Skip to content

Fix bug with wrongly checking all files inside archive input#814

Merged
brucemiller merged 1 commit into
brucemiller:masterfrom
dginev:archive-input-extra-ignores
Nov 14, 2016
Merged

Fix bug with wrongly checking all files inside archive input#814
brucemiller merged 1 commit into
brucemiller:masterfrom
dginev:archive-input-extra-ignores

Conversation

@dginev

@dginev dginev commented Nov 14, 2016

Copy link
Copy Markdown
Collaborator

This is an embarrassing oversight on my part - i had swapped the places of the grep/map pair, grepping on perl objects, instead of the file name held by them.

That lead to archives with no TeX file from the arXiv corpus to be falsely searched through, and in fact often .cls and .sty files were being converted in the absence of a real source, due to this error.

An example directory tree I just tested with looks like:

-rw-r--r-- 1 dreamweaver dreamweaver  41139 Jun 29  2004 llncs.cls
-rw-r--r-- 1 dreamweaver dreamweaver  12121 Jun 29  2004 mathpartir.sty
-rw-r--r-- 1 dreamweaver dreamweaver   4209 Jun 29  2004 welldef.bbl
-rw-r--r-- 1 dreamweaver dreamweaver  87850 Jun 29  2004 welldef.tex.cry

As the authors used an encrypted .cry file, this archive contains no usable tex source and should be marked invalid by latexml. Which was the intention all along, and is indeed what happens with the patch here. But with the current master, converting the archive holding these files identifies llncs.cls as the main TeX file and tries to convert it, failing with a Fatal.

I'd appreciate a quick merge, would be happy to rerun the unexpected Fatals, which look to be largely due to this oversight. There may be other harder to detect cases of this, but they'll come to light in further reruns.

@brucemiller
brucemiller merged commit cb55411 into brucemiller:master Nov 14, 2016
@brucemiller

Copy link
Copy Markdown
Owner

makes sense.

@dginev
dginev deleted the archive-input-extra-ignores branch February 25, 2018 18:07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants