Bug#828267: [Pkg-citadel-devel] Bug#828267: reopen, it is still valid
On 2016-12-11 21:07:14 [+0100], Michael Meskes wrote: > Sorry, wasn't precise enough it seems. I was wondering why you left out > the third possible option, namely the compatibility API in libssl 1.1. Ach. It does not work that way, this "compatibility" API is defined by the openssl build and should not be changed by a program. It only works for a few functions like SSL_load_error_strings() which point to the right functions. Those are enabled in the Debian build. The *main* problem is that structs like SSL are opaque and can't be dereferenced. There is no "compatibility" for that. Therefore the two options I mentioned are the valid candidates (a third might be to switch away from openssl to a different toolkit). > Michael Sebastian
Bug#828267: [Pkg-citadel-devel] Bug#828267: reopen, it is still valid
> > What are you trying to do here? Reopen 828267 and merge with > > itself? > > There is no other bug mentioned. What do I miss? > > Yeah. Not very smart. I intended to merge it with 846543. So there is already 846543 to track this problem. I don't really see the point in reopening 828267 and then merging it, but ok. > > Care to explain? The CFLAG change did make the package compile, > > install > > and run, so why reopen the bug? > > The bug was created and it was mentioned that this package needs > changes in > order to get it compiled against the new openssl ABI which is in > experimental. > Your upload of the "fixed" package was performed on 2016-10-30 and > was built > against openssl 1.0.2. So even if you would have done nothing, your Sure, because there was no 1.1.0 to build against for unstable. > package > would been built successfully *but* against a 1.0.2. The test should > have been > done against the version in experimental. Actually I did, but apparently something went wrong. > On 2016-11-01 openssl 1.1.0 was uploaded to unstable. From this point > in time > your package would have fail to build. So I *think* the change in the > CFLAGS > is a nop. It definitely is, at the very least because it contains a copy error, sigh. > So the bug was created because the package did not compile against > openssl > 1.1.0 and now #846543 was created which mostly a dupe of this one. Thus it would have worked to use that one. > That means you have a valid RC bug which should be solved in order to > get this > package ready for the release and I mentioned two options people are > doing. > Most of what I wrote is from the transition bug #827061. The > libssl1.0-dev is > provided by openssl 1.0.2 and is intended to provide the 1.0.2 API > for package > which can't be fixed in time for release. Sorry, wasn't precise enough it seems. I was wondering why you left out the third possible option, namely the compatibility API in libssl 1.1. Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Meskes at (Debian|Postgresql) dot Org Jabber: michael at xmpp dot meskes dot org VfL Borussia! Força Barça! SF 49ers! Use Debian GNU/Linux, PostgreSQL
Bug#828267: [Pkg-citadel-devel] Bug#828267: reopen, it is still valid
control: merge -1 846543 On 2016-12-11 09:54:42 [+0100], Michael Meskes wrote: > What are you trying to do here? Reopen 828267 and merge with itself? > There is no other bug mentioned. What do I miss? Yeah. Not very smart. I intended to merge it with 846543. > > from the change [0] you use say that a compatible API is used but the > > CFLAG change makes no sense. This is probably a miss understanding. > > Care to explain? The CFLAG change did make the package compile, install > and run, so why reopen the bug? The bug was created and it was mentioned that this package needs changes in order to get it compiled against the new openssl ABI which is in experimental. Your upload of the "fixed" package was performed on 2016-10-30 and was built against openssl 1.0.2. So even if you would have done nothing, your package would been built successfully *but* against a 1.0.2. The test should have been done against the version in experimental. On 2016-11-01 openssl 1.1.0 was uploaded to unstable. From this point in time your package would have fail to build. So I *think* the change in the CFLAGS is a nop. > > You > > have two choices: > > ... > > Says who? Again, without any explanation as to why, I don't see any > reason to act. So the bug was created because the package did not compile against openssl 1.1.0 and now #846543 was created which mostly a dupe of this one. That means you have a valid RC bug which should be solved in order to get this package ready for the release and I mentioned two options people are doing. Most of what I wrote is from the transition bug #827061. The libssl1.0-dev is provided by openssl 1.0.2 and is intended to provide the 1.0.2 API for package which can't be fixed in time for release. > Michael Sebastian
Bug#828267: [Pkg-citadel-devel] Bug#828267: reopen, it is still valid
> control: reopen -1 > control: merge -1 828267 What are you trying to do here? Reopen 828267 and merge with itself? There is no other bug mentioned. What do I miss? > from the change [0] you use say that a compatible API is used but the > CFLAG change makes no sense. This is probably a miss understanding. Care to explain? The CFLAG change did make the package compile, install and run, so why reopen the bug? > You > have two choices: > ... Says who? Again, without any explanation as to why, I don't see any reason to act. Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Meskes at (Debian|Postgresql) dot Org Jabber: michael at xmpp dot meskes dot org VfL Borussia! Força Barça! SF 49ers! Use Debian GNU/Linux, PostgreSQL