[inspection] License+copyrights in RC branch (was: [mpl] Metafunction terminology - revival?)
Gennaro Prota wrote:
On Mon, 24 Jul 2006 13:06:03 -0500, Rene Rivera
While your in there changing MPL docs, could you do us the favor of adding some of the missing copyright+license information. MPL is *1* in the number of those issues.
Rene, next time we should remember to run the inspect tool (and fix its discoveries :-)) before branching for release, so that we don't have to duplicate all corrections on head and branch.
Historically it hasn't made a difference when we run the inspect tool. People mostly ignored it and hence why we have an accumulation problems now. Now perhaps that it's in the form of a constant nagging email instead of a web page people will pay attention (cross-fingers). As for the amount of work I don't think libraries have diverged that much between HEAD and RC so it's likely not that much work since one only has to merge to the RC branch.
Or maybe the inspect tool should run once a week or so, in general.
Not sure about the schedule, but definitely it will running all the time from now on. I haven't enabled runs of it in HEAD yet as I expect that would generate a somewhat larger set of issues. But perhaps we could start such runs on HEAD on a weekly schedule? (Did a test run, and the *LC* email on HEAD would exceed the 75K limit. Only by a hair though.)
PS: maybe we want to highlight this in a separate thread?
Yep :-) -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo
On Mon, 24 Jul 2006 18:28:41 -0500, Rene Rivera <grafikrobot@gmail.com> wrote:
Historically it hasn't made a difference when we run the inspect tool. People mostly ignored it and hence why we have an accumulation problems now. Now perhaps that it's in the form of a constant nagging email instead of a web page people will pay attention (cross-fingers).
Yeah :-) If that doesn't solve the issue either, then we could integrate the check into regression tests: global failure if not inspect-free (and transcription to criminal record! ;-))
As for the amount of work I don't think libraries have diverged that much between HEAD and RC so it's likely not that much work since one only has to merge to the RC branch.
But it's still very tedious to switch to the branch, check the diffs etc. Or do you merge blindly? I'm actually tempted to do the work for the 1.34 branch only, and postpone it on the head. The point is, as you well know, that anything being tedious is error-prone. I'm doing the job in little pieces, every time in the day I *feel* like doing it :-)
Or maybe the inspect tool should run once a week or so, in general.
Not sure about the schedule, but definitely it will running all the time from now on. I haven't enabled runs of it in HEAD yet as I expect that would generate a somewhat larger set of issues.
Incidentally I've just found a file with *mixed* line endings. Is that a problem or it doesn't matter? (I can't tell off-hand which one, as I was doing a lot of modifications in one big sweep, so if that is important I'll go and look it up) -- [ Gennaro Prota, C++ developer for hire ]
Gennaro Prota wrote:
But it's still very tedious to switch to the branch, check the diffs etc. Or do you merge blindly? I'm actually tempted to do the work for the 1.34 branch only, and postpone it on the head. The point is, as you well know, that anything being tedious is error-prone. I'm doing the job in little pieces, every time in the day I *feel* like doing it :-)
Dude... Just keep two different checkouts of the Boost tree. One for HEAD and another for RC. No switching required, it takes me a very short, 2 minutes at most, to check into HEAD, merge to RC, check into RC, retag HEAD. Assuming there are no conflicts. And if I'm really paranoid I run WinMerge on the particular subdirs of files I merged between the HEAD and RC checkouts (no CVS mucking needed). Note, I'm a fairly paranoid person ;-)
Incidentally I've just found a file with *mixed* line endings. Is that a problem or it doesn't matter? (I can't tell off-hand which one, as I was doing a lot of modifications in one big sweep, so if that is important I'll go and look it up)
Yes it's a problem that should be getting detected. Is it not showing up on the emails? Or is this on HEAD? -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo
On Tue, 25 Jul 2006 09:59:19 -0500, Rene Rivera <grafikrobot@gmail.com> wrote:
Dude... Just keep two different checkouts of the Boost tree. One for HEAD and another for RC. No switching required
Yes, that's what I do. I'm using WinCVS now, and "switching" simply meant selecting the working directory in the Browse toolbar's combobox. My issue is probably that I want to see the diffs after the check-in, I can't resist.
Incidentally I've just found a file with *mixed* line endings. Is that a problem or it doesn't matter? (I can't tell off-hand which one, as I was doing a lot of modifications in one big sweep, so if that is important I'll go and look it up)
Yes it's a problem that should be getting detected. Is it not showing up on the emails? Or is this on HEAD?
Probably on HEAD. I'll go and check! -- [ Gennaro Prota, C++ developer for hire ]
On Tue, 25 Jul 2006 21:38:55 +0200, Gennaro Prota <gennaro_prota@yahoo.com> wrote:
Incidentally I've just found a file with *mixed* line endings. Is that a problem or it doesn't matter? [...]
Yes it's a problem that should be getting detected. Is it not showing up on the emails? Or is this on HEAD?
Probably on HEAD. I'll go and check!
As supposed. I've gone "a bit" further and fixed all inspector-Rex problems in boost/graph/, except a spurious min/max warning in maximum_cardinality_matching.hpp, actually on a comment line (I didn't feel like ruining the Tutte-Berge formula, sorry). That file also has a name longer than 31 characters. PS: I didn't make a total rewrite of plod_generator and two_bit_color_map :-) The diffs are just screwed up, probably because of the line-endings issues. -- [ Gennaro Prota, C++ developer for hire ]
On Mon, 24 Jul 2006 18:28:41 -0500, Rene Rivera <grafikrobot@gmail.com> wrote:
Historically it hasn't made a difference when we run the inspect tool. People mostly ignored it and hence why we have an accumulation problems now. Now perhaps that it's in the form of a constant nagging email instead of a web page people will pay attention (cross-fingers).
Since better preventing than treating, why we do not use commit triggers (commitinfo checks)? PS: I'll not be able to watch the list or work on the CVS for at least 6/7 hours. When back, I'd like to implement something like // <boost_inspect:nominmax> .... // </boost_inspect:nominmax> to avoid spurious warnings. Ideas, help, hardware and money donations are all welcome ;-) -- [ Gennaro Prota, C++ developer for hire ]
participants (2)
-
Gennaro Prota -
Rene Rivera