All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Antonin Godard" <antonin.godard@bootlin.com>
To: "Jose Quaresma" <jose.quaresma@oss.qualcomm.com>
Cc: <docs@lists.yoctoproject.org>
Subject: Re: [docs] [PATCH] ref-manual/system-requirements.rst: add ccache in distros
Date: Mon, 31 Aug 2026 14:26:32 +0200	[thread overview]
Message-ID: <DL34MCC6G8T0.12NMB0V503NE4@bootlin.com> (raw)
In-Reply-To: <CADWMC-bO57E5JZPfNXCp0rjOH-hGyWGyF8_bnBrvf8NrHRaDjg@mail.gmail.com>

On Mon Aug 31, 2026 at 12:48 PM CEST, Jose Quaresma wrote:
> On Mon, Aug 31, 2026 at 11:26 AM Antonin Godard <antonin.godard@bootlin.com>
> wrote:
>
>> Hi,
>>
>> On Mon Aug 31, 2026 at 11:58 AM CEST, Jose Quaresma via
>> lists.yoctoproject.org wrote:
>> > Looking at the Yocto docs, ccache is missing on Debian/Ubuntu
>> > and OpenSUSE but is in all the others.
>> >
>> > The ccache can improve significantly the builds speeds when we can't
>> > take advantage of the bitbake sstate cache, in such scenario we need
>> > o compile the code. The cache storage needs to be available from the
>> > previous builds so we need to store these ccache artifacts temporary
>> > on our permanent storage.
>> >
>> > The ccache is required to build "-native" packages in oe-core
>> > because it requires HOSTTOOLS += "ccache".
>>
>> Thanks for the patch. Actually, it makes me wonder if we should rather do
>> the
>> opposite and remove this for all distros, as this doesn't sound like a host
>> "requirement" to me.
>>
>
> Yeah, This is opcional and not a requirement as you said.
>
> However, this can greatly improve build time. So it fits in the same
> category as the
> python-websockets, which are not a requirement because upstream hashequiv
> server is not enabled by default, but it also appears listed.

websockets became a requirement when we made the "poky-with-sstate"
bitbake-setup configuration proposed as part of the choices when running
"bitbake-setup init". It would have made no sense to suggest a configuration
with a missing package otherwise.

> To take full advantage of the ccache.bbclass, you need to have the tool in
> the host;
> otherwise, you can only use it for the ":target" arch.

I think this should then be stated in the ccache class documentation, in
documentation/ref-manual/classes.rst. The host requirements should be kept
minimal, IMO.

Antonin


  reply	other threads:[~2026-08-31 12:26 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-31  9:58 [PATCH] ref-manual/system-requirements.rst: add ccache in distros Jose Quaresma
2026-08-31 10:26 ` [docs] " Antonin Godard
2026-08-31 10:48   ` Jose Quaresma
2026-08-31 12:26     ` Antonin Godard [this message]
2026-08-31 13:06       ` Jose Quaresma
2026-09-01 12:20         ` Antonin Godard
2026-09-01 12:39           ` Richard Purdie
2026-09-01 13:33             ` Jose Quaresma

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=DL34MCC6G8T0.12NMB0V503NE4@bootlin.com \
    --to=antonin.godard@bootlin.com \
    --cc=docs@lists.yoctoproject.org \
    --cc=jose.quaresma@oss.qualcomm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.