[Catalyst] PPM vs CPAN in a Windows Context
Nilson Santos Figueiredo Junior wrote: OK, don't mean to sound like a whiner here, and I haven't spent any time investigating the various GCC packages, but it's making me laugh that it's been suggested I download a C++ development package just so I can get my perl modules to install. Although I'm primarily a Windows user, I can't really deny that the only one to blame in this situation is Windows itself for not having a way to easily build stuff yourself. *Every* other OS has this sort of things. Even those Linux distros that are completely targetted at the end user (such as Ubuntu) and don't come with a bundled compiler can have it easily installed with a mere apt-get install gcc. I find it rather weird that the suggestion of downloading a C/C++ development package in order to compile C/C++ code makes you laugh. It's like saying that downloading Perl in order to run Perl application makes you laugh. Clearly you've got a wicked sense of humor. LOL. I'm not downloading a C/C++ development package to compile C/C++ code. If I had C/C++ code that I wrote that I wanted to compile, of course I would install a dev environment (in fact I would probably use the Visual Studio installed on my workstation). I'm downloading it to install the *perl* modules I need for my *web* app, and the perl module installation process wants to compile C/C++ modules, which it can not do without my installing this C/C++ compiler (not really an obvious dependency). I know if I wrote a C/C++ app, I wouldn't expect to be required to download a perl dev environment. I know... I'm supposed to do that all myself, and I guess I will, but for now I just don't have time. After downloading the setup file, it takes a double-click plus three aditional clicks in order to get it installed. I really can't imagine someone *this* busy. It's not the download or installation time. It's the analysis of the possible unintended consequences. I have a number of applications I support in my development environment and I try to make as few radical changes as possible, so that everything continues to work. I know I can download and install PPM's because they all go into the perl dev tree. I'm willing to try the CPAN shell as well, however installing all new dev environments with the effects they might have outside of C:\perl is risky. If it's *only* 4 clicks that's fine - if it's 4 clicks and 2 days of environment debugging that's not fine. I mean I only want to finish my small app. My boss is going to split his gut when I tell him first I need to download a C++ dev package so I can install the application framework that actually is written in perl. Catalyst's written in Perl. Most of its dependencies also are. But some of them have parts written in C/C++. How do you expect to compile C/C++ code without a C/C++ compiler? I understand this. What I would *like* is for someone to compile it for me and make it available in a PPM. I also understand that that's not a given and the big boys all use CPAN. One of these days I'll be a big boy also (probably when I have time to do more development work on my Linux box at home), but until then I'll take all the predigested pap I can get! I can't really figure what's the hassle about a 8MB download, some mouse clicks and about 1 or 2 minutes of setup time. I can't really figure out why your boss should even know or care about what you're using to get your job done. The hassle is that we are a Windows shop and my boss only cares about results. To roll out an .ASP application is only a matter of using the resources that are already installed in the development environment and on the production server (IIS, SQL Server 2000, etc.), so I'm already on thin ice with Apache and perl - although I've argued I can code more efficiently in perl than in the .NET environment. If I have to radically alter the production server beyond Perl and Apache/mod_perl then the limb I am going out on will bend substantially more. I do a lot of utility coding in Perl and when I was handed this assignment for a relatively simple web app I thought it would be an excellent opportunity to implement it in Catalyst, since I really don't like .NET. and don't know a thing about Ruby. While doing the initial problem analysis and design work I saw that (apparently) all the needed Catalyst modules were available as PPM's. If that was indeed the case, then rolling out this app would be as simple as installing ActiveState on the production server, using PPM to install the required modules on the production server, and installing and configuring Apache/mod_perl on the production server. I got permission to do that. Now hopefully PAR will work out, or the list of production dependencies will increase dramatically along with the corresponding risk increase in unintended consequences in the production environment. Looks like I'll be sticking to whatever Catalyst
Re: [Catalyst] PPM vs CPAN in a Windows Context
