CPAN::Smoke (was Re: Test::Harness::Straps)
On Thu, Sep 20, 2001 at 02:52:20PM -0400, Shane Landrum wrote: > My cursory look at CPAN::Smoke says that it's sort of what I want. > Except that I don't want to have to rely on having a CPAN mirror Look at CPAN::Smoke::Module and CPAN::Smoke::Author to set up modules without a CPAN index. > I guess I could build it myself, but perhaps a layer of abstraction > in CPAN::Smoke would be useful. CPAN::Smoke is still in the "just get the damned thing working" stage. -- Michael G. Schwern <[EMAIL PROTECTED]>http://www.pobox.com/~schwern/ Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One Your words scratch the back of my eyes! http://sluggy.com/d/010204.html
Re: Test::Harness::Straps
On Thu, Sep 20, 2001 at 11:56:57AM -0400, Michael G Schwern ([EMAIL PROTECTED]) wrote: > > Ok, I'll have to trap STDERR. It's going to be tricky to preserve the > interweaving of STDOUT and STDERR though, and to do it without > IPC::Open3. Thanks. In the meantime, I'll see what I can do as a user to get similar results. > > Also, I assume that T::H::S wants to be pointed at a module directory, > > where it can "perl Makefile.PL && make && make test"--- so that > > I can rebuild my embedded tests and get test results programmatically. > > That doesn't seem to be how it works; I'm doing a lot of that at > > user level instead. Seems somewhat clunky. > > That's out of Test::Harness's scope. For something like that you want > CPAN::Smoke. My cursory look at CPAN::Smoke says that it's sort of what I want. Except that I don't want to have to rely on having a CPAN mirror locally, using something like CPAN::Site. I may have a CPAN::Site at my site eventually, but not right now. I'd appreciate a smoking framework that doesn't rely on a local CPAN mirror. I guess I could build it myself, but perhaps a layer of abstraction in CPAN::Smoke would be useful. srl -- Shane Landrum (srl AT boston DOT com) Software Engineer, boston.com
RE: Test::Harness::Straps
From: Kirrily Robert [[EMAIL PROTECTED]] > In perl.qa, Schwern wrote: > > Hmmm, more people trying it, really, especially on the weird > > platforms. I rewrote all the test analysis logic and I > > still afraid I broke something. > We're probably going to start using it for e-smith's testing foo. > I've dinked around with it briefly, enough to know that it > does roughly what I want, but I'm not on any kind of unusual > platform or anything, so I don't know if that's at all helpful. It looks like I'm going to be putting it into use here as well for testing our Mason component architecture. x86 platform, web-based testing. Should be interesting. Unfortunately, I don't think I can get it installed as an alpha. I'll try to experiment on my own, though. R.
Re: Test::Harness::Straps
On Thu, Sep 20, 2001 at 12:17:08PM -0400, Kirrily Robert wrote: > We're probably going to start using it for e-smith's testing foo. > I've dinked around with it briefly, enough to know that it does roughly > what I want, but I'm not on any kind of unusual platform or anything, so > I don't know if that's at all helpful. Wait for 2.00_02. I just have to figure out CVS merging. -- Michael G. Schwern <[EMAIL PROTECTED]>http://www.pobox.com/~schwern/ Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One That which stirs me, stirs everything. -- Squonk Opera, "Spoon"
Re: Test::Harness::Straps
In perl.qa, you wrote: >On Thu, Sep 20, 2001 at 10:35:49AM -0400, Shane Landrum wrote: >> So, I've just found Test::Harness::Straps--- thanks to Skud >> for pointing me in the right direction. Anyone else using it? >> I'm working on using it to write a web-based test summarizer >> for my users. >> >> Schwern, do you have plans for releasing T::H 2.00 to CPAN >> as anything other than an alpha? What do you need to see >> before you release it? > >Hmmm, more people trying it, really, especially on the weird >platforms. I rewrote all the test analysis logic and I still afraid I >broke something. We're probably going to start using it for e-smith's testing foo. I've dinked around with it briefly, enough to know that it does roughly what I want, but I'm not on any kind of unusual platform or anything, so I don't know if that's at all helpful. K. -- Kirrily 'Skud' Robert - [EMAIL PROTECTED] - http://infotrope.net/ "Make a fist, Benno. No, that's too big." -- Renee (from the Netizen quotes file)
Re: Test::Harness::Straps
On Thu, Sep 20, 2001 at 11:30:38AM -0400, Shane Landrum wrote: > I want T::H::Straps to be able to tell me whether there were warnings > during my test run, and if so, what they were. Working modules > don't generally warn. Perhaps that should be switchable. Ok, I'll have to trap STDERR. It's going to be tricky to preserve the interweaving of STDOUT and STDERR though, and to do it without IPC::Open3. > Also, I assume that T::H::S wants to be pointed at a module directory, > where it can "perl Makefile.PL && make && make test"--- so that > I can rebuild my embedded tests and get test results programmatically. > That doesn't seem to be how it works; I'm doing a lot of that at > user level instead. Seems somewhat clunky. That's out of Test::Harness's scope. For something like that you want CPAN::Smoke. > Also, total_results doesn't yet work, and I want it to--- related > to the last paragraph. That's what I assume total_results will > give me when it's implemented. Nope, total_results is just a hash of file that have been analyzed and their results. -- Michael G. Schwern <[EMAIL PROTECTED]>http://www.pobox.com/~schwern/ Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One Stupid am I? Stupid like a fox!
Re: Test::Harness::Straps
On Thu, Sep 20, 2001 at 10:58:08AM -0400, Michael G Schwern ([EMAIL PROTECTED]) wrote: > > Hmmm, more people trying it, really, especially on the weird > platforms. I rewrote all the test analysis logic and I still afraid I > broke something. > > The interface needs to be gelled a bit. Half of it is unimplemented. Ok. 2 things: I want T::H::Straps to be able to tell me whether there were warnings during my test run, and if so, what they were. Working modules don't generally warn. Perhaps that should be switchable. Also, I assume that T::H::S wants to be pointed at a module directory, where it can "perl Makefile.PL && make && make test"--- so that I can rebuild my embedded tests and get test results programmatically. That doesn't seem to be how it works; I'm doing a lot of that at user level instead. Seems somewhat clunky. Also, total_results doesn't yet work, and I want it to--- related to the last paragraph. That's what I assume total_results will give me when it's implemented. srl -- Shane Landrum (srl AT boston DOT com) Software Engineer, boston.com
Re: Test::Harness::Straps
On Thu, Sep 20, 2001 at 10:35:49AM -0400, Shane Landrum wrote: > So, I've just found Test::Harness::Straps--- thanks to Skud > for pointing me in the right direction. Anyone else using it? > I'm working on using it to write a web-based test summarizer > for my users. > > Schwern, do you have plans for releasing T::H 2.00 to CPAN > as anything other than an alpha? What do you need to see > before you release it? Hmmm, more people trying it, really, especially on the weird platforms. I rewrote all the test analysis logic and I still afraid I broke something. The interface needs to be gelled a bit. Half of it is unimplemented. I also need to merge some fixes from the mainline into 2.00. I'll see if I can kick out 2.00_02 today. -- Michael G. Schwern <[EMAIL PROTECTED]>http://www.pobox.com/~schwern/ Perl6 Quality Assurance <[EMAIL PROTECTED]> Kwalitee Is Job One Schwern Unit: a positive but insignificant quantity