EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  <20192020  2021  2022  2023  2024  Index 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  <20192020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: a question on a commit comment
From: Jeong Han Lee via Core-talk <[email protected]>
To: "Johnson, Andrew N." <[email protected]>, "[email protected]" <[email protected]>
Date: Thu, 11 Apr 2019 01:02:00 +0200
Hi Andrew,

Thank you for more explanation about this commit. Due to your explanation, I can safely ignore this change, so I create a local change in our base. Then everything works well again.



On 4/3/19 6:37 PM, Johnson, Andrew N. via Core-talk wrote:
Hi Han,

On 4/3/19 4:15 AM, Jeong Han Lee via Core-talk wrote:
  I am interesting to hear the next sentence after your commit message with the following commit

https://github.com/epics-base/epics-base/commit/2206934ae2c64b5123874117d220480e8b2b1b0a

  "Might be unnecessary, but..."
"... I think it is worth the risk of making these changes to remove another set of warnings from the build."

  We (ESS) has our own ( ^__^ )  issue on this commit. I would like to understand your commit precisely, and do "some patch works" on the latest customized ESS environment.
Interesting, what are your symptoms?

It is not issue on our customized epics base installation, but several issues on "our own modules with our own building system which use partly EPICS base configuration files". I will investigate more why this command makes us and PSI trouble later.



The modification was made to ensure that the $(INSTALL_LIB) directory exists before we link any programs — $(TESTPRODNAME)$(PRODNAME) — or libraries — $(LIBNAME) $(SHRLIBNAME) $(LOADABLE_SHRLIBNAME) — because on macOS the linker flag -L generates a warning if you give it a directory path that doesn't exist. The use of the pipe character '|' in the additional dependency makes that directory target an order-only prerequisite, so it doesn't cause the downstream targets to be rebuilt if the directory's modification date is newer than the target, it just ensures that it exists before running the rule action that does the linking.

The rule change did not happen on the 3.15 branch; when I build the 3.15 libCom tree from scratch on a Mac I get these warnings:
kato$ make -s
ld: warning: directory not found for option '-L/Users/anj/Software/epics/base-3.15/lib/darwin-x86' ld: warning: directory not found for option '-L/Users/anj/Software/epics/base-3.15/lib/darwin-x86' ld: warning: directory not found for option '-L/Users/anj/Software/epics/base-3.15/lib/darwin-x86'

They are not displayed when I build the 7.0 branch.

I would like to help you resolve this — I wish sites would work with the core group if they need tweaks to the build system instead of making their own local changes, since that helps us understand your issues and makes us less likely to break your installations in the future. We intend to release Base-7.0 more frequently in the future (Michael wants releases quarterly) so there should be less of a delay between commits and their inclusion in a subsequent release.

Since we don't use the 100% EPICS standard building system, I think, we have to take over this pain ourselves. I had few ideas how I combine our own working model with the standard EPICS building system and did some evaluations before. But I don't have time anymore due to MTCA systems, which we have to deploy massively. :)

  And please understand if I can ask more stupid questions future.

  Thank you Andrew!

  Han





- Andrew

--
Arguing for surveillance because you have nothing to hide is no
different than making the claim, "I don't care about freedom of
speech because I have nothing to say." -- Edward Snowdon


References:
a question on a commit comment Jeong Han Lee via Core-talk
Re: a question on a commit comment Johnson, Andrew N. via Core-talk

Navigate by Date:
Prev: Jenkins build is still unstable: epics-7.0 » linux64 #111 APS Jenkins via Core-talk
Next: Jenkins build is still unstable: epics-7.0 » mac #111 APS Jenkins via Core-talk
Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  <20192020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: a question on a commit comment Johnson, Andrew N. via Core-talk
Next: Excessive scan times for periodic scans Konrad, Martin via Core-talk
Index: 2002  2003  2004  2005  2006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  <20192020  2021  2022  2023  2024 
ANJ, 10 Apr 2019 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·