted table columns. 

* It has a wonderful manual
<http://mirrors.ctan.org/graphics/pgf/contrib/pgfplots/doc/pgfplots.pdf>, with
complete examples for every plot and several nice tutorials. 

* Finally, because it is generated by the same graphical engine as the paper,
I would argue that it plots are much more aesthetically pleasing to the eye.
Although I do acknowledge that this last point may be subjective ;-).

So, generally I agree with you that PGFPlots isn't commonly used my many
people and we don't want to increase the entry barrier any more than it
currently is. So its great we fixed this bug. 

But we should really try our best to demonstrate the usefulness of PGFPlots
(from the objective point of reproducibility problems of tools like Matplotlib
and encouraging data sharing and also the subjective points) in future
tutorials and example ;-).

    _______________________________________________________

Reply to this item at:

  <https://savannah.nongnu.org/bugs/?58719>

_______________________________________________
  Message sent via Savannah
  https://savannah.nongnu.org/




reply via email to

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