- Boost - lists.boost.org

Re: [boost] Formal Review: Indexed Set
by JOAQUIN LOPEZ MU?Z 28 Mar '04

28 Mar '04

Hello Jeff, thanks for reviewing the indexed_set! ----- Mensaje original ----- De: Jeff Garland <jeff(a)crystalclearsoftware.com> Fecha: Domingo, Marzo 28, 2004 5:21 am Asunto: [boost] Formal Review: Indexed Set > Let me start by saying thanks to Joaquín for the contribution of > this library. You are welcome. The experience has been a lot of fun to me. [...] > > * Do you think the library should be accepted as a Boost library? > Yes. great :) thanx! > > ********* > Now for some questions / comments: > > 1) Dead code in index_node.hpp? > > I don't think the ':' is legal below -- I assume this is dead code? > > template<typename Super> struct > index_node_trampoline:index_node_impl{}; > template<typename Super> > struct index_node:Super,index_node_trampoline<Super> > ... > This is not dead code. index_node_treampoline:index_node_impl is short for index_node_treampoline:public index_node_impl. My coding style is admiteddly a little too compact. > 2) bidirectional_map > > My first interest in this library started when it was just a bi- > directionalmap. Now it has become something much broader and more > powerful. Curiously, > the bi-directional map part seems to have taken a back seat -- > being pushed > into the examples as an 'exercise for the reader'. > > I think that this is unfortunate. I like to see bidirectional_map > as an > official part of the library. A few reasons for this: > > a) bidirectional_map will be documented up front as a part of > the library. > b) For me it is my current primary use of this library -- even > though it is > now so flexible I'll have to rethink how I use containers in designs. > c) Complex code is pushed on the user. For example, the code > below is from > examples/bimap.cpp > > template<typename FromType,typename ToType> > struct bidirectional_map > { > typedef std::pair<FromType,ToType> value_type; > > #if defined(BOOST_NO_POINTER_TO_MEMBER_TEMPLATE_PARAMETERS) ||\ > defined(BOOST_MSVC)&&(BOOST_MSVC<1300) ||\ > > defined(BOOST_INTEL_CXX_VERSION)&&defined(_MSC_VER)&&(BOOST_INTEL_CXX_VERSION<=700) > > .... > > It is obvious from this that there are some portability issues. > Putting this > sort of detail on library users is too mcch to expect from users > in my opinion. > > But it is only a suggestion since it took me all of 5 minutes to > pull out the > core of the example to a header file. I don't think the bidirectional map I provide in the examples section is good enough for officialization (for one, it lacks operator[]). I only wrote it as an example of use which I think will be common enough. IMHO a much more user-friendly bidirectional map could be developed based on indexed_set, but we are talking probably of a separate library. Also, when I first introduced bimap to Boost a year ago, people expressed interest in an extensible approach, which bimap's rigid interface certainly did not provide. As for the macro stuff, you probably shouldn't be concerned if the compiler you work with is reasonably conformant. > > 3) Serialization support > In future directions you say "Once Robert Ramey's serialization > library gets > accepted into Boost..." > > I'd love to see Joaquín at least attempt this soon -- it would > serve as an > excellent review of serialization which is upcoming shortly :-) Good idea, if time permits I'll engage into the review and try to sketch some serialization support for indexed_set. > > 4) Should the SGI/HP copyrights be included in the source files > which are > based off the stl_tree.h implementation? I don't know. The original pieces of code (with lots of changes) are scattered across some five headers or so. Basically it is the rb-tree algorithms that I have retained. Maybe some of the licensing wizards here at the list can offer advice on this issue. > > 5) Naming > > For the namespace/library name I suggest > boost::multiindex_containers. > > Rationale: > -- The library provides several stl-like container types (set, > list, map) -- > hence the plural 'containers'. > -- The containers all have multiple indexes -- hence multiindex. > Maybe boost::container::multi_index would please more people: it is as descriptive as your proposal, but (1) it is into boost::container and (2) avoids the abhorred "ii". Best, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

Re: [boost] Re: Formal Review: Indexed Set
by JOAQUIN LOPEZ MU?Z 28 Mar '04

28 Mar '04

----- Mensaje original ----- De: David Abrahams <dave(a)boost-consulting.com> Fecha: Domingo, Marzo 28, 2004 7:41 am Asunto: [boost] Re: Formal Review: Indexed Set > JOAQUIN LOPEZ MU?Z <joaquin(a)tid.es> writes: > > > Hi Jeff > > > > ----- Mensaje original ----- > > De: Jeff Garland <jeff(a)crystalclearsoftware.com> > > Fecha: Sábado, Marzo 27, 2004 4:55 pm > > Asunto: Re: [boost] Formal Review: Indexed Set > > > > > >> // why doesn't this work? > >> //typedef nth_iterator_type<date_to_string_bimap,2>::type > to_map_itr;> > > It is > > > > typedef nth_iterator_type<date_to_string_bimap,1>::type to_map_itr; > > > > Index numbering is 0-based. I think this is the > > only problem in the program. If you have further > > problems please tell me so. > > Not having looked at the proposal at all, I still think I should raise > the usual concerns about "metafunctions" not accepting type > parameters, which don't play nice with MPL and lambda expressions. > IMHO nth_iterator_type and familiy are user-level facilities (similar vg to Boost.Tuple get<>) and accepting mpl integral constants instead of raw constants does not make much sense here. Please somebody correct me if I'm wrong. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

