1.34 release: What about the inspection failures?
Hi there After the call for volunteers to clear the regression failures (can't help with those, currently no access to the relevant platforms), I think it's also time to get busy with the remaining inspection failures: - The *X* report doesn't look too bad, all of the few remaining problems seem to be in libraries that are currently actively maintained. - The *LC* report however does not look good at all. There are a large number of problems and some of them are in public areas (e.g. people) or not actively maintained libs (e.g. pool). What should we do about the *LC* problems? Is it worth bothering with those? Regards, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.
Andreas Huber wrote:
Hi there
After the call for volunteers to clear the regression failures (can't help with those, currently no access to the relevant platforms), I think it's also time to get busy with the remaining inspection failures: - The *X* report doesn't look too bad, all of the few remaining problems seem to be in libraries that are currently actively maintained. - The *LC* report however does not look good at all. There are a large number of problems and some of them are in public areas (e.g. people) or not actively maintained libs (e.g. pool).
What should we do about the *LC* problems? Is it worth bothering with those?
Um, don't know, in theory yes, but there are a lot truly trivial issues showing up, for example there are dozens of html redirect files, that have no content except a http-refresh redirection, we could just mark these so they don't show up in the report I guess? I'm more concerned about all the broken links reported in the X report, some of our documentation looks to be in a very poor state :-( If someone would like to look at this, and/or nag the authors concerned that would be great! John.
John Maddock wrote: [snip]
What should we do about the *LC* problems? Is it worth bothering with those?
Um, don't know, in theory yes, but there are a lot truly trivial issues showing up, for example there are dozens of html redirect files, that have no content except a http-refresh redirection, we could just mark these so they don't show up in the report I guess?
Sure, how can I mark them? What about the People pages? Do they need license and copyright or should we just mark them too?
I'm more concerned about all the broken links reported in the X report, some of our documentation looks to be in a very poor state :-(
Ok.
If someone would like to look at this, and/or nag the authors concerned that would be great!
Aren't people automatically nagged about this anyway? If not I can surely do some personalized nagging :-). Regards, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.
Andreas Huber wrote:
John Maddock wrote: [snip]
What should we do about the *LC* problems? Is it worth bothering with those?
Um, don't know, in theory yes, but there are a lot truly trivial issues showing up, for example there are dozens of html redirect files, that have no content except a http-refresh redirection, we could just mark these so they don't show up in the report I guess?
Sure, how can I mark them?
I don't remember and I can't find it in the docs :-( You may have to dig into the source or ask Rene!
What about the People pages? Do they need license and copyright or should we just mark them too?
Yep I guess they do need license and copyright.
I'm more concerned about all the broken links reported in the X report, some of our documentation looks to be in a very poor state :-(
Ok.
If someone would like to look at this, and/or nag the authors concerned that would be great!
Aren't people automatically nagged about this anyway? If not I can surely do some personalized nagging :-).
I think the messages are only going to the mailing list at present. I guess in order of importance we need to fix: 1) Regressions. 2) Coding guidelines violations (min/max/namespaces etc). 3) Missing copyright on source files. 4) Broken links in docs. 5) Everything else :-) Oh and somewhere into all that, we need to decide whether users will be building 1.34 with bbv2 or bbv1. If the former then we need a new set of getting started instructions at a minimum. If you can instigate some personalised nagging on any of the above and/or fix anything trivial you come across that would certainly be most welcome :-) Thanks! John.
John Maddock wrote:
Andreas Huber wrote:
John Maddock wrote: [snip]
What should we do about the *LC* problems? Is it worth bothering with those? Um, don't know, in theory yes, but there are a lot truly trivial issues showing up, for example there are dozens of html redirect files, that have no content except a http-refresh redirection, we could just mark these so they don't show up in the report I guess? Sure, how can I mark them?
I don't remember and I can't find it in the docs :-(
You may have to dig into the source or ask Rene!
Use "boostinspect:nolicense" or "boostinspect:nocpoyright". But...
What about the People pages? Do they need license and copyright or should we just mark them too?
Yep I guess they do need license and copyright.
All files, including the redirection files, need copyright and license, legally it's what's best.
Aren't people automatically nagged about this anyway? If not I can surely do some personalized nagging :-).
I think the messages are only going to the mailing list at present.
Correct.
I guess in order of importance we need to fix:
1) Regressions. 2) Coding guidelines violations (min/max/namespaces etc). 3) Missing copyright on source files. 4) Broken links in docs. 5) Everything else :-)
Seems right.
Oh and somewhere into all that, we need to decide whether users will be building 1.34 with bbv2 or bbv1.
Given that we haven't been doing regressions with BBv1 I think we've already decided that we'll be building with BBv2. One area of help I would appreciate on this one would be user level testing of the build+install on the various platforms.