Nilson Santos Figueiredo Junior wrote: On 6/30/06, Hugh Lampert [EMAIL PROTECTED] wrote: The hassle is that we are a Windows shop and my boss only cares about results. To roll out an .ASP application is only a matter of using the resources that are already installed in the development environment and on the production server (IIS, SQL Server 2000, etc.), so I'm already on thin ice with Apache and perl - although I've argued I can code more efficiently in perl than in the .NET environment. If I have to radically alter the production server beyond Perl and Apache/mod_perl then the limb I am going out on will bend substantially more. If you've already got a .NET environment then why don't you use for compiling Perl modules? It's a much better choice than GCC when under Windows since VS is the default compiler for AS Perl. GCC is used as a fallback method since buying VS just for the C compiler is a little overkill. Don't do that. Bad things will happen. Always compiled your modules with the same compiler used for the perl install itself on Windows. To that point, you could compile perl in .NET, then do the modules that way too. -=Chris signature.asc Description: OpenPGP digital signature ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN in a Windows Context
Nilson Santos Figueiredo Junior wrote: The problem is that I never managed to get Apache to run mod_perl properly without crashing. But maybe that's just me, since I've seen other people reporting the opposite. But it works fine enough for my current purposes under Apache::Registry. We've definitely got live users there since one of them reported a bug (long-since-fixed) with DBIx::Class wrt. thread-safety :) -- Matt S Trout Offering custom development, consultancy and support Technical Directorcontracts for Catalyst, DBIx::Class and BAST. Contact Shadowcat Systems Ltd. mst (at) shadowcatsystems.co.uk for more information + Help us build a better perl ORM: http://dbix-class.shadowcatsystems.co.uk/ + ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN in a Windows Context
On 6/30/06, Christopher H. Laco [EMAIL PROTECTED] wrote: Don't do that. Bad things will happen. Always compiled your modules with the same compiler used for the perl install itself on Windows. To that point, you could compile perl in .NET, then do the modules that way too. FUD. VS.NET 2003 compiles everything that's compilable successfully for AS Perl. That's what I've been using for at least two years without any problems. As GCC also does compile everything successfully for AS Perl. In fact, I don't even know if AS Perl is still compiled using the VC6. -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN in a Windows Context
On 6/30/06, Matt S Trout [EMAIL PROTECTED] wrote: I think that's the point - that AS has switched to gcc and it's *generally* preferable to use the same compiler as your perl binary was built with. There's nothing in the release notes indicating that they've done this (they've recently switched to GCC on other platforms but not a single mention of Windows). What they did was to start supporting GCC under Windows. -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Nilson Santos Figueiredo Junior wrote: On 6/28/06, Hugh Lampert [EMAIL PROTECTED] wrote: Err, yes... gcc... seems to be a bit of a problem. Hate to impose on the members of the list, but can anyone point me in the direction of a good win32 binary GCC package that doesn't require Cygwin or other environments? the CPAN module was kind enough to download and install NMAKE from Microsoft itself (that was nice), when I upgraded it. Although it's somewhat like killing a mosquito using a shotgun, I usually install the Dev-Cpp open-source IDE for Windows. It already comes with everything you need to self-compile your modules (GCC/MinGW, etc) and works out-of-the-box. It's not a big download (8mb I think) so it's something pretty reasonable. OK, don't mean to sound like a whiner here, and I haven't spent any time investigating the various GCC packages, but it's making me laugh that it's been suggested I download a C++ development package just so I can get my perl modules to install. There's a couple of nice modules for SQLite that include the SQLite executable maybe it's wishful thinking but it sure would be nice if there was a GCC module that included a binary GCC and any required libraries and headers. I know... I'm supposed to do that all myself, and I guess I will, but for now I just don't have time. I mean I only want to finish my small app. My boss is going to split his gut when I tell him first I need to download a C++ dev package so I can install the application framework that actually is written in perl. Looks like I'll be sticking to whatever Catalyst modules are available in PPM form for now. Not because I'm afraid of installing GCC, but because I can't imagine altering the roll-out environment to the point of installing UNIX emulation layers or C++ development packages just to put this app into production. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
