Hi, I didn't understand why in (f) jrhacker should refine the series, and not just the commit message...
2014-12-08 21:47 GMT-02:00 Thien-Thi Nguyen <[email protected]>: > After many changes in LibreDWG: > > 5dca683 2014-12-05 [v] Fix bug: Collapse individual tuple-item > assignments. > 29abb49 2014-12-04 [v] Make testsuite/testcases progs link against > in-tree libredwg.la. > 387d011 2014-12-03 [v] Explicitly name testsuite/testcases support files. > 88b7f13 2014-11-29 [v] Distribute testsuite/testcases support files. > 812ea8e 2014-11-27 [v] Declare helper.py encoding. > 6f18b96 2014-11-26 [maint] Remove *.pyc files from repo. > cad9247 2014-11-26 [v] Streamline invocation of sub-program txttoxml.py. > 03ac140 2014-11-23 [v] Use Python interpreter discovered by configure > script. > 78a0b97 2014-11-21 [v] Make xmlsuite data files non-executable. > fd111c9 2014-11-21 [v] Distribute xmlsuite support files. > 356315d 2014-11-19 [v] Use ‘LDADD’ and full libredwg.la filename. > 174ceaf 2014-11-16 [v] Don't forget to distribute common.c. > 0cdbb4e 2014-11-16 [v] Don't forget to distribute suffix.c. > 38cea3f 2014-10-22 [build] Determine libxml2 flags at configure time. > ddba219 2014-10-22 [build] Make configure script check for pkg-config(1). > 79b2a82 2014-10-22 [admin] Update ignorance; nfc. > > and in the LibreDWG configuration for Hydra: > > 4c7787c 2014-11-25 libredwg: Add libxml2Python as a dependency. > 3171825 2014-11-12 libredwg: Don't try to build on darwin. > e558e01 2014-11-09 libredwg: add dejagnu as dependency > 0146a76 2014-10-28 libredwg: "fix-up" factor dependencies list (again) > e92e4b7 2014-10-27 libredwg: "fix-up" factor dependencies list (again) > dd9d096 2014-10-27 libredwg: "fix-up" factor dependencies list > 8f7ffb7 2014-10-21 libredwg: factor dependencies list > 36519a3 2014-10-16 libredwg: add dejagnu as dependency > 75b5503 2014-10-13 libredwg: delete EOL ws; set maintainer email address > 157659e 2014-10-13 libredwg: add libxml2 as dependency > > i am happy to report that LibreDWG succesfully builds the tarball on > Hydra, and makes a good show of building the "make check" prerequisites > for several platforms on Hydra. Now, a good show is not success and > even less so "make check" success, so it is not yet time to celebrate. > > This message is intended, instead, to invite the rest of the LibreDWG > hackers to fix up the warnings and errors (and segfaults!) unearthed by > these builds. To start, have a look at: > > http://hydra.nixos.org/jobset/gnu/libredwg-master > > Then, follow the links (eventually) to a "raw" link, such as: > > http://hydra.nixos.org/build/17704711/nixlog/1/raw > > Lastly, search the contents of that page for "error" or "warning" or > "segmentation fault", find something you can fix, and fix it, locally. > > But why locally? Why not just push the fix? > > Well, i am going to be blunt here: I'm not happy w/ the quality of some > of the more recent commit messages. They do not conform to the GNU > coding standards. IMHO, it's too much of a pain to change them now, > but we should make a strong effort to improve ourselves going forward. > > So, i propose to serve as a mentor for this very narrow aspect of > LibreDWG hacking (commit message construction), and explicitly not for > any other aspect (design, implementation). The work flow is like this: > > (a) jrhacker finds PROBLEM (see above) > (b) jrhacker signals mailing list: "i'm on it; PROBLEM is mine!" > (c) jrhacker fixes PROBLEM and commits to local repo > (d) jrhacker posts patch series (see git-format-patch(1)) > (e) ttn critiques the commit message(s) > (f) jrhacker rolls eyes, refines the series, reposts > (g) goto (e) unless ttn finally shuts up > (h) jrhacker pushes to the savannah repo > > BTW, if there are design / implementation critique from non-ttn > hackers as part of (e), then that's fine, too. This way, we get > (technically) the fixes and proper commit messages, and (socially) > a chance to get to know each other better. All in the name of GNU, > of course! > > What do people think? > > -- > Thien-Thi Nguyen > GPG key: 4C807502 > (if you're human and you know it) > read my lisp: (responsep (questions 'technical) > (not (via 'mailing-list))) > => nil >