If the former then we need a new set of getting started instructions at a minimum.
Dave has been promising he is working on those. In case you want to nag him directly ;-) -- -- 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
Rene Rivera <grafikrobot@gmail.com> writes:
John Maddock wrote:
Oh and somewhere into all that, we need to decide whether users will be building 1.34 with bbv2 or bbv1.
Given that we haven't been doing regressions with BBv1 I think we've already decided that we'll be building with BBv2. One area of help I would appreciate on this one would be user level testing of the build+install on the various platforms.
I'm running regressions with BBv1, I believe. At least, I'm not doing anything to enable a BBv2 build. I'm following the instructions at tools/regression/xsl_reports/runner/instructions.html The only mention of BBv2 I can see is the --v2 option to regression.py, which isn't even documented in the regression test instructions, only in regression.py itself. Anthony -- Anthony Williams Software Developer Just Software Solutions Ltd http://www.justsoftwaresolutions.co.uk
Rene Rivera wrote: [snip]
All files, including the redirection files, need copyright and license, legally it's what's best.
That'll be my task then. [snip]
I guess in order of importance we need to fix:
1) Regressions. 2) Coding guidelines violations (min/max/namespaces etc). 3) Missing copyright on source files. 4) Broken links in docs. 5) Everything else :-)
Seems right.
Ok, I can't currently do anything about 1, but I will nag people about 2-4 and try to fix all the stuff in the "public" areas (like the people pages). Regards, -- Andreas Huber When replying by private email, please remove the words spam and trap from the address shown in the header.
Peter Dimov wrote:
Rene Rivera wrote:
All files, including the redirection files, need copyright and license, legally it's what's best.
Debatable. Claiming copyright for a redirection file seems of questionable legality to me.
I feel the same way... Although really someone did originally compose those files. But as has been demonstrated here before corporate lawyers seem uncomfortable with anything but complete coverage. As I've learned in the past few weeks, because of some personal issues, law is rarely logical :-( -- -- 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
Peter Dimov wrote:
Debatable. Claiming copyright for a redirection file seems of questionable legality to me.
If you do not put the copyright into the file, the file automatically is copyrighted to the writer of the file. Consequentially the copyright must be explicitly stated or none is allowed to copy the file. Roland
Roland Schwarz wrote:
Peter Dimov wrote:
Debatable. Claiming copyright for a redirection file seems of questionable legality to me.
If you do not put the copyright into the file, the file automatically is copyrighted to the writer of the file.
Only if it constitutes an original work. If I create a file with an exact copy of something you wrote, I do not and cannot automatically obtain copyright on it. I need to introduce non-trivial original changes in order to be able to claim copyright on them. If the only thing I introduce is a copyright line that pronounces that I'm the author, I would definitely be on shaky moral and legal ground.
Peter Dimov wrote:
Only if it constitutes an original work. If I create a file with an exact copy of something you wrote, I do not and cannot automatically obtain copyright on it. I need to introduce non-trivial original changes in order to be able to claim copyright on them. If the only thing I introduce is a copyright line that pronounces that I'm the author, I would definitely be on shaky moral and legal ground.
It seems as if we are talking about different things. I am saying that we need the copyright in all files, else there is nothing that gives a possible downloader the right to do so. I was not about to claim that simply copying by typing and adding me as the author changes anything on the rights of the original author. I just wanted to remind, that it is not necessary to claim copyright. You just have it, even if not explicitely stated. (How to prove this fact is another issue.) I reiterate: Boost (or the original authors) need to be explicit in giving the right to users to copy the files. Roland
Hi, Rene Rivera wrote:
John Maddock wrote:
Andreas Huber wrote:
John Maddock wrote: I guess in order of importance we need to fix:
1) Regressions. 2) Coding guidelines violations (min/max/namespaces etc). 3) Missing copyright on source files. 4) Broken links in docs. 5) Everything else :-)
Agreed.
Seems right.
Oh and somewhere into all that, we need to decide whether users will be building 1.34 with bbv2 or bbv1.
Given that we haven't been doing regressions with BBv1 I think we've already decided that we'll be building with BBv2. One area of help I would appreciate on this one would be user level testing of the build+install on the various platforms.
BBv2 is going to be the only supported build system for 1.34.0
If the former then we need a new set of getting started instructions at a minimum.
Dave has been promising he is working on those. In case you want to nag him directly ;-)
Consider this done. The nagging that is. Thomas -- Thomas Witt witt@acm.org
participants (7)
-
Andreas Huber -
Anthony Williams -
John Maddock -
Peter Dimov -
Rene Rivera -
Roland Schwarz -
Thomas Witt