List of papers that are not exported correctly

[villalbamartin]

There's a list of papers that fail to export, and I'm opening this ticket to keep track of what's going on. Most of these papers fail with the error NoMethodError: undefined method 'title' for nil:NilClass, which is an error that doesn't make sense because those papers do have titles. Seeing as how most of them end in 000 (see the list down below), I believe there's an issue with papers that represent volumes rather than single entries.

I'll look deeper into this.

Here's the list of paper IDs that fail to export properly:

• C69: 0100, 0101, 0200, 0201, 0300, 0301, 0400, 0401, 0500, 0501, 0600, 0601, 0700, 0701, 0800, 0801, 0900, 0901
• J00: 1000, 2000, 3000, 4000
• J01: 1000, 2000, 3000, 4000
• J02: 1000, 2000, 3000, 4000
• J03: 1000, 2000, 3000, 4000
• J04: 1000, 2000, 3000, 4000
• J05: 1000, 2000, 3000, 4000
• J06: 1000, 2000, 3000, 4000
• J07: 1000, 2000, 3000, 4000
• J08: 1000, 2000, 3000, 4000
• J09: 1000, 2000, 3000, 4000
• J10: 1000, 2000, 3000, 4000
• J11: 1000, 2000, 3000, 4000
• J12: 1000, 2000, 3000, 4000
• J13: 1000, 2000, 3000, 4000
• J14: 1000, 2000, 3000, 4000
• J15: 1000, 2000, 3000, 4000
• J16: 1000, 2000, 3000, 4000
• J17: 1000, 2000, 3000, 4000
• J18: 1000, 2000, 3000, 4000
• J79: 1000
• J80: 1000, 2000, 3000
• J81: 1000, 2000, 3000, 4000
• J82: 1000, 2000, 3000
• J83: 1000, 2000, 3000
• J84: 1000, 2000, 3000
• J85: 1000, 2000, 4000
• J86: 1000, 2000, 3000, 4000
• J87: 1000, 3000
• J88: 1000, 2000, 3000, 4000
• J89: 1000, 2000, 3000, 4000
• J90: 1000, 2000, 3000, 4000
• J91: 1000, 2000, 3000, 4000
• J92: 1000, 2000, 3000, 4000
• J93: 1000, 2000, 3000, 4000
• J94: 1000, 2000, 3000, 4000
• J95: 1000, 2000, 3000, 4000
• J96: 1000, 2000, 3000, 4000
• J97: 1000, 2000, 3000, 4000
• J98: 1000, 2000, 3000, 4000
• J99: 1000, 2000, 3000, 4000

[knmnyn]

We probably want to exclude the J and Q volumes from this process since they (CL and TACL) don't have such frontmatter typically accompanies both workshop and conference venues. That should be pretty easy a fix to add a conditional for those.

[mjpost]

Thanks for cataloguing this! Ideally we would not print the PDF (or bib) links if the URLs don't resolve.

[mbollmann]

The C69 issue is a consequence of faulty XML, in my opinion. In the database import script, IDs of the form "x000" get interpreted as proceedings volumes, except for workshops, where it is IDs of the form "xx00". In C69.xml, volumes follow the workshop format, which leads to the first nine papers being ignored until a "proper" volume ID (1000) is found.

However, IMHO, the actual issue is that each paper has its own proceedings entry, which doesn't seem correct or useful to me. I believe the file should have a single proceedings entry with ID 1000, and the individual papers should be renumbered 0101 -> 1001, 0201 -> 1002, 0301 -> 1003, and so on.

[mjpost]

This seems right to me about C69. I just created an issue (#147) for it specifically, since it will require hand-correction.

[mjpost]

I believe these were all fixed with #324, in which I checked whether the frontmatter PDF existed and removed the <url> tag if not.