Ticket #1684 (closed enhancement: wontfix)

Opened 15 months ago

Last modified 8 weeks ago

Option to disable comments building

Reported by: robe Owned by: pramsey
Priority: medium Milestone: PostGIS 2.1.0
Component: postgis Version: trunk
Keywords: Cc:

Description

I know I was the one that asked for it in the first place, so it's ironic I'm asking to please take this out and remove xsltproc as a requirement.

It's driving me nuts also please remove the validity checking. That is really slowing things down and fails half the time if the network download gets interrupted. Lately my compile has been failing because the validity checks complain about things that are valid but presumable the mathml download is paetially there and then I have to start all over again. This has been happening fo rhte past week or so.

Change History

Changed 15 months ago by strk

Are comments installed with 'CREATE EXTENSION' ?

Changed 15 months ago by robe

Yes they are but my feeling is that these should be packaged as part of the tar ball because they are the same regardless of OS. So I personally would be fine if extensions don't build if you disable comment building.

Alternatively we could have the include comments part of extension be conditional on if you have xsltproc. Whether the comments are present or not, it's still considered the same extension by postgresql since comments are an appendage of functions similar to how function specific variable settings are.

Changed 15 months ago by strk

ok, back on the analysis: my "make" isn't descending in the doc/ directory. is yours ? consequently, "make install" doesn't try either.

"make check" is the only one that tries, due to an explicit request about testing comments, which I think makes sense for those running "make check".

So, this ticket sounds kind of misleading as we don't really need an option to disable comments building because comments building are built explicitly. That's to say the ticket should mention running "make check" or something. Or are you really having problems unrelated to "make check" ?

Changed 15 months ago by robe

I think make check is something people would want to do anyway regardless if they are building comments or note. I'm thinking with my package maintainer hat on.

As a package maintainer 1) want to be able to build from the tar ball and not have to have requirements that are not necessary. e.g if comments is the same across all OS, it should be in the tar ball and if extensions are the same across all OS it should be in the tar ball.

xsltproc is not so bad except for that nasty MathML junk (that is hard to configure). If the whole purpose is to build comments, I really shouldn't be required to sacrifice a goat just to do a make check.

2) I would want to be able to do make check to make sure my package is sound.

Changed 12 months ago by robe

  • milestone changed from PostGIS 2.0.1 to PostGIS 2.1.0

Changed 6 months ago by pramsey

Having removed MathML (r10663) I find makes things a lot faster... close?

Changed 8 weeks ago by robe

  • status changed from new to closed
  • resolution set to wontfix

this is less annoying now that we got rid of that damn mathML. So I'll go with I can live with it.

Note: See TracTickets for help on using tickets.