Skip Menu |

This queue is for tickets about the CPANPLUS CPAN distribution.

Report information
The Basics
Id: 4702
Status: resolved
Priority: 0/
Queue: CPANPLUS

People
Owner: Nobody in particular
Requestors: daxelrod-cpanplus-rt [...] cpan.danonline.net
RSOD [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Normal
Broken in: 0.048
Fixed in: (no value)



Subject: Declining sending test results for a distribution prevents me from *ever* sending results for that distribution
If I tell CPANPLUS not to report the testing result for a module to CPAN Testers, it causes CPANPLUS to think I *have* sent results for this module, even though it does not send them. This means that it will not allow me to send results of a subsequent test of that module distribution. To reproduce: 0. Make sure the configuration option 'cpantest' is set to '1', and that CPANPLUS has the necessary prerequisites to send test reports to CPAN-Testers. 1. Install any module using cpanplus, default shell, by typing "i Module::Name", where Module::Name is any module. 2. When CPANPLUS asks "Report Module-Name-0.01's testing result?", answer "n". No report will be sent, which is the correct behavior. 3. Type "t Module::Name". In fact, you can trigger this behavior whenever you test a module, decline to send the test results, and then test it again. Expected Result: Once the module is done being tested, CPANPLUS *again* asks "Report Module-Name-0.01's testing result?", because Module-Name-0.01's testing result has not yet been sent to CPAN Testers. Observed Result: Once the module is done being tested again, CPANPLUS says "You have already sent a report for Module-Name-0.01, skipping." and does not give me the option of sending a report. This becomes a problem when a module fails to install because of some mistake of mine, rather than the author's. I decline to send a bogus report that the module has failed. I then fix my mistake, and rerun the module install. CPANPLUS will not allow me to report the module's testing result this time around, even though I never reported the test results. CPANPLUS is a wonderful, wonderful module, and a pleasure to use. Thank you for developing it. A transcript: Show quoted text
_________________ [Max:~] qwertyd% perl -v This is perl, v5.6.0 built for darwin [Max:~] qwertyd% uname -a Darwin Max.local. 6.5 Darwin Kernel Version 6.5: Mon Apr 7 17:05:38 PDT 2003; root:xnu/xnu-344.32.obj~1/RELEASE_PPC Power Macintosh powerpc [Max:~] qwertyd% cpanp CPANPLUS::Shell::Default -- CPAN exploration and modules installation (v0.03) *** Please report bugs to <cpanplus-bugs@lists.sourceforge.net>. *** Using CPANPLUS::Backend v0.048. ReadLine support enabled. CPAN Terminal> i Acme::Tango Checking if source files are up to date Retrieving /Users/qwertyd/.cpanplus/mailrc.2.08.stored Retrieving /Users/qwertyd/.cpanplus/dslip.2.08.stored Retrieving /Users/qwertyd/.cpanplus/packages.2.08.stored Installing: Acme::Tango #...Snip.... Report Acme-Tango-0.04's testing result (PASS)?: [y/N]: n Successfully installed Acme::Tango All modules installed successfully CPAN Terminal> t Acme::Tango Testing: Acme::Tango #...Snip.... You have already sent a report for Acme-Tango-0.04, skipping. Successfully tested Acme::Tango All modules tested successfully
Subject: A lot of things prevents me from *ever* sending results for that distribution
From: rrwo [...] cpan.org
[guest - Wed Dec 17 18:55:07 2003]: Show quoted text
> If I tell CPANPLUS not to report the testing result for a module to > CPAN Testers, it causes CPANPLUS to think I *have* sent results for > this module, even though it does not send them. This means that it > will not allow me to send results of a subsequent test of that > module distribution.
Similarly, if CPANPLUS decides on its own not to send a test result, it marks the test result as having been sent already. And, if a test result has failed (and been sent out), and one re-runs the test successfully after diagnosing the problem (suppose a prereq was not installed, but one sent the test report out anyway), one cannot send the new test result out. And, if a test result bombs completely (unable to compile C code, or failed pre-reqs) and skips sending the test out, one is unable to send any tests out for other modules one tests... it will always say problem with prereqs and skipping. One has to quit CPANPLUS and re-enter it. And (mentioned in a closed issue), if the send fails due to a problem with Test::Reporter or Mail::Mailer, it assumes the test has been sent out and will not let one resend the test. A workaround should at least let one use a special command to resend test results, with maybe a check against website and an "Are you really really really sure?" question or two.
As of 0.050_02 (not out yet), t --force Acme::Tango submits the test results every time. - R. [guest - Fri Jun 4 14:58:43 2004]: Show quoted text
> [guest - Wed Dec 17 18:55:07 2003]: >
> > If I tell CPANPLUS not to report the testing result for a module to > > CPAN Testers, it causes CPANPLUS to think I *have* sent results for > > this module, even though it does not send them. This means that it > > will not allow me to send results of a subsequent test of that > > module distribution.
> > Similarly, if CPANPLUS decides on its own not to send a test result, it > marks the test result as having been sent already. > > And, if a test result has failed (and been sent out), and one re-runs > the test successfully after diagnosing the problem (suppose a prereq was > not installed, but one sent the test report out anyway), one cannot send > the new test result out. > > And, if a test result bombs completely (unable to compile C code, or > failed pre-reqs) and skips sending the test out, one is unable to send > any tests out for other modules one tests... it will always say problem > with prereqs and skipping. One has to quit CPANPLUS and re-enter it. > > And (mentioned in a closed issue), if the send fails due to a problem > with Test::Reporter or Mail::Mailer, it assumes the test has been sent > out and will not let one resend the test. > > A workaround should at least let one use a special command to resend > test results, with maybe a check against website and an "Are you really > really really sure?" question or two.