Unify the format of the write_error_msg

[betanalpha]

I had previously updated the text accompanying errors in command.hpp (which gets called for errors during initialization) but not the accompanying text in base_hamiltonian.hpp (which gets called during sampling). Only constant text has been modified -- no implementation details are touched so no tests should be affected.

Not an immediate priority but would be nice to add before 2.0 and could easily be wrapped up into another request to avoid another round of testing.

[stan-buildbot]

• code review
  
  Can one of the admins verify this patch?

[bob-carpenter]

My bet is that this will not cut down the number
of messages to the list about broken models due
to this message and it's not going to cut down the
pain it causes in the R output (you lose track of
which chain you're on and where).

But hey, everyone else wants it in there, so we'll
just deal with the mailing list as best we can and
hope users don't abandon Stan because they think it's
buggy (my deeper worry). You can see a recent exchange
where I explained this and the user said they were
still nervous about seeing so many messages (I really
can't blame them).

• Bob

On 9/11/13 9:31 AM, Michael Betancourt wrote:

I had previously updated the text accompanying errors in command.hpp (which gets called for errors during
initialization) but not the accompanying text in base_hamiltonian.hpp (which gets called during sampling). Only constant
text has been modified -- no implementation details are touched so no tests should be affected.

Not an immediate priority but would be nice to add before 2.0 and could easily be wrapped up into another request to
avoid another round of testing.

---

    You can merge this Pull Request by running

git pull https://github.com/stan-dev/stan feature/unify_write_error_msg

Or view, comment on, or merge it at:

#213

    Commit Summary

• Unify the format of the write_error_msg text during initialization and sampling.

  File Changes
  

• M src/stan/gm/command.hpp https://github.com/stan-dev/stan/pull/213/files#diff-0 (5)

• M src/stan/mcmc/hmc/hamiltonians/base_hamiltonian.hpp https://github.com/stan-dev/stan/pull/213/files#diff-1 (17)

  Patch Links:
  

• https://github.com/stan-dev/stan/pull/213.patch

• https://github.com/stan-dev/stan/pull/213.diff

[betanalpha]

This just makes the output consistent -- I make no claim that it solves the current problem with scaring users.

On Sep 11, 2013, at 6:30 PM, Bob Carpenter notifications@github.com wrote:

My bet is that this will not cut down the number
of messages to the list about broken models due
to this message and it's not going to cut down the
pain it causes in the R output (you lose track of
which chain you're on and where).

But hey, everyone else wants it in there, so we'll
just deal with the mailing list as best we can and
hope users don't abandon Stan because they think it's
buggy (my deeper worry). You can see a recent exchange
where I explained this and the user said they were
still nervous about seeing so many messages (I really
can't blame them).

• Bob

On 9/11/13 9:31 AM, Michael Betancourt wrote:

I had previously updated the text accompanying errors in command.hpp (which gets called for errors during
initialization) but not the accompanying text in base_hamiltonian.hpp (which gets called during sampling). Only constant
text has been modified -- no implementation details are touched so no tests should be affected.

Not an immediate priority but would be nice to add before 2.0 and could easily be wrapped up into another request to
avoid another round of testing.

---

You can merge this Pull Request by running

git pull https://github.com/stan-dev/stan feature/unify_write_error_msg

Or view, comment on, or merge it at:

#213

Commit Summary

• Unify the format of the write_error_msg text during initialization and sampling.

File Changes

• M src/stan/gm/command.hpp https://github.com/stan-dev/stan/pull/213/files#diff-0 (5)
• M src/stan/mcmc/hmc/hamiltonians/base_hamiltonian.hpp https://github.com/stan-dev/stan/pull/213/files#diff-1 (17)

Patch Links:

• https://github.com/stan-dev/stan/pull/213.patch
• https://github.com/stan-dev/stan/pull/213.diff

—
Reply to this email directly or view it on GitHub.

[syclik]

Jenkins, retest this please.

[stan-buildbot]

Test PASSed.
Refer to this link for build results: http://d1m1s1b1.stat.columbia.edu:8080/job/Stan%20Github%20Pull%20Requests/273/