Update build_script

[JensTimmerman]

ps make targets don't exist anymore in the new doc Makefile

[coveralls]

Coverage remained the same when pulling 17c632d on JensTimmerman:patch-1 into 7b31baf on pyca:master.

[coveralls]

Coverage remained the same when pulling b3e3991 on JensTimmerman:patch-1 into 7b31baf on pyca:master.

[exarkun]

Thanks.

Removing ps makes sense to me. There is no ps target at all and so that can never work.

Removing man makes less sense to me. Maybe setup.py should define a doc extra that declares a Sphinx >= 1.0 dependency. On the other hand, pyOpenSSL includes no man pages. So perhaps it doesn't matter.

On the other hand, I have no idea what rpm/build_script is actually used for. I don't know why it's in the repository or what the benefits of having it working (or the benefit of keeping it at all) are...

Since you noticed this problem, perhaps you know what it's good for? 😄

[JensTimmerman]

In our environment we create rpm's of python packages on a few build servers, using python setup.py bdist_rpm
These then get put in our repositories and installed on our servers.

setup.py bdist_rpm triggers the run of this build_script, which failed because of the nonexisting ps target.

I was able to create a working rpm of my branch, (so without the ps target)
The man wasn't there in the first place, I tried adding it, but it failed because of a dependency, and I didn't want to put time into installing the dependency on our build servers. (and this would indeed require patching of the setup.py)

So, on the down side, using bdist_rpm still doesn't create man pages in the rpm (but I do not really need them, not if there is a text and html version on the system)
on the up side, something that doesn't work without this patch now works.

I am willing to rebase these 3 commits into 1 if you prefer this, I was quickly testing stuff using the web interface, resulting in 3 tries.

[exarkun]

Thanks for explaining that!

It sounds like another good change to make would be to have travis run setup.py bdist_rpm to ensure this continues to work in the future. I ... think ... I know how to do that easily, I'll submit a PR against this branch to add that change if I can work it out.

I am willing to rebase these 3 commits into 1 if you prefer this, I was quickly testing stuff using the web interface, resulting in 3 tries.

That's okay. I generally prefer accurate history to idealized, fictionalized history. 😄 I can see the benefits to the latter but they mostly require everyone to create perfect histories and that's a lot more work I'm not going to ask every pyOpenSSL contributor to do.

[hynek]

Could you rebase and add a target to tox.ini that exercises bdist_rpm please?

[coveralls]

Coverage increased (+1.24%) to 96.27% when pulling 35204f7 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage remained the same at 95.03% when pulling 35204f7 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[JensTimmerman]

@hynek : squached and rebased on top of master.
I'm not sure how to create a tox.ini that works on all systems, python setup.py bdist_rpm depends on the 'rpm' binary being installed, I'm not familliar enought with tox to figure out how to set the rpm binary as a dependency.

[hynek]

are you on an rpm-based OS? if not, there’s packages for all of them so I would just presume the presence of rpm and add the dep to Travis (which is Ubuntu based).

[coveralls]

Coverage decreased (-2.35%) to 92.67% when pulling fb06467 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage decreased (-2.35%) to 92.67% when pulling fb06467 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage decreased (-2.35%) to 92.67% when pulling 2db672e on JensTimmerman:patch-1 into 389f78b on pyca:master.

[hynek]

Ugh, can you read that failure? Unless it’s trivial to fix I think it’s not worth the trouble and we can't have nice things tests for it. :(

[JensTimmerman]

I'm not very familliar with the tox.ini/travis syntax, it works on my end with the requires='/' but the quotes around the / seem to disappear for some reason, might need to escape them?

[coveralls]

Coverage remained the same at 95.03% when pulling 03b0a3f on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage remained the same at 95.03% when pulling 03b0a3f on JensTimmerman:patch-1 into 389f78b on pyca:master.

[hynek]

ah meh reset to 35204f7 and we’ll merge it. This amount of complexity is not worth it. if it breaks, someone will complain (like you did :)).

[coveralls]

Coverage remained the same at 95.03% when pulling 076dd10 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage remained the same at 95.03% when pulling 076dd10 on JensTimmerman:patch-1 into 389f78b on pyca:master.

[coveralls]

Coverage increased (+1.24%) to 96.27% when pulling 6b2f911 on JensTimmerman:patch-1 into cc44fcd on pyca:master.

[coveralls]

Coverage remained the same at 95.03% when pulling 6b2f911 on JensTimmerman:patch-1 into cc44fcd on pyca:master.

[JensTimmerman]

@hynek reset, I've left it at cc44fcd, since I can't figure out why lynx would be needed, I don't have it on my system and it works.

[hynek]

thanks!