RE: [boost] Re: Standard C Library and C++ / BOOST
by Reece Dunn 28 Mar '04

28 Mar '04

David Abrahams wrote: >"Klaus Nowikow" <nowikow(a)decomsys.com> writes: >You're wrong. *this is copied before it is thrown. Still, the idiom >above seems a bit suspicious, unless you somehow prevent derivation >from errorcheck. So is it possible to prevent derivation? If not, then at least make it harder to do so (and put a warning in the docs)? Also, would a non-derivable class be a good addition to C++0x? Regards, Reece _________________________________________________________________ It's fast, it's easy and it's free. Get MSN Messenger today! http://www.msn.co.uk/messenger

3 3

RE: [boost] Re: Re: Standard C Library and C++ / BOOST
by Reece Dunn 28 Mar '04

28 Mar '04

Pavel Vozenilek wrote: >"Reece Dunn" <msclrhd(a)hotmail.com> wrote > > So is it possible to prevent derivation? If not, then at least make it > > harder to do so (and put a warning in the docs)? >Yes. >Legal way: >http://www.research.att.com/~bs/bs_faq2.html#no-derivation Cool! I have adapted the example to the errorcheck class, so it will report a clean error message. The code has been tested on borland 5.5 and vc-6, vc-7 and vc-7.1 and should work on other compilers. If there is interest, where in the boost-sandbox would be best to put it? My gut feeling is boost/errorcheck.hpp, but if anyone has any better suggestions... Regards, Reece _________________________________________________________________ Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo

1 0

Re: Formal Review: Indexed Set
by Arkadiy Vertleyb 28 Mar '04

28 Mar '04

Hi Joaquin, Sorry for joining this discussion so late, but here are my comments: - What is your evaluation of the design? - What is your evaluation of the implementation? The code looks clean. One minor question: the index class template's equal_range method forwards to lower/upper_bound. Is it really the best you can achieve here or is it just a temporary shortcut? - What is your evaluation of the documentation? I read the "tutorial" part. I believe the documentation is excelent -- does a great job of explaining things, and provides excelent examples to just be copy-pasted into the code for one to get started. - What is your evaluation of the potential usefulness of the library? At least once in my previous experience I had to maintain multiple synchronized collections. This project would have greatly benefitted from a library like the indexed set. I also think that the "compatible predicate" feature, so unfortunately missing from the std::set, is very useful. We are currently using the library to provide a set-based implementation for our RTL tables. This is not posted yet, but we hope to post it soon. We plan to eventially make the Indexed set-based table implementation default for our tables. - Did you try to use the library? With what compiler? Did you have any problems? Yes, with VC7.1. I did not have any problems related to the Indexed set. - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I read the "tutorial" part of the documentation. Then I wrote Indexed set-based RTL table implementation, and made it default for RTL tables. After this I (eventially) successfully compiled and ran RTL tests, which, by that time all internally utilized the indexed set. I had to solve a few issues that showed up, but turned out to be our own issues (thanks for helping us to discover them). - Are you knowledgeable about the problem domain? I use STL collections a lot but never attempted to implement any non-trivial collection (I don't know if this is qualifies as domain knowledge). - Do you think the library should be accepted as a Boost library? Yes, definitely. Arkadiy

1 0

What should be called? [was: BOOST_CHECK_EQUAL fails...]
by Stefan Slapeta 28 Mar '04

28 Mar '04

Sorry for the repost, but there was no answer to this (http://tinyurl.com/3yba8) and I'm still not sure if it is a compiler or a boost defect. Which f should be called here? void f(const char*) { std::cout << "function called!" << std::endl; } template <typename T> void f(const T&) { std::cout << "function template called!" << std::endl; } int main() { char buf[10] = { 0 }; f(buf); } In my opinion, it should be the function template (which whould also mean that EDG works correctly and VC 7.1 not) as "f(char[10])" is no exact match for the function. However, this would also mean that there is a defect in boost.test (see link above). br, Stefan *********************************** Stefan Slapeta Wien/Vienna, Austria http://www.slapeta.com

