[vlc-devel] [PATCH 0/3] Expand contrib

[Steve Lhomme]

On Mon, Jun 12, 2017 at 11:37 AM, Rémi Denis-Courmont <remi at remlab.net> wrote:
> On June 12, 2017 11:18:49 AM GMT+03:00, Steve Lhomme <robux4 at videolabs.io> wrote:
>>As I often build the contribs and update the MSVC patches, it's always
>>a PITA
>>to find where it failed to merge old patches. I realised if there was a
>>way to
>>prepare all the code (download the tarballs and extract them) that
>>would save
>>me quite a bit of time during that phase. So I added "make expand" to
>>the
>>contribs.
>>
>>It could also be used by the build bots instead of make fetch. That
>>would also
>>make the failures of compilation easier to find.
>>
>>Steve Lhomme (3):
>>  contrib: add a make expand target
>>  contrib: more consistent unpack target names
>>  contrib: fix the expand rule for headers-only contribs
>>
>> contrib/src/a52/rules.mak       |  4 ++--
>> contrib/src/ass/rules.mak       |  4 ++--
>> contrib/src/crystalhd/rules.mak |  1 +
>> contrib/src/d3d11/rules.mak     |  1 +
>> contrib/src/d3d9/rules.mak      |  1 +
>> contrib/src/dca/rules.mak       |  4 ++--
>> contrib/src/dvbpsi/rules.mak    |  4 ++--
>> contrib/src/ebml/rules.mak      |  4 ++--
>> contrib/src/freetype2/rules.mak |  4 ++--
>> contrib/src/gcrypt/rules.mak    |  4 ++--
>> contrib/src/gme/rules.mak       |  4 ++--
>> contrib/src/gpg-error/rules.mak |  4 ++--
>> contrib/src/help.txt            |  1 +
>> contrib/src/kate/rules.mak      |  4 ++--
>> contrib/src/mad/rules.mak       |  4 ++--
>> contrib/src/main.mak            | 10 +++++++++-
>> contrib/src/matroska/rules.mak  |  4 ++--
>> contrib/src/modplug/rules.mak   |  4 ++--
>> contrib/src/mpcdec/rules.mak    |  8 ++++----
>> 19 files changed, 43 insertions(+), 31 deletions(-)
>>
>>--
>>2.12.1
>>
>>_______________________________________________
>>vlc-devel mailing list
>>To unsubscribe or modify your subscription options:
>>https://mailman.videolan.org/listinfo/vlc-devel
>
> To find why a build fails, I do:
> # make -k || make -j1
>
> Then the failure is right before the end of the execution.

I do that too.

> I don't understand how adding an explicit decompression target helps here.

Here's an example of a build on Jenkins using make -j4. Everything is
mixed together. There's no failure here. But it decompressing is mixed
with compiling there's a higher chance to not get the actual context
of what's being compiled easily.
https://jenkins.videolan.org/job/vlc-nightly/job/vlc-nightly-win64-x86_64/412/consoleFull

> --
> Rémi Denis-Courmont
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel