I'm trying to get a handle on the current state of the various boost migrations which have been underway. a) migration to GIT b) migration to CMake c) status of rypll when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot? Robert Ramey
Robert Ramey wrote:
I'm trying to get a handle on the current state of the various boost migrations which have been underway.
a) migration to GIT b) migration to CMake c) status of rypll
when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot?
Robert Ramey
For example a) https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus has a bunch of dead links b) what used to be www.rypll.org now seems to be https://github.com/ryppl/ryppl c) this looks like cmake for boost https://github.com/boost-cmake c) https://bitbucket.org/boostorg looks to contain the same information or ? In general, very, very confusing Robert Ramey
On 23/06/13 22:59, Robert Ramey wrote:
Robert Ramey wrote:
I'm trying to get a handle on the current state of the various boost migrations which have been underway.
a) migration to GIT b) migration to CMake c) status of rypll
when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot?
Robert Ramey
For example
a) https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus has a bunch of dead links b) what used to be www.rypll.org now seems to be https://github.com/ryppl/ryppl c) this looks like cmake for boost https://github.com/boost-cmake c) https://bitbucket.org/boostorg looks to contain the same information or ?
In general, very, very confusing
AFAIK migration to git (with modularized repositories) is ongoing, validation of all contributors has been requested. The other projects are on hold until the migration to git is done.
on Wed Jun 26 2013, Mathias Gaunard
On 23/06/13 22:59, Robert Ramey wrote:
Robert Ramey wrote:
I'm trying to get a handle on the current state of the various boost migrations which have been underway.
a) migration to GIT
b) migration to CMake c) status of rypll
when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot?
Robert Ramey
For example
a) https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus has a bunch of dead links b) what used to be www.rypll.org now seems to be https://github.com/ryppl/ryppl c) this looks like cmake for boost https://github.com/boost-cmake c) https://bitbucket.org/boostorg looks to contain the same information or ?
In general, very, very confusing
AFAIK migration to git (with modularized repositories) is ongoing, validation of all contributors has been requested. The other projects are on hold until the migration to git is done.
Correct. * FWIW, http://ryppl.org (only one "l") is still live. * http://bitbucket.org/boostorg contains the same information as http://github.com/boostorg, **as we originally announced**, because it has a potentially more-useful history browser that some might prefer using to review the modularization. Also someone recently checked some enormous files into the SVN sandbox, which prevents us from pushing the sandbox module to github. -- Dave Abrahams
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Monday, July 15, 2013 4:46 PM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
on Wed Jun 26 2013, Mathias Gaunard
wrote: On 23/06/13 22:59, Robert Ramey wrote:
Robert Ramey wrote:
I'm trying to get a handle on the current state of the various boost migrations which have been underway.
a) migration to GIT
b) migration to CMake c) status of rypll
when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot?
Robert Ramey
For example
a) https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus has a bunch of dead links b) what used to be www.rypll.org now seems to be https://github.com/ryppl/ryppl c) this looks like cmake for boost https://github.com/boost-cmake c) https://bitbucket.org/boostorg looks to contain the same information or ?
In general, very, very confusing
AFAIK migration to git (with modularized repositories) is ongoing, validation of all contributors has been requested. The other projects are on hold until the migration to git is done.
Correct.
* FWIW, http://ryppl.org (only one "l") is still live.
* http://bitbucket.org/boostorg contains the same information as http://github.com/boostorg, **as we originally announced**, because it has a potentially more-useful history browser that some might prefer using to review the modularization. Also someone recently checked some enormous files into the SVN sandbox, which prevents us from pushing the sandbox module to github.
This all sounds very promising but can I ask a further question. Has anyone given any thought to how Boost documentation will work in GIT? I've recently converted some documentation and though the conversion itself was quite painless, the links to icons, examples, tests, other Boost library documentation and files caused me some trouble, and would appear likely to be disrupted by GITerization, potentially causing some serious work, especially for big libraries (no prizes for guessing which I particularly have in mind ;). However, don't let me distract from the major task of getting the code and history converted. I'm just raising a flag that this might be a significant issue for the future. Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow@hetp.u-net.com
on Mon Jul 15 2013, "Paul A. Bristow"
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Monday, July 15, 2013 4:46 PM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
on Wed Jun 26 2013, Mathias Gaunard
wrote: On 23/06/13 22:59, Robert Ramey wrote:
Robert Ramey wrote:
I'm trying to get a handle on the current state of the various boost migrations which have been underway.
a) migration to GIT
b) migration to CMake c) status of rypll
when I try to track this down I find many dead links, repeated information. I also don't find some blogs I used to be able to find on these subjects. Anyone up to provide an current status snapshot?
Robert Ramey
For example
a) https://svn.boost.org/trac/boost/wiki/CMakeModularizationStatus has a bunch of dead links b) what used to be www.rypll.org now seems to be https://github.com/ryppl/ryppl c) this looks like cmake for boost https://github.com/boost-cmake c) https://bitbucket.org/boostorg looks to contain the same information or ?
In general, very, very confusing
AFAIK migration to git (with modularized repositories) is ongoing, validation of all contributors has been requested. The other projects are on hold until the migration to git is done.
Correct.
* FWIW, http://ryppl.org (only one "l") is still live.
* http://bitbucket.org/boostorg contains the same information as http://github.com/boostorg, **as we originally announced**, because it has a potentially more-useful history browser that some might prefer using to review the modularization. Also someone recently checked some enormous files into the SVN sandbox, which prevents us from pushing the sandbox module to github.
This all sounds very promising but can I ask a further question.
Has anyone given any thought to how Boost documentation will work in GIT?
*Any* thought? Of course!
I've recently converted some documentation
Converted it to what?
and though the conversion itself was quite painless, the links to icons, examples, tests, other Boost library documentation and files caused me some trouble, and would appear likely to be disrupted by GITerization, potentially causing some serious work, especially for big libraries (no prizes for guessing which I particularly have in mind ;).
Yes, we might want to create a repository for common resources that individual libraries can include as a submodule. But I do expect paths to need adjustment.
However, don't let me distract from the major task of getting the code and history converted.
I'm just raising a flag that this might be a significant issue for the future.
Maybe I'm missing something, but nothing you're describing sounds like it could possibly be a significant issue compared to the general problem of Git conversion and modularization. -- Dave Abrahams
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Tuesday, July 16, 2013 4:46 AM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
This all sounds very promising but can I ask a further question.
Has anyone given any thought to how Boost documentation will work in GIT?
*Any* thought? Of course!
I've recently converted some documentation
Converted it to what?
To Quickbook, Doxygen generated C++ reference section from the code, and Autoindexed with links to examples, header files, other Boost libraries etc.
and though the conversion itself was quite painless, the links to icons, examples, tests, other Boost library documentation and files caused me some trouble, and would appear likely to be disrupted by GITerization, potentially causing some serious work, especially for big libraries (no prizes for guessing which I particularly have in mind ;).
Yes, we might want to create a repository for common resources that individual libraries can include as a submodule.
But I do expect paths to need adjustment.
That might be quite a lot of work ;-) And take some time, especially for older libraries with intermittent maintenance. But keep up the good work. Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow@hetp.u-net.com
on Tue Jul 16 2013, "Paul A. Bristow"
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Tuesday, July 16, 2013 4:46 AM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
This all sounds very promising but can I ask a further question.
Has anyone given any thought to how Boost documentation will work in GIT?
*Any* thought? Of course!
I've recently converted some documentation
Converted it to what?
To Quickbook, Doxygen generated C++ reference section from the code, and Autoindexed with links to examples, header files, other Boost libraries etc.
OK...
and though the conversion itself was quite painless, the links to icons, examples, tests, other Boost library documentation and files caused me some trouble, and would appear likely to be disrupted by GITerization, potentially causing some serious work, especially for big libraries (no prizes for guessing which I particularly have in mind ;).
Yes, we might want to create a repository for common resources that individual libraries can include as a submodule.
But I do expect paths to need adjustment.
That might be quite a lot of work ;-)
OK then, I suppose, it will. But still practically nothing compared to figuring out how to do the conversion.
And take some time, especially for older libraries with intermittent maintenance.
But keep up the good work.
Thanks :-) I hope most of my work is nearly done now. -- Dave Abrahams
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Thursday, July 18, 2013 8:20 AM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
on Tue Jul 16 2013, "Paul A. Bristow"
wrote: Yes, we might want to create a repository for common resources that individual libraries can include as a submodule.
But I do expect paths to need adjustment.
That might be quite a lot of work ;-)
OK then, I suppose, it will. But still practically nothing compared to figuring out how to do the conversion.
And take some time, especially for older libraries with intermittent maintenance.
My original enquiry was to know if you have yet sketched out how the docs for each library would be built and made available, and if they would be available in some sort of combined way for reviewed and accepted libraries, a bit like the Boostbook collection http://www.boost.org/doc/libs/1_54_0/doc/html/index.html, but for *all* libraries, preferably in some standardish format. I also see a need to be able to provide links to other Boost libraries - Boost is highly, and increasingly, incestuous ;-) But if you haven't started consideration of this yet, then getting the main files into GIT format is obviously has total priority.
I hope most of my work is nearly done now.
:-) I'm sure we are all looking forward to the new structure - even if it needs some work to reconfigure etc. Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow@hetp.u-net.com
on Thu Jul 18 2013, "Paul A. Bristow"
My original enquiry was to know if you have yet sketched out how the docs for each library would be built and made available, and if they would be available in some sort of combined way for reviewed and accepted libraries, a bit like the Boostbook collection http://www.boost.org/doc/libs/1_54_0/doc/html/index.html, but for *all* libraries, preferably in some standardish format.
I also see a need to be able to provide links to other Boost libraries - Boost is highly, and increasingly, incestuous ;-)
But if you haven't started consideration of this yet, then getting the main files into GIT format is obviously has total priority.
Actually my plan is to ask other people (like you) to consider those questions and solve those problems. -- Dave Abrahams
On Sat, 20 Jul 2013, at 06:02 PM, Dave Abrahams wrote:
on Thu Jul 18 2013, "Paul A. Bristow"
wrote: My original enquiry was to know if you have yet sketched out how the docs for each library would be built and made available, and if they would be available in some sort of combined way for reviewed and accepted libraries, a bit like the Boostbook collection http://www.boost.org/doc/libs/1_54_0/doc/html/index.html, but for *all* libraries, preferably in some standardish format.
I also see a need to be able to provide links to other Boost libraries - Boost is highly, and increasingly, incestuous ;-)
But if you haven't started consideration of this yet, then getting the main files into GIT format is obviously has total priority.
Actually my plan is to ask other people (like you) to consider those questions and solve those problems.
As far as I'm concerned the plan is to adjust the header links to the new location (probably using a one-off perl script) to get things working quickly. And then try to do something better later. Linking in boostbook is problematic, because xslt has no real support for paths, so I expect the better solution might be to write some kind of post-processor (for html and fop markup) which will adjust custom urls depending on the situation. The urls would be similar to the current 'boost:' urls, but with a way to specify the module, and they'd work for images. Something like 'boost://math/doc/html/quaternions.html' (btw. it would be better if math had forwarding file at places like 'libs/math/quaternion/index.html'). We'd also perhaps distribute a version of the html documentation from before the mapping, so that distributions can adjust the links as they see best. Should probably use a better schema than 'boost', hopefully there's an official convention for custom url schemas. And then, in a future version of quickbook, we could rewrite relative urls to this custom url schema so that they'd behave sanely.
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Daniel James Sent: Saturday, July 20, 2013 5:50 PM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
On Sat, 20 Jul 2013, at 06:02 PM, Dave Abrahams wrote:
on Thu Jul 18 2013, "Paul A. Bristow"
wrote: My original enquiry was to know if you have yet sketched out how the docs for each library would be built and made available, and if they would be available in some sort of combined way for reviewed and accepted libraries, a bit like the Boostbook collection http://www.boost.org/doc/libs/1_54_0/doc/html/index.html, but for *all* libraries, preferably in some standardish format.
I also see a need to be able to provide links to other Boost libraries - Boost is highly, and increasingly, incestuous ;-)
But if you haven't started consideration of this yet, then getting the main files into GIT format is obviously has total priority.
Actually my plan is to ask other people (like you) to consider those questions and solve those problems.
As far as I'm concerned the plan is to adjust the header links to the new location (probably using a one- off perl script) to get things working quickly. And then try to do something better later.
Linking in boostbook is problematic, because xslt has no real support for paths, so I expect the better solution might be to write some kind of post-processor (for html and fop markup) which will adjust custom urls depending on the situation. The urls would be similar to the current 'boost:' urls, but with a way to specify the module, and they'd work for images. Something like 'boost://math/doc/html/quaternions.html' (btw. it would be better if math had forwarding file at places like 'libs/math/quaternion/index.html'). We'd also perhaps distribute a version of the html documentation from before the mapping, so that distributions can adjust the links as they see best. Should probably use a better schema than 'boost', hopefully there's an official convention for custom url schemas.
And then, in a future version of quickbook, we could rewrite relative urls to this custom url schema so that they'd behave sanely.
