This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH]: Updates to Ada sources, part 1 (longish)


Date: Fri, 04 Jun 2004 14:31:57 -0400
From: Andrew Cagney <cagney@gnu.org>

> > That'd be fine with me, but I'd also suggest to have a pointer to
> ChangeLog.GNAT in gdb/ChangeLog, right where the import of Ada changes
> is recorded. Something like "See ChangeLog.GNAT for details of the
> changes."


Please don't do this. ChangeLog.XXX is good for branches (as with even readline/ChangeLog.gdb) but not for what is ment to be the mainline.


I agree with the principle, but I don't think it is such a sacred one
that would justify asking the GNAT people to invest such a large
effort.

We need the information to be there, and the suggested compromises
achieve that with a reasonable effort.

We've on a number of occasions seen large through to extreemly large merges (HP comes to mind) and on each occason the contributor, unprompted, ensure that a correct ChangeLog was included.


The ChangeLog entry should provide a summary of what was added/changed
at this point in time - stuff like listing the new functions and summary of changed functions. Can we do that?


How many man-hours would you say is reasonable for such an effort?
1? 10? 100? 1000?  Where, if at all, do we draw the line?

We're talking an hour or perhaphs two, and something we're _all_ expected to do willingly.


The way to do this is to ignore the history and just look at the final diff. A bit of sed'n'sort will in a matter of minutes give the functions added/deleted bit, leaving just terse verbage of the other functions changed. Can we do that?

Much of the stuff in that GNAT ChangeLog will no longer be relevant.


I understand that the problem is to translate the irrelevant suff into
something that is relevant.  I suggested that at first, but the GNAT
people tell that it will take a lot of work.


I'm told Diego took rougly a day to prepare his tree-SSA ChangeLog entry.


Who is Diego and how is the tree-SSA stuff relevant to our case?

Diego did a very very large (read 4 years of work x many people) merge of GCC's mainline. When committed it included a very large (but terse) ChangeLog entry.


Andrew



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]