Stream: Beginner Questions

Topic: Github


view this post on Zulip Huub Vromen (Jul 24 2020 at 12:17):

Isabelle theory files in Github sometimes use notation like "\<lfloor>\<P> \<N>\<E>\<rfloor>" . Is it possible to convert this into "normal" code, without having to do it manually?

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:19):

Isabelle/jEdit should replace it when opening the file. The generated documentation will also have symbols.

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:21):

Isabelle/jEdit is internally mapping the symbols like \<P> to the utf8 output, whereas github shows the raw file.

view this post on Zulip Josh Chen (Jul 24 2020 at 12:22):

Yeah, the whole Isabelle toolchain uses these custom encoding things, so unfortunately it's a thing you have to get used to if your idea of "normal" code is "UTF-8 encoded".

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:26):

As a side remark, do not attempt to replace the commands by the associated utf8 symbol. I did it once and it leads to very strange errors (isabelle build did not work, while isabelle jedit worked).

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:26):

For my own theories, I have a jenkins instance running to produce the documentation to be able to look at it, if I want to.

view this post on Zulip Josh Chen (Jul 24 2020 at 12:30):

Mathias Fleury said:

As a side remark, do not attempt to replace the commands by the associated utf8 symbol. I did it once and it leads to very strange errors (isabelle build did not work, while isabelle jedit worked).

I think there was a discussion on these things in an early thread on here. Do you know if it's a parser-related thing?

view this post on Zulip Huub Vromen (Jul 24 2020 at 12:31):

Thanks for your answers. I hoped there would be a way to get "UTF-8 codes". I do not encounter the problem with theories in the AFP.

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:32):

Josh Chen said:

I think there was a discussion on these things in an early thread on here. Do you know if it's a parser-related thing?

I did not try to debug it. I thought that Isabelle/jEdit maps all UTF8 symbols back to their representation, but I have no idea…

view this post on Zulip Mathias Fleury (Jul 24 2020 at 12:33):

Huub Vromen said:

Thanks for your answers. I hoped there would be a way to get "UTF-8 codes". I do not encounter the problem with theories in the AFP.

Chapter 3 of https://isabelle.in.tum.de/dist/Isabelle2020/doc/system.pdf to produce HTML code…

view this post on Zulip Huub Vromen (Jul 24 2020 at 12:56):

I have found the answer: when I reload the theory Jedit replaces the backslash-codes with the usual symbols.

view this post on Zulip Manuel Eberl (Jul 25 2020 at 08:19):

Sometimes, they stay there for no apparent reason, and you have to do "File → Reload with Encoding → UTF8-Isabelle"

view this post on Zulip Manuel Eberl (Jul 25 2020 at 08:19):

I think this sometimes happens when you load Isabelle theories that have actual UTF8 in them (as opposed to ASCII-encoded Isabelle symbols).


Last updated: Apr 23 2024 at 20:15 UTC