Re: [ccache] ccache error: Failed to create temporary file... Read-only file system

2021-10-24 Thread Bernd Petrovitsch via ccache

Hi all!

On 24/10/2021 15:12, Raju C K via ccache wrote:
[...]

The system worked fine earlier, in the same environment, where .ccache 
directory was read-write mounted system. No changes have been made to the 
configuration.


Strange - I don't think - I fail to see any somewhat valid reason that - 
it makes sense to try to get ccache to work with a read-only cache 
directory 


Even if one disables all statistic stuff (in case that is possible - I 
never tried that) 



Error came up as a complete surprise, and I just wanted to verify whether any 
changes were made to ccache-4 that would throw up that warning.
I will keep investigating and will report back here, if I overcome this hurdle.


Kind regards,
Bernd
--
Bernd Petrovitsch  Email : be...@petrovitsch.priv.at
 LUGA : http://www.luga.at

___
ccache mailing list
ccache@lists.samba.org
https://lists.samba.org/mailman/listinfo/ccache


Re: [ccache] ccache error: Failed to create temporary file... Read-only file system

2021-10-24 Thread Bernd Petrovitsch via ccache

Hi all!

On 22/10/2021 18:48, Raju C K via ccache wrote:
[...]

I am trying to build bootimage from kernel sources for android. I have ccache 
enabled with max_size 90G.
The ccache version is 4.2.1

Compilation stops, throwing a "ccache error: Failed to create temporary file 
/home/ck/.ccache/tmp/tmp.a.Xdafd Read only file system".


ccache *caches* something and writes the cache to some filesystem.
Obviously, that should be - thus - writable.
And obviously, the home directory of that use (or at least the
/home/ck/.ccache directory) is on a read-only mounted filesystem.

You need to tell more about your compiling environment (I never built 
something for Android anywhere ...).



Can anyone help or show some resources to read/refer ?


Any basic Unix book/documentation should mention "rw vs ro" mounts of 
filesystems (of any type) ...


Kind regards,
Bernd
--
Bernd Petrovitsch  Email : be...@petrovitsch.priv.at
 LUGA : http://www.luga.at

___
ccache mailing list
ccache@lists.samba.org
https://lists.samba.org/mailman/listinfo/ccache


Re: [ccache] gmake and ccache conspiring together in creating gremlins

2021-02-08 Thread Bernd Petrovitsch via ccache
Hi all!

On 07/02/2021 16:45, Sam Varshavchik via ccache wrote:
[...]
> In the interim I discovered a workaround that seems to work, for
> anyone who's also affected by this. Instead of setting CCACHE_PREFIX
> to distcc, I pointed CCACHE_PREFIX to this:
> 
> #! /bin/sh
> 
> unset MFLAGS ${!MAKE@}
> exec distcc "$@"
> 
> This clears make variables from the environment and when the linker
> invokes make for ...whatever it needs to invoke it for, against a
> manufactured makefile it generates on the fly... it doesn't go look
> for the job server file descriptors, and go sideways.

Hmm, with the whole thread: Shouldn't better the linker (or whatever
tool) cleanup the environment variables before invoking `make` (or
whatever programs) to fix the - IMHO - core cause of the bug?

MfG,
Bernd
-- 
Bernd Petrovitsch  Email : be...@petrovitsch.priv.at
 LUGA : http://www.luga.at

___
ccache mailing list
ccache@lists.samba.org
https://lists.samba.org/mailman/listinfo/ccache


Re: [ccache] ccache on generated C++ files...

2019-06-25 Thread Bernd Petrovitsch via ccache
On 25/06/2019 00:28, Malfettone, Kris via ccache wrote:
> So what I seem to be observing is only when I do a ninja -t clean or make 
> clean.  This removes all of my generated files (.h and .cpp files) and of 
> course the normal build outputs.  A subsequent run is then mostly cache 
> misses.   If I say manually do a rm $(find -name "*.o") then I get a 100% 
> cache hit rate.  I had assumed it was due to the fact that the generated 
> files would have a newer mtime and ctime.  My assumption must be wrong.

TTBOMK ccache ignores mtime/ctime/... (as make/ninja/... use that anyway
before) but only the preprocessed content matters.

IIUC, the `make clean` removes all *.o and all generated .h/.c/.hpp/.cpp
files.
The rm $(find -name "*.o") removes only *.o.

So the analysis above is at least incomplete.

Are you really sure the newly generated .h/.c/.hpp/.cpp files are
identical (enough)?
No static variables or expanded preprocessor #defines with the build
date+time or similar in there?

[ Fullquote deleted ]
MfG,
Bernd
-- 
Bernd Petrovitsch  Email : be...@petrovitsch.priv.at
 LUGA : http://www.luga.at

___
ccache mailing list
ccache@lists.samba.org
https://lists.samba.org/mailman/listinfo/ccache