Re: [PATCH] scripts/ci: enable use of cached base repository for fast build

[Jan Kiszka <jan.kiszka@siemens.com>]

On 29.11.18 13:33, [ext] Henning Schild wrote:
> Am Thu, 29 Nov 2018 15:02:06 +0300
> schrieb "Maxim Yu. Osipov" <mosipov@ilbers.de>:
> 
>> On 11/29/18 2:33 PM, Henning Schild wrote:
>>> Am Thu, 29 Nov 2018 08:19:18 +0100
>>> schrieb "Maxim Yu. Osipov" <mosipov@ilbers.de>:
>>>    
>>>> Signed-off-by: Maxim Yu. Osipov <mosipovd@ilbers.de>
>>>> ---
>>>>    scripts/ci_build.sh | 7 +++++++
>>>>    1 file changed, 7 insertions(+)
>>>>
>>>> diff --git a/scripts/ci_build.sh b/scripts/ci_build.sh
>>>> index 059df21..20b77b7 100755
>>>> --- a/scripts/ci_build.sh
>>>> +++ b/scripts/ci_build.sh
>>>> @@ -86,7 +86,14 @@ fi
>>>>    if [ -n "$FAST_BUILD" ]; then
>>>>        # Start build for the reduced set of configurations
>>>>        # Enforce cross-compilation to speed up the build
>>>> +    # Enable use of cached base repository
>>>>        sed -i -e 's/ISAR_CROSS_COMPILE ?= "0"/ISAR_CROSS_COMPILE ?=
>>>> "1"/g' conf/local.conf
>>>> +    bitbake $BB_ARGS -c cache_base_repo \
>>>> +        multiconfig:qemuarm-stretch:isar-image-base \
>>>> +        multiconfig:qemuarm64-stretch:isar-image-base \
>>>> +        multiconfig:qemuamd64-stretch:isar-image-base
>>>> +    sudo rm -rf tmp
>>>
>>> That rm is not in line with the user manual, should that be updated
>>> as well?
>>
>> Yes, after commit de68185 "conf: Move base-apt to downloads directory"
>> one may remove the whole tmp directory. I'll update the user manual.
>>
>>>> +    sed -i -e 's/#ISAR_USE_CACHED_BASE_REPO ?=
>>>> "1"/ISAR_USE_CACHED_BASE_REPO ?= "1"/g' conf/local.conf bitbake
>>>
>>> I will not propose "ifconfig down" but this here to cover cutting
>>> off the network for the debian side.
>>
>>
>>> # set invalid proxy variables to enforce "offline"
>>> export http_proxy=http://localhost:3128
>>> export https_proxy=$http_proxy
>>> export ftp_proxy=$http_proxy
>>
>> Have you tested this in your environment?
> 
> No, this is just the long version of "your patch does not actually go
> offline and should not be merged".
> 

FWIW, if we decide to test offline build via invalidating proxies, we need to 
have a separate cache test for the same images in place as well. Is that the 
case already?

Jan

-- 
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux