[1.32 release] Regressions clean-up
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state: - Date/time, http://tinyurl.com/563b7 - Filesystem, http://tinyurl.com/4kshr - UBLAS, http://tinyurl.com/3vwcz - Thread, http://tinyurl.com/4s58j Also, most of the new libraries, namely - algorithm/string - numeric/conversion - range - program_options are either missing explicit markup, or unexpectedly failing. I'd appreciate if the corresponding library authors / maintainers (or simply volunteers) looked over these failures and cleaned things up a little, brigning us closer to the sacred "all green" field. Thank you, -- Aleksey Gurtovoy MetaCommunications Engineering
Aleksey Gurtovoy wrote:
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state:
- Date/time, http://tinyurl.com/563b7 - Filesystem, http://tinyurl.com/4kshr - UBLAS, http://tinyurl.com/3vwcz - Thread, http://tinyurl.com/4s58j
Also, most of the new libraries, namely
- algorithm/string - numeric/conversion - range - program_options
are either missing explicit markup, or unexpectedly failing.
FWIW, I've mentioned to you that program_options has much more markup than shows up in summary page. For example: http://tinyurl.com/4plgu shows a note for a particular failure, but the summary page: http://tinyurl.com/4ghk7 has no hint on that. The _dll failures is something which is still in works. - Volodya
Vladimir Prus writes:
Aleksey Gurtovoy wrote:
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state:
- Date/time, http://tinyurl.com/563b7 - Filesystem, http://tinyurl.com/4kshr - UBLAS, http://tinyurl.com/3vwcz - Thread, http://tinyurl.com/4s58j
Also, most of the new libraries, namely
- algorithm/string - numeric/conversion - range - program_options
are either missing explicit markup, or unexpectedly failing.
FWIW, I've mentioned to you that program_options has much more markup than shows up in summary page. For example:
shows a note for a particular failure, but the summary page:
has no hint on that.
It has been put on the TODO list; sorry for not making it clear. In my above message, I wasn't referring to *these* failures, though -- after all, they are green.
The _dll failures is something which is still in works.
OK, good to know. Thanks for clarifying, -- Aleksey Gurtovoy MetaCommunications Engineering
I'm unable to contact tinyurl.com. Do you also experience problems ? Aleksey Gurtovoy wrote:
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state:
- Date/time, http://tinyurl.com/563b7 - Filesystem, http://tinyurl.com/4kshr - UBLAS, http://tinyurl.com/3vwcz - Thread, http://tinyurl.com/4s58j
Also, most of the new libraries, namely
- algorithm/string - numeric/conversion - range - program_options
are either missing explicit markup, or unexpectedly failing.
I'd appreciate if the corresponding library authors / maintainers (or simply volunteers) looked over these failures and cleaned things up a little, brigning us closer to the sacred "all green" field.
Thank you,
On Tue, 13 Jul 2004 03:18:38 -0500, Aleksey Gurtovoy wrote
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state:
- Date/time, http://tinyurl.com/563b7
We have been and will continue to work this. However, we added wstream and other features (awhile ago now) which are simply not supported in many compilers. So we will just be marking these failed. In addition, we have stopped 'covering up' for bad compilers / standard libraries. In the past we had macros that would swap out some tests on VC6 and other compilers to make the tests look better. We've now removed these and are letting these tests fail. So again we need to mark more tests as failing. I believe this will help people better understand the limits of using old compilers, etc... As for the boost book conversion -- all the xml is all checked in. I just need to do the changes in the top level. Probably later today. Jeff
Jeff Garland writes:
On Tue, 13 Jul 2004 03:18:38 -0500, Aleksey Gurtovoy wrote
According to http://tinyurl.com/25ar3, the following libraries appear to be in a particularly untidy state:
- Date/time, http://tinyurl.com/563b7
We have been and will continue to work this. However, we added wstream and other features (awhile ago now) which are simply not supported in many compilers. So we will just be marking these failed.
OK, sounds good.
In addition, we have stopped 'covering up' for bad compilers / standard libraries. In the past we had macros that would swap out some tests on VC6 and other compilers to make the tests look better. We've now removed these and are letting these tests fail. So again we need to mark more tests as failing. I believe this will help people better understand the limits of using old compilers, etc...
Sure. Looking forward towards the marked-up picture.
As for the boost book conversion -- all the xml is all checked in. I just need to do the changes in the top level. Probably later today.
Excellent. Thanks for the update! -- Aleksey Gurtovoy MetaCommunications Engineering
Aleksey Gurtovoy <agurtovoy <at> meta-comm.com> writes:
Also, most of the new libraries, namely
- range
are either missing explicit markup, or unexpectedly failing.
This is one of the libraries which should probably not have a link on the main page yet. I did take a quick look at the regression-tests and saw all compilers without partial template spcialization fail. Some of these errors are a bit hard to combat when one does not have the compilers that are failing. Therefore I hope that people who care about boost.range on their XX compiler could try finding the fixes, sometimes probably just different compiler flags could do it. br Thorsten
participants (5)
-
Aleksey Gurtovoy -
Jeff Garland -
Thorsten Ottosen -
Toon Knapen -
Vladimir Prus