On 6/29/06, Hugh Lampert [EMAIL PROTECTED] wrote: OK, don't mean to sound like a whiner here, and I haven't spent any time investigating the various GCC packages, but it's making me laugh that it's been suggested I download a C++ development package just so I can get my perl modules to install. Although I'm primarily a Windows user, I can't really deny that the only one to blame in this situation is Windows itself for not having a way to easily build stuff yourself. *Every* other OS has this sort of things. Even those Linux distros that are completely targetted at the end user (such as Ubuntu) and don't come with a bundled compiler can have it easily installed with a mere apt-get install gcc. I find it rather weird that the suggestion of downloading a C/C++ development package in order to compile C/C++ code makes you laugh. It's like saying that downloading Perl in order to run Perl application makes you laugh. Clearly you've got a wicked sense of humor. I know... I'm supposed to do that all myself, and I guess I will, but for now I just don't have time. After downloading the setup file, it takes a double-click plus three aditional clicks in order to get it installed. I really can't imagine someone *this* busy. I mean I only want to finish my small app. My boss is going to split his gut when I tell him first I need to download a C++ dev package so I can install the application framework that actually is written in perl. Catalyst's written in Perl. Most of its dependencies also are. But some of them have parts written in C/C++. How do you expect to compile C/C++ code without a C/C++ compiler? I can't really figure what's the hassle about a 8MB download, some mouse clicks and about 1 or 2 minutes of setup time. I can't really figure out why your boss should even know or care about what you're using to get your job done. Looks like I'll be sticking to whatever Catalyst modules are available in PPM form for now. Not because I'm afraid of installing GCC, but because I can't imagine altering the roll-out environment to the point of installing UNIX emulation layers or C++ development packages just to put this app into production. There's no need of installing UNIX emulation layers. If your worries are deployment related, you could easily use PAR, which would pack the already compiled DLLs in a single package. Production servers shouldn't need this installed. Also, unless it's something absolutely necessary, I'd suggest you against deploying it in a Windows server. It's somewhat of a hassle to get mod_perl or FastCGI working correctly under Windows, the best I've got so far is running Catalyst under Apache::Registry, since mod_perl crashes when using PerlModule directives and I can't manage to even compile FastCGI and it's related Perl module and the built-in server becomes really slow if you need to support IE clients directly connecting thanks to the necessary -k switch. -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Hugh Lampert wrote: I mean I only want to finish my small app. My boss is going to split his gut when I tell him first I need to download a C++ dev package so I can install the application framework that actually is written in perl. Looks like I'll be sticking to whatever Catalyst modules are available in PPM form for now. Not because I'm afraid of installing GCC, but because I can't imagine altering the roll-out environment to the point of installing UNIX emulation layers or C++ development packages just to put this app into production. Some modules have bits that need to be compiled so they can hook into the interpreter and/or OS at a lower level than perl code. You don't have to deploy the development tools. Would your boss be surprised that you wanted a copy of the Visual Studio compiler and toolchain in order to write .Net code? Why is that any different from needing a C compiler to compile loadable modules for a language whose runtime is written in C? -- Matt S Trout Offering custom development, consultancy and support Technical Directorcontracts for Catalyst, DBIx::Class and BAST. Contact Shadowcat Systems Ltd. mst (at) shadowcatsystems.co.uk for more information + Help us build a better perl ORM: http://dbix-class.shadowcatsystems.co.uk/ + ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
You may want to look into Vanilla / Strawberry perl as an alternative to ActivePerl. It includes the mingw (gcc) compiler and nmake, and the perl included is compiled from scratch with mingw, rather than ms compilers. http://win32.perl.org/wiki/index.php?title=Vanilla_Perl The files are here (get the .exe) : http://sourceforge.net/project/showfiles.php?group_id=158775package_id=178164release_id=393299 Vanilla Perl is officially experimental because until a few months ago some of the core modules were a bit flakey on windows, but I use it full time for development and running catalyst under fastcgi / apache and have no problems. The quite recent site http://win32.perl.org has a news item added today by Adam Kennedy, saying he's hoping to get a new release of Vanilla Perl and also an initial alpha of Strawberry Perl both out today. Strawberry perl in just Vanilla Perl, but with an up-to-date Bundle::CPAN, and IO / LWP modules - so it's considered a more realistic 'basic' version, rather than vanilla, which is really targeted at people wanting to do CPAN testing. http://win32.perl.org/wiki/index.php?title=Strawberry_Perl Anyway, as I said, I use Vanilla Perl, and have had very little trouble getting everything installed using CPAN.pm - no more trouble than occasionally crops up with other platforms - and the problems that exist have been getting fixed with-a-vengence these last few months. It doesn't come with PPM.pm, but I've written a script that will download PPM's from Kobes' repository and install them for me - I think the only modules I need to do that for are DBD::mysql and Image::Magick. PPM's are just archive files though, so it's easy enough to extract the files from. The next distribution will be Chocolate Perl, which will include a lot more useful modules, including everything that comes with ActivePerl (including PPM.pm) - hopefully we'll get an alpha of this out this year. On 29/06/06, Nilson Santos Figueiredo Junior [EMAIL PROTECTED] wrote: Also, unless it's something absolutely necessary, I'd suggest you against deploying it in a Windows server. It's somewhat of a hassle to get mod_perl or FastCGI working correctly under Windows, the best I've got so far is running Catalyst under Apache::Registry, since mod_perl crashes when using PerlModule directives and I can't manage to even compile FastCGI and it's related Perl module and the built-in server becomes really slow if you need to support IE clients directly connecting thanks to the necessary -k switch. I use a binary fastcgi apache module which I downloaded from the fastcgi website. I can't remember whether I had problems compiling FCGI.pm - maybe that's one of the few I had to get a PPM for. If you use PPM.pm, make sure you add Randy Kobes' cpan mirror repository. If anyone has problems with compiling/installing modules on windows, it'd be really appreciated if you could report the problem on rt.perl.org, and post a note on the Compatibility List of Perl Modules wiki page, so people know to chase it up. http://win32.perl.org/wiki/index.php?title=Compatibility_List_of_Perl_Modules Cheers, Carl ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Nilson Santos Figueiredo Junior wrote: On 6/26/06, Hugh Lampert [EMAIL PROTECTED] wrote: Just to let you know something amusing - I tried the preconfigured CPAN.pm that comes with ActiveState Perl 5.8.8. First, it took about 10 minutes to figure out I needed to type Enter+Space instead of just Enter to get my commands accepted. Then, upon trying the i command, I was informed there was a newer CPAN bundle available and that it was suggested I upgrade in place. After typing install Bundle::CPAN the most amazing chain of downloads, dialogs, makes, compiles, test outputs, etc. was initiated I was truly in fear for my life that something was going to get severely screwed up when it started downloading and compiling Crypt and PGP modules. In the end, after the upgrade I was forced to go through the CPAN configuration dialog anyway. Everything seems to have been successfully installed however and nothing seems to be amiss. I DO think for my own sanity that I will stick with PPM's where available, as THAT process seems to be mostly just download and copy to the appropriate location. This is rather weird. You shouldn't need to type Enter+Space. In fact, I'm not even sure if I understood you correctly. The CPAN shell is a regular command shell. You type your commands and press Enter as in any other shell. It IS weird - but on my Windows XP workstation the CPAN shell command interpreter does not accept commands when I hit Enter unless the Enter key is followed by a space. I have no idea what kind of parser is involved. It's not really important though because it DOES work. It's kind of like the Perl debugger not being restartable in ActiveState Perl (it gives some kind of POSIX constant not defined error)... it's annoying to a minor extent but still workable. You didn't really need to upgrade the CPAN shell but when you do it when it asks you if you're ready for manual configuration all you need to do is type no and it will auto-configure itself. Now, I know *this* is rather counter-intuitive. But the rest seems pretty intuitive to me. Coming from the Windows world, I am severely suspicious of allowing ANYTHING to configure itself. I worked through the dialog, it really wasn't a problem, just not what I wanted to do with my boss breathing down my neck regarding my choice of Catalyst as an application platform. The thing about using the CPAN shell is that you're able to use more modules and you'll usually have newer versions of the modules. You'll be able to see that some modules have optional dependencies and choose wether to install them or not (PPM doesn't follow optional dependencies). Of course it takes a little bit longer when installing through the CPAN shell because it actually runs the test cases on your machines so you're even safer when using it (it's perfectly possible that a module worked alright at ActiveState's server but doesn't work correctly on your machine due to various circumstances). I think you were a little bit frightened probably because it installed *a lot* a modules. This probablu happened because, since you previously only used PPM, you had a lot of outdates modules and when installing Bundle::* from the CPAN it installs everything in that bundle unless you've got the newest version. When installing normal modules it will only install a newer version if it's a module requirement. After installing some heavy-weight modules such as Task::Catalyst, DBIx::Class, Bundle::CPAN, PAR and POE most of your installs will end up being single module installs since you'll already have most of the usual dependencies installed. BTW, Why is it Task::Catalyst and not Bundle::Catalyst? I want to install this but it does not run, getting an NMAKE fatal error U1077, errors looking for GPG, etc. This is why I like the PPM packages... I'm assuming that anything that fails to make does not get installed in the perl lib tree, correct? -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Hugh Lampert wrote: Nilson Santos Figueiredo Junior wrote: On 6/26/06, Hugh Lampert [EMAIL PROTECTED] wrote: Just to let you know something amusing - I tried the preconfigured CPAN.pm that comes with ActiveState Perl 5.8.8. First, it took about 10 minutes to figure out I needed to type Enter+Space instead of just Enter to get my commands accepted. Then, upon trying the i command, I was informed there was a newer CPAN bundle available and that it was suggested I upgrade in place. After typing install Bundle::CPAN the most amazing chain of downloads, dialogs, makes, compiles, test outputs, etc. was initiated I was truly in fear for my life that something was going to get severely screwed up when it started downloading and compiling Crypt and PGP modules. In the end, after the upgrade I was forced to go through the CPAN configuration dialog anyway. Everything seems to have been successfully installed however and nothing seems to be amiss. I DO think for my own sanity that I will stick with PPM's where available, as THAT process seems to be mostly just download and copy to the appropriate location. This is rather weird. You shouldn't need to type Enter+Space. In fact, I'm not even sure if I understood you correctly. The CPAN shell is a regular command shell. You type your commands and press Enter as in any other shell. It IS weird - but on my Windows XP workstation the CPAN shell command interpreter does not accept commands when I hit Enter unless the Enter key is followed by a space. I have no idea what kind of parser is involved. It's not really important though because it DOES work. It's kind of like the Perl debugger not being restartable in ActiveState Perl (it gives some kind of POSIX constant not defined error)... it's annoying to a minor extent but still workable. You didn't really need to upgrade the CPAN shell but when you do it when it asks you if you're ready for manual configuration all you need to do is type no and it will auto-configure itself. Now, I know *this* is rather counter-intuitive. But the rest seems pretty intuitive to me. Coming from the Windows world, I am severely suspicious of allowing ANYTHING to configure itself. I worked through the dialog, it really wasn't a problem, just not what I wanted to do with my boss breathing down my neck regarding my choice of Catalyst as an application platform. Once you've got CPAN configured and nmake and a gcc installed, http://shadowcatsystems.co.uk/static/cat-install will install Catalyst itself plus deps hands-off via CPAN (with a little help from ppm on windows) BTW, Why is it Task::Catalyst and not Bundle::Catalyst? I want to install this but it does not run, getting an NMAKE fatal error U1077, errors looking for GPG, etc. This is why I like the PPM packages... I'm assuming that anything that fails to make does not get installed in the perl lib tree, correct? See http://search.cpan.org/~adamk/Task-1.01/lib/Task.pm for an explanation. -- Matt S Trout Offering custom development, consultancy and support Technical Directorcontracts for Catalyst, DBIx::Class and BAST. Contact Shadowcat Systems Ltd. mst (at) shadowcatsystems.co.uk for more information + Help us build a better perl ORM: http://dbix-class.shadowcatsystems.co.uk/ + ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Matt S Trout wrote: Once you've got CPAN configured and nmake and a gcc installed, http://shadowcatsystems.co.uk/static/cat-install will install Catalyst itself plus deps hands-off via CPAN (with a little help from ppm on windows) Err, yes... gcc... seems to be a bit of a problem. Hate to impose on the members of the list, but can anyone point me in the direction of a good win32 binary GCC package that doesn't require Cygwin or other environments? the CPAN module was kind enough to download and install NMAKE from Microsoft itself (that was nice), when I upgraded it. BTW, Why is it Task::Catalyst and not Bundle::Catalyst? I want to install this but it does not run, getting an NMAKE fatal error U1077, errors looking for GPG, etc. This is why I like the PPM packages... I'm assuming that anything that fails to make does not get installed in the perl lib tree, correct? See http://search.cpan.org/~adamk/Task-1.01/lib/Task.pm for an explanation. Thanks, that was very informative! (well beyond my level, but still very informative). -- Hugh ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
On Wed, Jun 28, 2006 at 01:09:13PM -0400, Hugh Lampert wrote: Err, yes... gcc... seems to be a bit of a problem. Hate to impose on the members of the list, but can anyone point me in the direction of a good win32 binary GCC package that doesn't require Cygwin or other environments? Depends how strictly we interpret your question. If you can handle having a runtime DLL, you should be able to install gcc and binutils without needing a full Cygwin/UWIN/MinGW32 environment. If you want something fully native, you'll still not have anything providing (most of) the POSIX function set just using MS' libraries, AFAIK. http://www.mingw.org/x86-win32-ports.shtml Your best bet will be to install something from that list, I guess. Perhaps somebody should write a gcc-compatible C compiler in Perl. /joel ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
Hugh Lampert wrote: Matt S Trout wrote: Once you've got CPAN configured and nmake and a gcc installed, http://shadowcatsystems.co.uk/static/cat-install will install Catalyst itself plus deps hands-off via CPAN (with a little help from ppm on windows) Err, yes... gcc... seems to be a bit of a problem. Hate to impose on the members of the list, but can anyone point me in the direction of a good win32 binary GCC package that doesn't require Cygwin or other environments? the CPAN module was kind enough to download and install NMAKE from Microsoft itself (that was nice), when I upgraded it. From the top of that script - # This is the Shadowcat Catalyst installer. Its purpose is to make it easier # and quicker to get started with Catalyst development. In order to use it, # make sure you have perl 5.8.1+, a make and a compiler, (nmake and dev-c++ # are good on windows), a configured CPAN.pm and Module::Build installed. -- Matt S Trout Offering custom development, consultancy and support Technical Directorcontracts for Catalyst, DBIx::Class and BAST. Contact Shadowcat Systems Ltd. mst (at) shadowcatsystems.co.uk for more information + Help us build a better perl ORM: http://dbix-class.shadowcatsystems.co.uk/ + ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN
On 6/28/06, Hugh Lampert [EMAIL PROTECTED] wrote: Err, yes... gcc... seems to be a bit of a problem. Hate to impose on the members of the list, but can anyone point me in the direction of a good win32 binary GCC package that doesn't require Cygwin or other environments? the CPAN module was kind enough to download and install NMAKE from Microsoft itself (that was nice), when I upgraded it. Although it's somewhat like killing a mosquito using a shotgun, I usually install the Dev-Cpp open-source IDE for Windows. It already comes with everything you need to self-compile your modules (GCC/MinGW, etc) and works out-of-the-box. It's not a big download (8mb I think) so it's something pretty reasonable. The only manual configuration I remeber having to do is to add the GCC /bin dir to my system PATH variable. And then the CPAN shell from ActiveState will automatically configure itself and work using GCC. Since you've upgraded your CPAN shell using a version from the CPAN, I don't know if the magic still works and you might have to manually configure your compiler parameters (but that's a one time thing). -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/
Re: [Catalyst] PPM vs CPAN (Was: Problem with Catalyst Authorization)
On 6/26/06, Hugh Lampert [EMAIL PROTECTED] wrote: Thanks for responding! Perhaps you can forward this to the mailing list, as I am unable to reach it from my work (the mail server will not accept my relays.) Done. I'll answer your message without removing anything you wrote. I looked at the CPAN shell and it seems to be harder to use and less flexible than the PPM shell (Couldn't find an easy way find module descriptions like the PPM describe command, for example). Also, PPM's are pretty much guaranteed by the packagers to work with ActiveState Perl - I've been hesitant to install modules from CPAN due to fear of blowing up my Perl installs. Perhaps that's a foolish fear? I mean development is time intensive enough without debugging problems introduced by incompatible modules. Your discussion of the problem with Module::Install is a good example that I'd prefer to not have to worry about if I didn't have to. I also see that the CPAN shell is soon to be deprecated in favor of CPANPLUS in Perl 5.10 - Is it worth it for me to switch from PPM to something new that won't be applicable shortly? I really can't understand what you mean by flexible, then. I agree that it's a little bit easier to use the PPM shell but it's really *a little bit* (basically, at the CPAN shell you'll have to confirm the install of dependencies [even when you set your preferences to automatically follow them] while the PPM shell will do this automatically - I could argue this is one of the lack of flexibilities of the PPM shell). It really is a foolish fear, IMO. There's no such thing as blowing up your Perl install. The same sense of safety provided by the PPM shell regarding module compatibility is also available at the CPAN shell (i.e. version checking of dependencies, because that's pretty much the compatibility that is verified). I think this subject really can be summed up like this: the things that are problematic when dealing with the CPAN shell are usually not even available when you're restricted to the PPM shell. So, when you get an error when using the CPAN shell, probably that module (or module version) isn't even available at ActiveState's repositories. At least, you'll have a chance to fix it yourself and make it work instead of not even having the choice to do it. -Nilson Santos F. Jr. ___ List: Catalyst@lists.rawmode.org Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/ Dev site: http://dev.catalyst.perl.org/