Unify thesis under a same document class

[jpbarraca]

Unify the document to use a single document class, instead of two (cover and matter).
This may be necessary to comply with current template and to allow better handling of PDF metadata.

Currently tracked in the unified-memoir branch

[danielvcorreia]

This unified approach does not seem to work on Overleaf... Is the implementation with the temporary file too CPU expensive?

[jpbarraca]

Currently we are creating two different documents, using two different document classes. This happened due to historical reasons, and has no effective support.
The cover was using the report class, while the main thesis used memoir.

The current template provided by the University suggests the addition of the index and references following the same format used in the cover. At the same time, it is desirable to have metadata directly in the PDF. Both point towards having all information easily available. Currently, the abstract is in the cover, while the lists of chapters, sections, figures, etc... is in the main document.

This work aims to unify everything under the same document. With the feature, we could easily output PDF metadata without side tasks, and effectively comply with the University template.

[danielvcorreia]

Are intermediate files still needed?

When using the work from the branch referenced in #56 we do not need intermediate files, as all the information is available. Can you check it?

• I confirm that this unification does not require intermediate files. I opened a new PR Configure pdf information dictionary #57 that uses this branch.
• The previous PR Added dynamic configuration of the .pdf metadata with hyperref #42 can be closed?!

Tokens not allowed in a PDF string

Some tokens in the abstract text are not allowed and generate errors. In the PR I'm using a static string for now....
These are some possible solutions:

1. Remove these tokens that are not supported
2. Use the \pdfstringdefDisableCommands command to redefine commands that are usually not supported
3. Don't use the abstract text

There are still some compilation errors in this branch

• \IndexContents

Class memoir Error: \footskip (28.45274pt) is too large for \lowermargin (-10.40627pt) by 38.85901pt.

• The usage of minted package in file chapter1.tex

Package minted Error: Missing Pygments output;

[jpbarraca]

#57 is a better approach. Lets close #42

Regarding the Abstract: Which tokens are not supported?

Regarding the compilation:
It's a devel branch with compilation and format errors. This was one of the reasons why memoir was not used for the cover. The definition of the margins and text area is different from the report class. So consider it to be WIP.
If you have the opportunity, PRs are welcome.

[danielvcorreia]

I think all commands (\anything) are not supported as a PDF string.
For example: \selectlanguage{}, \begin{}, \\

[jpbarraca]

We can leave a note warning students to not use commands on the abstract.

[danielvcorreia]

I didn't notice a comment above the abstract that says: %%% Don't remove the \selectlanguage statement or you will end with incorrect hyphenation

Because of this, we need to use the \pdfstringdefDisableCommands command to possibilitate certain commands to be in the Abstract... I added the support for \selectlanguage, \texttt, and \\ commands for now. I also left a note not to use additional commands.

Another aspect is that the Abstract is a list of Abstracts in different languages, so the last added element is the one used in the PDF information dictionary. Should the Portuguese version always be used?