Looks like a couple of latex-related failures, I'll retry them with texlive-full tomorrow.

!rebuild models/04-diffeqbio_II_networkproperties.jmd

Created pipeline: see it here.

Same error even with texlive-full.

Must be a Weave issue.

If the PDF fails, does the full build not upload?

Correct. We could do something like julia -e '...' || true to ignore failures but that seems bad. Maybe the weave_file function could catch errors and warn instead of erroring?

Maybe we just shouldn't weave the PDFs since the markdown files probably are more useful

Seeing as failure seems pretty rare, maybe we ought to just put the pdf part in a try/catch?

Yeah we can do that too

!rebuild models/04-diffeqbio_II_networkproperties.jmd

Created pipeline: see it here.

!rebuild models/04b-diffeqbio_III_steadystates.jmd

Created pipeline: see it here.

Would be nice to be able to do multiple individual files like !rebuild models/first.jmd models/second.jmd

Looks like that last job removed all the output from the notebook (-6000 lines)

https://github.com/SciML/DiffEqTutorials.jl/blob/rebuild/b4b24f65/notebook/models/04-diffeqbio_II_networkproperties.ipynb

edit: but the notebook for models/04b-diffeqbio_III_steadystates.jmd turned out fine

The notebooks don't really need the output

Is there a reason why we only get one .md?

Hm, that's interesting. But there's only one file in https://github.com/SciML/DiffEqTutorials.jl/tree/master/markdown/models not including one for the now-moved pendulum tutorial so it seems like a preexisting issue.

The markdown files are new so the only ones that exist are ones that I manually created about three months ago or something like that, but one of the things I hope to happen is that they all get generated in this new round of generation.

Ah, I see. And now that I think of it, I know why they're not being pushed -- it only stages updated files, not new (and hence untracked) files to avoid committing junk. I think maybe we can just stage everything and tweak the gitignore as necessary.

!rebuild models

this should include new files now

Created pipeline: see it here.

Hahahahahahahaha yikes. The CI setup puts the Julia depot in the current directory to facilitate caching. I can work around that though.

Waiting on JuliaGPU/gitlab-ci#20 and then I can address JuliaGPU/gitlab-ci#25 which will make it okay to stage all files.

!rebuild models

ok let's see if i fixed that properly

Created pipeline: see it here.

!rebuild models

we're almost there

!rebuild models

oops

Created pipeline: see it here.

08-kolmogorov_equations.jmd has some errors https://gitlab.com/JuliaGPU/DiffEqTutorials.jl/-/jobs/616218638#L3325

All jobs except for 08 are successful, that one exposed a small bug that I've fixed -- and the tutorial itself seems to have some bugs.
Maybe we can merge this now and rebuild 08 separately.

Yeah, 08 we never fully tested it (@ashutosh-b-b)