xiaoxiang781216 commented on pull request #4193: URL: https://github.com/apache/incubator-nuttx/pull/4193#issuecomment-885033582
> > > > > It should be done and fully tested on real HW before it comes in to master. > > > > > Event the printf change will increase code size with FS_LARGEFILE off. > > > > > > > > > > Why? I don't see any reason that the code size if FS_LARGEFILE off. Could you explain more? > > > > > > > > > 1953165 - 1953157 = 8 > > > This is about code space and useless bloat. If it is Off it need to cost 0 > > > > > > The difference you identify is due to the cast in syslog, I can add #if/#else/#endif in the code, but all these places are the debugging log, do you think it's worth to make the code like spaghetti to save several bytes(especially if we consider that the debuging log already consume much more memory than the production code). > > > > the effect can be seen in this PR it cost 8 bytes (yes it is a debug path, but it still need to fit when needed) > > Of course is it the cast. I and already said it was the debug path. > > No we do not want any more if def rash. (Maybe some PRI and type magic?) The is the point is. What it the total cost of this change in resources when it is OFF? > > Will debug builds still fit on constrained SoC that do not need 64Bit offsets for their files systems? > Ok, done, please try it. > The changes from 10.0-10.1+ has already added 3.7K if the offset is referenced 25 times in debug code that that is another 1/4 K. > As @acassis suggest, it's better to integrate some basic size check to CI. > All I am asking you do is not kill the small of Nuttx. Please see what you can do here. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
