b2 headers is unreliable on 1.56 on Windows
Anyone else found that b2 headers in 1.56 randomly misses things to hard link on Windows? If you run it four or five times it eventually hard links everything through monte-carlo, but it is a bit annoying. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/
Hi Niall, Am Mittwoch, 30. Juli 2014, 16:49:37 schrieb Niall Douglas:
Anyone else found that b2 headers in 1.56 randomly misses things to hard link on Windows?
Already reported by Eric in http://lists.boost.org/Archives/boost/2014/07/215266.php
If you run it four or five times it eventually hard links everything through monte-carlo, but it is a bit annoying.
See my reply in http://lists.boost.org/Archives/boost/2014/07/215276.php for more details. Short:
It seems that the hardlinks for the directories created in the first run are replaced with new directories containing file hardlinks in the second run. This sometimes fails horrible.
The first run creates all files. The next runs create a big pile... I send this to the Boost.Build mailing list too. As 1.56.0 should come with pregenerated headers, this should not be a showstopper. Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! * voice: ++49 4257 300 ! Fährstraße 1 * fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen * jhunold@gmx.eu ! Germany
On 30 Jul 2014 at 18:09, Jürgen Hunold wrote:
Anyone else found that b2 headers in 1.56 randomly misses things to hard link on Windows?
Already reported by Eric in
Oh good.
I send this to the Boost.Build mailing list too. As 1.56.0 should come with pregenerated headers, this should not be a showstopper.
It is for anyone with a library in the review queue :( In our documentation we're going to have to ask people installing our library as an additional Boost module to repeatedly do b2 headers until it doesn't error out. It also creates problems for those running CIs for per-commit testing, because the CI needs to bootstrap the test code into Boost and do b2 headers to update. I've already seen problems with my CI testing Boost.Thread on Windows due to this, it currently randomly fails due to wrong or missing headers. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/
participants (2)
-
Jürgen Hunold -
Niall Douglas