Add :PRINT-CASE argument to quickload, set :UPCASE by default

[jessymilare]

Some systems don't load correctly when *PRINT-CASE* isn't :UPCASE.

[quicklisp]

This is really interesting! What systems are affected and how did you find out?

…

On Wed, Jun 12, 2019 at 11:15 AM Jéssica Milaré ***@***.***> wrote: Some systems don't load correctly when *PRINT-CASE* isn't :UPCASE. ------------------------------ You can view, comment on, or merge this pull request online at: #186 Commit Summary - Add :PRINT-CASE argument to quickload, set :UPCASE by default File Changes - *M* quicklisp/client.lisp <https://github.com/quicklisp/quicklisp-client/pull/186/files#diff-0> (8) Patch Links: - https://github.com/quicklisp/quicklisp-client/pull/186.patch - https://github.com/quicklisp/quicklisp-client/pull/186.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#186?email_source=notifications&email_token=AACPNLJFYEI55VQERIPI6KTP2EHJNA5CNFSM4HXKBSF2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4GZC5VFA>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACPNLMKW5KDSO5TDGZDV7LP2EHJNANCNFSM4HXKBSFQ> .

[jessymilare]

I found it out because I don't like a REPL that looks like it is screaming lol. So I set *print-case* to :downcase, which makes SLIME command <M-.> stop working after updating slime because it looks for symbols with downcased names. I had to reset the variable `print-case' and delete ".slime/" directory for it to work again.

[quicklisp]

Thanks for the context. I'm hesitant to add this because there are likely a bunch of other variables that would affect things in a similar way. The first one that comes to mind is *read-default-float-format*. The most consistent way to handle all of them is let-binding the special around quickload (or any other operation). Should they all be rolled into quickload's lambda list?

[quicklisp]

Also, style-wise, I'd prefer something like this: &key ... ((:print-case *print-case*) *print-case*) ...

[jessymilare]

I don't see any problem including some (maybe many) specials into quickload's lambda list. Some standard functions (e.g. write ) and some libraries (e.g. bordeaux-threads) do that. If you wrap the special around quickload, the user loses control.

I'll change the style.

[quicklisp]

If you wrap the special around quickload, the user loses control.

I don't understand - how is (let ((*print-case* :upcase)) (ql:quickload ...)) reducing user control?

[jessymilare]

If you wrap the special around quickload, the user loses control.

I don't understand - how is (let ((*print-case* :upcase)) (ql:quickload ...)) reducing user control?

If the user has to type that, that is a lot to type and doesn't integrate well with SLIME, LEM editor, etc. IMHO, the default way of calling the function should work as-is.

[quicklisp]

Ahh, thanks, I see the argument. Despite the clearly written title and clear code, I misinterpreted this as providing an argument to bind the special variable without changing it on the way in.

Now that I understand, I'm not comfortable changing *print-case* by default via quickload. Things that break with a non-default *print-case* should be fixed or let-binding should be used instead.