Bruno Le Floch

Provide alternative names for regex functions

1

Currently, all `l3regex` functions start with `\regex` and their first argument is a regex. This leads to a sometimes quite unnatural order of arguments. I propose providing additional commands that...

\iow_wrap:nnnN and non-printable bytes in pdfTeX

2

When pdfTeX prints a message to the terminal (and not any time earlier in the processing, such as in \detokenize) characters with character code from 0 to 31 are output...

Nesting message redirection

2

Andrew Stacey at TeX.sx wishes to [suppress `xparse` log messages locally](http://tex.stackexchange.com/q/118677/2707). The package can indeed be quite verbose at times, and packages built upon it may want to redefine some...

Add \file_if_exist_input:nnTF (note second 'n'!)

6

We've had many discussions on how to provide a functionality similar to LaTeX2e's `\InputIfFileExists`. The discussion inevitably hits the question of whether the true branch should be performed before or...

Loading files with \ExplSyntaxOff

7

See #419 on use of internals in outside code. `xunicode-addon.sty` and the `ctex` bundle use `\l__kernel_expl_bool` to correctly restore the `\ExplSyntaxOn` status after loading a file (cannot be done in...

Document TF and non-TF versions of functions together

3

@wspr points out in #527 that `\prop_pop:NnN` and `\prop_pop:NnNTF` are documented 2 pages apart from each other. Nowadays the TF versions are the “preferred” ones so it makes sense to...

Catcodes in code environment

12

In another issue, Joseph raises the question of whether to keep `&` special in a code environment. To recapitulate, here are the catcodes in code (I've restricted myself to the...

Unicode support in l3regex

6

This issue aims to summarize what needs to be done to make `l3regex` Unicode-aware. Longer discussions should be moved to separate issues. - [ ] Internal code change: add to...

Redundant checks in debug mode

1

In some cases we end up checking multiple times the same properties of variables. Concretely, I came across the following behaviour. ``` \RequirePackage[enable-debug]{expl3} \ExplSyntaxOn \cs_show:N \tl_put_right:Nn % starts with \__kernel_chk_var_local:N...

Internally to the "write" handling we use a badly-named command `\__kernel_iow_with:Nnn` which sets an integer `#1` to the value `#2`, then runs the code `#3`, then restores the original value...