2016-03-26 15:57 GMT+03:00 Fat-Zer fatzer2@gmail.com:
subj;
The complete build log is attached.
Note the build failure is triggered by "-fkeep-inline-functions".
On 03/26/2016 09:57 PM, Fat-Zer wrote:
subj;
The complete build log is attached.
Hi Alex, I just made a new build of tdelibs (commit 491c2e4, same as your log) in Debian Stretch and had no issue. Attached the log file for reference. At a very first sight, it looks like there is some kind of order difference in the building. "Linking CXX shared library libtdecore.so" is reported at 42% on your log and 14% on mine.
Hope this helps :-) Cheers Michele
2016-03-28 7:25 GMT+03:00 Michele Calgaro michele.calgaro@yahoo.it:
On 03/26/2016 09:57 PM, Fat-Zer wrote: Hi Alex, I just made a new build of tdelibs (commit 491c2e4, same as your log) in Debian Stretch and had no issue. Attached the log file for reference.
As I noted above the build failure is triggered by "-fkeep-inline-functions", so you won't see it unless you turn off your pretty release flags and add this. I generally use this configurations for debug, but mostly for small projects. As it reveals now most complicated sets of interlinked libraries with lots of inlines (like Qt4 and kdelibs) are end up with FTBFS in such cases, so you may consider this a false alarm... Particularly this FTBFS is caused by a bit sloppy inclusion of tdeui headers to tdelibs, but this will be a bit hard one to workaround, so nevermind...
At a very first sight, it looks like there is some kind of order difference in the building. "Linking CXX shared library libtdecore.so" is reported at 42% on your log and 14% on mine.
Sure it is, I've build with make -j5...
On 2016/03/28 02:50 PM, Fat-Zer wrote:
2016-03-28 7:25 GMT+03:00 Michele Calgaro michele.calgaro@yahoo.it:
On 03/26/2016 09:57 PM, Fat-Zer wrote: Hi Alex, I just made a new build of tdelibs (commit 491c2e4, same as your log) in Debian Stretch and had no issue. Attached the log file for reference.
As I noted above the build failure is triggered by "-fkeep-inline-functions", so you won't see it unless you turn off your pretty release flags and add this. I generally use this configurations for debug, but mostly for small projects. As it reveals now most complicated sets of interlinked libraries with lots of inlines (like Qt4 and kdelibs) are end up with FTBFS in such cases, so you may consider this a false alarm... Particularly this FTBFS is caused by a bit sloppy inclusion of tdeui headers to tdelibs, but this will be a bit hard one to workaround, so nevermind...
Ah sorry, I must have missed that. I will try again with "-fkeep-inline-functions" (when I have some free time for this) and see what happens. As long as we can reproduce it properly, let's try to fix it :-) Feel free to open a report in bugszilla ;-)
At a very first sight, it looks like there is some kind of order difference in the building. "Linking CXX shared library libtdecore.so" is reported at 42% on your log and 14% on mine.
Sure it is, I've build with make -j5...
Ok, that makes sense then.
Cheers Michele
As I noted above the build failure is triggered by "-fkeep-inline-functions", so you won't see it unless you turn off your pretty release flags and add this. I generally use this configurations for debug, but mostly for small projects. As it reveals now most complicated sets of interlinked libraries with lots of inlines (like Qt4 and kdelibs) are end up with FTBFS in such cases, so you may consider this a false alarm... Particularly this FTBFS is caused by a bit sloppy inclusion of tdeui headers to tdelibs, but this will be a bit hard one to workaround, so nevermind...
Ok, I can reproduce the exact same FTBFS in my system as well when I use "-fkeep-inline-functions". Question: based on your experience, do you think it is worth spending time to look and fix this kind of problems across all TDE code? I am thinking in terms of time required vs. benefit. Cheers Michele
2016-03-28 16:59 GMT+03:00 Michele Calgaro michele.calgaro@yahoo.it:
Ok, I can reproduce the exact same FTBFS in my system as well when I use "-fkeep-inline-functions". Question: based on your experience, do you think it is worth spending time to look and fix this kind of problems across all TDE code? I am thinking in terms of time required vs. benefit.
Short answer: No. A bit longer answer: There won't likely be a lot of such issues, most of them supposed to be in tedlibs, but those in tdelibs may be tricky to fix and also their fixing may require to break the API. The benefits are quite little: a bit more code clearance... So IMHO it's priority is much lower than e.g. fixing those tones of warnings which are present even on the default level...
Sorry I've reported the issue before I get into it.. So it came to a big noise with mostly no reason...
On 2016/03/29 02:53 AM, Fat-Zer wrote:
2016-03-28 16:59 GMT+03:00 Michele Calgaro michele.calgaro@yahoo.it:
Ok, I can reproduce the exact same FTBFS in my system as well when I use "-fkeep-inline-functions". Question: based on your experience, do you think it is worth spending time to look and fix this kind of problems across all TDE code? I am thinking in terms of time required vs. benefit.
Short answer: No. A bit longer answer: There won't likely be a lot of such issues, most of them supposed to be in tedlibs, but those in tdelibs may be tricky to fix and also their fixing may require to break the API. The benefits are quite little: a bit more code clearance... So IMHO it's priority is much lower than e.g. fixing those tones of warnings which are present even on the default level...
Sorry I've reported the issue before I get into it.. So it came to a big noise with mostly no reason...
Ok, thanks for the feedback and sorry for the late reply (I have been quite busy in the last few days). Cheers Michele
-
Fat-Zer -
Michele Calgaro