Hi Seb,
Sebastien Lelong wrote: > Hi Rob, > > I was about to ask for permission ;-) to update the device files one > more time. MPLAB 8.15 came out and it has some improvements. I generated > new device files and saw several changes. Most are however of minor > importance. Some of the more important changes apply mostly to the newer > (maybe even future) PICs which are probably not used yet. Since some > errors which I 'fixed' in the .dev files are back in 8.15 I need some > time to be sure everything is OK. > So I guess 'no more device files updates' when we release soon, and we > keep some improvements in the pipeline! > > > Maybe we can deal with this issue a better way. Here's what I have in > mind... > > If device files are updated, we should ideally perform tests again, just > to be sure( and I'll modify "jallib" script to also include device files > revision to log history). Now, as you say, changes may be minor. Maybe > there's even no changes at all ! As the hex file is the last form before > going to the PIC (jal, asm, etc,...), we could run md5sum on hex files, > and compare these sums between previous and current device files. This > will point to the files which result in changes, and hopefully show > those which did not. After all, changes might occur in part of device > files that we may not even use... Although the device files may validate and compile without errors, actually most of the declared registers and variables are not touched in our samples and test programs. So comparing hex files will not catch 'subtile' differences. But when people are going to use these for a 'real' programs they may encounter bugs. In the meantime I have run a DIFF program to search for differences between different versions of the device files. That showed clearly enough to me that there are relatively few improvements and essential fixes are mostly in the 'unpopular' PICs (the very new or exotic or even 'future product' types of which there is even not a datasheet published). On the other hand there are still errors in MPLAB 8.15 which I had corrected manually in 8.14. All by all I came to the conclusion it is not worth to take the risk to replace the device files now. But your idea seems good for regression tests. > I don't even know if everybody here has run tests against your last > devive files updates. I personally did not have time yet. Since we > currently have no log/track of our test results, we just cannot know > (but using jallib script will fix). So I'd say to update them... I withdraw my intended request for permission! > The real problem is the validation process is far too long, and block > the unvalidated map for too long time. We need to find a way to make it > faster. I don't understand. Validation of all device files together with generation, validation and compilation of all blink-an-LED programs takes a few minutes here. You'll understand I do this locally after a checkout: no blocking of the repository at all. You must do it significantly different.... how? Regards. -- Rob Hamerling, Vianen, NL (http://www.robh.nl/) --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jallib" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/jallib?hl=en -~----------~----~----~----~------~----~------~--~---
