search

mpaland / printf

Tiny, fast, non-dependent and fully loaded printf implementation for embedded systems. Extensive test suite passing.
MIT License
2.46k stars 453 forks source link

Use more appropriate types for base, precision and width #116

Open eyalroz opened 2 years ago

eyalroz commented 2 years ago

The only base values the library support are: 2, 8, 10, 16. And yet - we pass it around in larger types, and not even the same type everywhere.

So, why should we not just use a single small type for the base? Say, uint_least8_t or uint_fast8_t?

eyalroz commented 2 years ago

@mickjc750 : This issue report is inspired by your commit b907595b09bc7b03ed0803f731ec5aeffa80b91d , where you've done something similar. Which of the types do you believe is more appropriate?

mickjc750 commented 2 years ago

For base, just a simple uint8_t, as there's no arithmetic performed on base which would require masking it down to 8bits on a 32bit cpu. For precision and width, uint_fast8_t would provide a slightly smaller (and faster) executable than uint8_t , as there are places where precision is decremented.

In the ntoa(), I'd recommend using literal values %2 %8 %10 %16 rather than %base, as it saves performing a division. It's a minor size increase for a big performance increase.

On Sun, Aug 1, 2021 at 6:03 AM Eyal Rozenberg @.***> wrote:

@mickjc750 https://github.com/mickjc750 : This issue report is inspired by your commit b907595 https://github.com/mpaland/printf/commit/b907595b09bc7b03ed0803f731ec5aeffa80b91d , where you've done something similar. Which of the types do you believe is more appropriate?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/mpaland/printf/issues/116#issuecomment-890398279, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABJ2OLLVCCELVM7VD5DVY2TT2RJIBANCNFSM5BKKNNDA .

--

Michael Clift - Design Engineer.

Clift Innovations Pty Ltd Suit 10/13 Walkers Road Nunawading, VIC 3131

ABN 24 286188328

Tel: +61 3 9877 3109 Fax: +61 3 9999 1406 Mob: 0417 613 680 Email: @.*** Web: www.clift.com.au

PRIVATE AND CONFIDENTIAL This e-mail and any files transmitted with it are intended solely for the use of the addressee(s) and may contain information that is confidential or privileged. If you receive this e-mail and you are not the addressee(s), please disregard the contents of the e-mail, delete the e-mail and notify the author immediately.

eyalroz commented 2 years ago

@mickjc750 : You've packaged several issues into one... anyway, are we sure width always fits in a uint8_t ?

Also, please try to avoid commenting by replying to the email and quoting everything. If you reply, please quote nothing. Otherwise we get a huge quoted-part of your comment which doesn't add any new information.

mickjc750 commented 2 years ago

Not 100% sure. It really comes down to opinion as to what a reasonable width limit is. Printf is generating text, 255 sounds like a reasonable limit for the width of a number. Or the width of a field with a %s aligned in it.