2 2

Re: [boost] Formal Review: Indexed Set
by JOAQUIN LOPEZ MU?Z 28 Mar '04

28 Mar '04

Hi Tom, thank you for taking the effort to review indexed_set. ----- Mensaje original ----- De: Tom Brinkman <reportbase(a)yahoo.com> Fecha: Domingo, Marzo 28, 2004 7:46 am Asunto: [boost] Formal Review: Indexed Set > * What is your evaluation of the design? > * What is your evaluation of the implementation? > > This is a very creative and impressive library. It > might become as useful > as the boost::lambda, boost::spirit and > boost::function libraries are for me. > > This library has many features that would be useful > generally to many enterprise > applications. > > * What is your evaluation of the documentation? > > Again, excellant. Great work. > > * What is your evaluation of the potential usefulness > of the library? > > Being able to sort sequences of elements by more than > one sorting criteria is useful > to me. I also am intrigued by the lookup operations > and updating functionality, > mentioned in the help file, although I did not test > these. > > * Did you try to use the library? > > I compiled the samples successfully with gcc. I > re-worked the basic sample to get > a feel for some of the functionality. > > * How much effort did you put into your evaluation? > > Compiled samples and read help files. I created a > small test application. > > * Are you knowledgeable about the problem domain? > > Yes. > > * Do you think the library should be accepted as a > Boost library? > > Yes. > Thank you! If you feel in the mood, you might want to contribute your opinions about naming issues (I'll try to compose a unified naming scheme pleasing as many people as possible) The points under discussion are listed at http://groups.yahoo.com/group/boost/files/indexed_set_review_notes.html Regards, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

1 0

Formal Review: Indexed Set
by Tom Brinkman 28 Mar '04

28 Mar '04

* What is your evaluation of the design? * What is your evaluation of the implementation? This is a very creative and impressive library. It might become as useful as the boost::lambda, boost::spirit and boost::function libraries are for me. This library has many features that would be useful generally to many enterprise applications. * What is your evaluation of the documentation? Again, excellant. Great work. * What is your evaluation of the potential usefulness of the library? Being able to sort sequences of elements by more than one sorting criteria is useful to me. I also am intrigued by the lookup operations and updating functionality, mentioned in the help file, although I did not test these. * Did you try to use the library? I compiled the samples successfully with gcc. I re-worked the basic sample to get a feel for some of the functionality. * How much effort did you put into your evaluation? Compiled samples and read help files. I created a small test application. * Are you knowledgeable about the problem domain? Yes. * Do you think the library should be accepted as a Boost library? Yes. Tom Brinkman __________________________________ Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time. http://taxes.yahoo.com/filing.html

1 0

Re: [boost] Formal Review: Indexed Set
by JOAQUIN LOPEZ MU?Z 28 Mar '04

28 Mar '04

Hi Jeff ----- Mensaje original ----- De: Jeff Garland <jeff(a)crystalclearsoftware.com> Fecha: Sábado, Marzo 27, 2004 4:55 pm Asunto: Re: [boost] Formal Review: Indexed Set > // why doesn't this work? > //typedef nth_iterator_type<date_to_string_bimap,2>::type to_map_itr; It is typedef nth_iterator_type<date_to_string_bimap,1>::type to_map_itr; Index numbering is 0-based. I think this is the only problem in the program. If you have further problems please tell me so. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

3 2

RE: [boost] Re: Boost.Thread and BoostBook
by Reece Dunn 28 Mar '04

28 Mar '04

Michael Glassford wrote: >Douglas Gregor wrote: >Well, that gets me a little farther, but now xsltproc.exe crashes. The >output is attached to the end of the message. >I noticed it was defaulting to gcc, which I don't have installed, so I also >added "using msvc ;" to user-config.jam. It warns me that it found two >versions, 6 and 7.1, and that it will use 7.1, but still crashes. If I >instead add "using msvc : : c:/msvs6/vc98/bin ;" to force it to use MSVC6 >(because version 7.1 unfortunately isn't my main compiler, though we're >thankfully finally moving that way), I get the same result. I think it uses the gcc toolset regardless as a target to put the XIncluded document file in so it will work within bjam/BBv2 (I could have this wrong, as I am not a Boost.Build expert). >Any ideas about this, now? Have you checked the catalog.xml file to see if the namespace aliases are pointing to the correct locations? Does it work when you don't set the XSL/DTD environment variables, but download from the internet? (This will be slow, but it may show if the error is with the program settings or the XSL/DTD settings). Are you building from the Windows or cygwin/Linux shell? If you are using cygwin/Linux, you could try something like: using boostbook : /cygdrive/d/devel/libraries/docbook/xsl/ : /cygdrive/d/devel/libraries/docbook/dtd/ ; Regards, Reece _________________________________________________________________ Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo

2 1