All sounds plausible but the proof of the pudding ... It is moving from experimental to sandbox to trunk to Boostbook collections that has caused me (and John even) some grief. This is where a ' official convention' on where everything lives might help. I feel that it would be very useful to be able to produce links easily even to the level of section or paragraphs or anchors within other libraries, and of course to their header, test and example files too. I feel these links are a major benefit to readers now that Boost has become so large (and further expansion seems inevitable). It would be nice if the pdf version links work too, but that may be asking too much. Some people like the PDF format for its self-contained nature - and that you can use the Find function in the whole document, something that you can't easily do with html - only the html page that you are on. Some form of local and global index would also be useful. I still find trouble finding things that I know are there, despite googling from boost.org. Your expert input will be totally invaluable of course :-) Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow@hetp.u-net.com
On Sat, 20 Jul 2013, at 07:26 PM, Paul A. Bristow wrote:
All sounds plausible but the proof of the pudding ...
It is moving from experimental to sandbox to trunk to Boostbook collections that has caused me (and John even) some grief. This is where a ' official convention' on where everything lives might help.
Hopefully, git should make this easier, since you can work in a git branch within the normal boost setup.
I feel that it would be very useful to be able to produce links easily even to the level of section or paragraphs or anchors within other libraries, and of course to their header, test and example files too. I feel these links are a major benefit to readers now that Boost has become so large (and further expansion seems inevitable).
Right now the best way to do that is using 'boost:' links. You might be able to use docbook's 'olink' tag, I think that's more of a build problem than anything. http://www.sagehill.net/docbookxsl/Olinking.html
It would be nice if the pdf version links work too, but that may be asking too much. Some people like the PDF format for its self-contained nature - and that you can use the Find function in the whole document, something that you can't easily do with html - only the html page that you are on.
You should be able to setup 'boost:' links to link to the site in pdf documentation. I think by setting 'boost.url.prefix'.
Some form of local and global index would also be useful. I still find trouble finding things that I know are there, despite googling from boost.org.
The big problem there is dealing with different documentation formats. Even quickbook based documentation has a variety of reference formats (e.g. asio, geometry, spirit). Could maybe combine the output from autoindex, John would be the person to ask about that, and it's independent of modularization really.
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Daniel James Sent: Saturday, July 20, 2013 7:09 PM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
On Sat, 20 Jul 2013, at 07:26 PM, Paul A. Bristow wrote:
All sounds plausible but the proof of the pudding ...
It is moving from experimental to sandbox to trunk to Boostbook collections that has caused me (and John even) some grief. This is where a ' official convention' on where everything lives might help.
Hopefully, git should make this easier, since you can work in a git branch within the normal boost setup.
I feel that it would be very useful to be able to produce links easily even to the level of section or paragraphs or anchors within other libraries, and of course to their header, test and example files too. I feel these links are a major benefit to readers now that Boost has become so large (and further expansion seems inevitable).
Right now the best way to do that is using 'boost:' links. You might be able to use docbook's 'olink' tag, I think that's more of a build problem than anything.
http://www.sagehill.net/docbookxsl/Olinking.html
It would be nice if the pdf version links work too, but that may be asking too much. Some people like the PDF format for its self-contained nature - and that you can use the Find function in the whole document, something that you can't easily do with html - only the html page that you are on.
You should be able to setup 'boost:' links to link to the site in pdf documentation. I think by setting 'boost.url.prefix'.
Some form of local and global index would also be useful. I still find trouble finding things that I know are there, despite googling from boost.org.
The big problem there is dealing with different documentation formats. Even quickbook based documentation has a variety of reference formats (e.g. asio, geometry, spirit). Could maybe combine the output from autoindex, John would be the person to ask about that, and it's independent of modularization really.
Thanks for these suggestions. I'll go back to using boost: and see if I can get it to work right with pdfs as you outline. Paul
-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Dave Abrahams Sent: Saturday, July 20, 2013 5:02 PM To: boost@lists.boost.org Subject: Re: [boost] Status of various Boost initiatives?
on Thu Jul 18 2013, "Paul A. Bristow"
wrote: My original enquiry was to know if you have yet sketched out how the docs for each library would be built and made available, and if they would be available in some sort of combined way for reviewed and accepted libraries, a bit like the Boostbook collection http://www.boost.org/doc/libs/1_54_0/doc/html/index.html, but for *all* libraries, preferably in some standardish format.
I also see a need to be able to provide links to other Boost libraries - Boost is highly, and increasingly, incestuous ;-)
But if you haven't started consideration of this yet, then getting the main files into GIT format is obviously has total priority.
Actually my plan is to ask other people (like you) to consider those questions and solve those
problems. Ah - a cunning plan ;-) Cheers Paul
participants (5)
-
Daniel James -
Dave Abrahams -
Mathias Gaunard -
Paul A. Bristow -
Robert Ramey