Re: [R] Disable LazyLoading mechanism completely
On Sat, 1 Apr 2006, Duncan Murdoch wrote: On 3/31/2006 6:03 PM, Paul Roebuck wrote: On Fri, 31 Mar 2006, Duncan Murdoch wrote: On 3/31/2006 5:15 PM, Paul Roebuck wrote: Is there a global option somewhere that can completely disable the LazyLoad option? I want all my packages in source format for searching purposes, crippled by the conversion to database format. Why not just keep the source? Maybe I misunderstood where you're thinking of searching, but I think even without LazyLoad, you'll lose comments in the functions. My thought is something along the lines of options(lazy.load = FALSE) which would override the setting in the DESCRIPTION file as though every package had set it as: LazyLoad: no I want the ability to patch/search the code by default. LazyLoad removes this capability. I know there will be a minor speed hit and accept that in exchange for the above capability. Lazy loading just replaces each object with a promise to load it, right? So if you need to do a search/patch, couldn't you force those promises during the search? As far as searching, think outside the bun - grep, mdfind. Patching-wise, used to be able to just correct package.R and try again, which no longer exists as a compilation of the R source files from the package in text format. Much faster development-wise than constantly reinstalling the package. -- SIGSIG -- signature too long (core dumped) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On Sat, 1 Apr 2006, Prof Brian Ripley wrote: On Fri, 31 Mar 2006, Dirk Eddelbuettel wrote: On 31 March 2006 at 16:15, Paul Roebuck wrote: | Is there a global option somewhere that can completely | disable the LazyLoad option? I want all my packages | in source format for searching purposes, crippled by | the conversion to database format. AFAICT setting LazyLoad=No in DESCRIPTION might do the trick. For reasons I fail to understand I have to set it to that value for Debian builds of the Rcmdr package, and the resulting package does indeed have R code as plain sources. I'd agree with you that being able to read, and hence search, the code is a good thing but I also trust the R masters on the advantages of lazy loading. There may be an option to diable it globally ... Quite a large number of packages (including almost all using S4 methods) *do not work* without lazyloading or saved images. So this option would just break other people's work. Can you elaborate a bit on this statement? S4 predates LazyLoad and used to work just fine before that mechanism was introduced. Indeed, several of my own packages have S4 methods and work with neither LazyLoading nor saved images. Should that be read as since package authors made one-way changes to support LazyLoad? Is there some way to tell whether a package can do without it? -- SIGSIG -- signature too long (core dumped) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On 3/31/2006 6:03 PM, Paul Roebuck wrote: On Fri, 31 Mar 2006, Duncan Murdoch wrote: On 3/31/2006 5:15 PM, Paul Roebuck wrote: Is there a global option somewhere that can completely disable the LazyLoad option? I want all my packages in source format for searching purposes, crippled by the conversion to database format. Why not just keep the source? Maybe I misunderstood where you're thinking of searching, but I think even without LazyLoad, you'll lose comments in the functions. My thought is something along the lines of options(lazy.load = FALSE) which would override the setting in the DESCRIPTION file as though every package had set it as: LazyLoad: no I want the ability to patch/search the code by default. LazyLoad removes this capability. I know there will be a minor speed hit and accept that in exchange for the above capability. Lazy loading just replaces each object with a promise to load it, right? So if you need to do a search/patch, couldn't you force those promises during the search? Duncan Murdoch __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Disable LazyLoading mechanism completely
Is there a global option somewhere that can completely disable the LazyLoad option? I want all my packages in source format for searching purposes, crippled by the conversion to database format. -- SIGSIG -- signature too long (core dumped) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On 3/31/2006 5:15 PM, Paul Roebuck wrote: Is there a global option somewhere that can completely disable the LazyLoad option? I want all my packages in source format for searching purposes, crippled by the conversion to database format. Why not just keep the source? Maybe I misunderstood where you're thinking of searching, but I think even without LazyLoad, you'll lose comments in the functions. Duncan Murdoch __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On Fri, 31 Mar 2006, Duncan Murdoch wrote: On 3/31/2006 5:15 PM, Paul Roebuck wrote: Is there a global option somewhere that can completely disable the LazyLoad option? I want all my packages in source format for searching purposes, crippled by the conversion to database format. Why not just keep the source? Maybe I misunderstood where you're thinking of searching, but I think even without LazyLoad, you'll lose comments in the functions. My thought is something along the lines of options(lazy.load = FALSE) which would override the setting in the DESCRIPTION file as though every package had set it as: LazyLoad: no I want the ability to patch/search the code by default. LazyLoad removes this capability. I know there will be a minor speed hit and accept that in exchange for the above capability. -- SIGSIG -- signature too long (core dumped) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On 31 March 2006 at 16:15, Paul Roebuck wrote: | Is there a global option somewhere that can completely | disable the LazyLoad option? I want all my packages | in source format for searching purposes, crippled by | the conversion to database format. AFAICT setting LazyLoad=No in DESCRIPTION might do the trick. For reasons I fail to understand I have to set it to that value for Debian builds of the Rcmdr package, and the resulting package does indeed have R code as plain sources. I'd agree with you that being able to read, and hence search, the code is a good thing but I also trust the R masters on the advantages of lazy loading. There may be an option to diable it globally ... Hth, Dirk -- Hell, there are no rules here - we're trying to accomplish something. -- Thomas A. Edison __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Disable LazyLoading mechanism completely
On Fri, 31 Mar 2006, Dirk Eddelbuettel wrote: On 31 March 2006 at 16:15, Paul Roebuck wrote: | Is there a global option somewhere that can completely | disable the LazyLoad option? I want all my packages | in source format for searching purposes, crippled by | the conversion to database format. AFAICT setting LazyLoad=No in DESCRIPTION might do the trick. For reasons I fail to understand I have to set it to that value for Debian builds of the Rcmdr package, and the resulting package does indeed have R code as plain sources. I'd agree with you that being able to read, and hence search, the code is a good thing but I also trust the R masters on the advantages of lazy loading. There may be an option to diable it globally ... Quite a large number of packages (including almost all using S4 methods) *do not work* without lazyloading or saved images. So this option would just break other people's work. Rcmdr comes with LazyLoad: no in its DESCRIPTION file (so I have no idea why you need to do anything different). John Fox and I worked out why this was needed, and that lazy loading was not helpful as that package does access almost all its code at startup. -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
